US 20070064943 A1
The present invention provides systems and methods for secure transaction management and electronic rights protection. Electronic appliances such as computers equipped in accordance with the present invention help to ensure that information is accessed and used only in authorized ways, and maintain the integrity, availability, and/or confidentiality of the information. Such electronic appliances provide a distributed virtual distribution environment (VDE) that may enforce a secure chain of handling and control, for example, to control and/or meter or otherwise monitor use of electronically stored or disseminated information. Such a virtual distribution environment may be used to protect rights of various participants in electronic commerce and other electronic or electronic-facilitated transactions. Distributed and other operating systems, environments and architectures, such as, for example, those using tamper-resistant hardware-based processors, may establish security at each node. These techniques may be used to support an all-electronic information distribution, for example, utilizing the “electronic highway.”
91. A method of using a governed item at a processing arrangement, including the following steps:
receiving a first rule at the processing arrangement;
receiving a second rule at the processing arrangement, the second rule being received independently from the first rule and from the governed item;
wherein one or both of the first rule and the second rule specify at least one permitted or prohibited use of the governed item; and
at the processing arrangement, employing the first rule and the second rule to securely govern at least one aspect of access to or other use of the governed item.
92. The method of
the first rule is directly or indirectly received from a first entity; and
the second rule is directly or indirectly received from a second entity different from the first entity.
93. The method of
the step of receiving the first rule includes receiving the governed item along with the first rule.
94. The method of
at least a portion of the governed item is received in an encrypted state; and
the step of employing the first rule and the second rule to securely govern at least one aspect of access to or other use of the governed item includes decrypting at least a portion of the governed item.
95. The method of
96. The method of
storing audit-related information relating to the access to or other use of the governed item.
97. A method for using a governed item, the method including the following steps:
encrypting at least a portion of the governed item;
storing the governed item in a memory of a first processing arrangement located at a first site;
receiving, at the first processing arrangement, a first rule set made up of one or more rules, the first rule set being received directly or indirectly from a second processing arrangement located at a second site located remotely from the first site;
wherein the first rule set includes at least one rule specifying at least one permitted or prohibited use of the governed item;
at the first processing arrangement, decrypting at least a portion of the governed item, the decryption being governed at least in part by one or more of the first rule set rules;
at the first processing arrangement, making a use of the governed item, the use being governed at least in part by one or more of the first rule set rules; and
at the first processing arrangement, storing audit-related information relating to the use of the governed item.
98. The method of
using at least a portion of the audit-related information to determine a payment.
99. The method of
100. The method of
101. The method of
102. The method of
receiving, at the first processing arrangement, a second rule set made up of one or more rules, the second rule set being received separately from the first rule set.
103. The method of
the second rule set is directly or indirectly received from a third processing arrangement located at a third site remote from the first site and from the second site.
104. The method of
using a rule from the second rule set to at least in part govern an aspect of access to or use of the governed item.
105. The method of
the step of using a rule from the second rule set includes at least in part governing an attempt to transfer at least a portion of the governed data item from the first data processing arrangement to a different data processing arrangement.
106. A method of governing an operation at a processing arrangement, the method including:
at the processing arrangement, receiving a first control directly or indirectly from a first party;
at the processing arrangement, receiving a second control directly or indirectly from a second party;
at the processing arrangement, using the first control and the second control to at least in part govern a use of an item; and
wherein one or both of the first control and the second control specify at least one permitted or prohibited use of the item.
107. The method of
storing a first type of audit-related information relating to the use of the item, the storing being required by the first control; and
storing a second type of audit-related information relating to the use of the item, the storing being required by the second control.
108. The method of
109. A method including:
storing a first control in a memory of a processing arrangement;
at the processing arrangement, receiving a data item, the data item being at least partially encrypted;
at the processing arrangement, receiving a second control;
wherein one or both of the first control and the second control specify at least one permitted or prohibited use of the data item;
using the first or second control to at least in part govern the decryption of at least a portion of the data item; and
using the first control or the second control to govern an aspect of access to or other use of the data item.
110. The method of
111. A method of controlling an operation at a processing arrangement including a memory, a removable memory reader, and a communications port, the method including:
inserting a removable memory into the removable memory reader;
detecting a governed item stored in the removable memory, the governed item being at least in part encrypted;
receiving a first digital control through the communications port, the first digital control specifying at least one permitted or prohibited use of the governed item;
using at least a portion of the governed item, the use being governed at least in part by the first digital control.
112. The method of
113. The method of
114. The method of
115. The method of
This is a continuation of application Ser. No. 09/678,252, filed Oct. 3, 2000, which is a continuation of application Ser. No. 09/328,671, filed Jun. 9, 1999, now U.S. Pat. No. 6,389,402, which is a continuation of application Ser. No. 08/964,333, filed Nov. 4, 1997, now U.S. Pat. No. 5,982,891, which is a continuation of application Ser. No. 08/388,107, filed Feb. 13, 1995, now abandoned, all of which are incorporated herein by reference.
This invention generally relates to computer and/or electronic security.
More particularly, this invention relates to systems and techniques for secure transaction management. This invention also relates to computer-based and other electronic appliance-based technologies that help to ensure that information is accessed and/or otherwise used only in authorized ways, and maintains the integrity, availability, and/or confidentiality of such information and processes related to such use.
The invention also relates to systems and methods for protecting rights of various participants in electronic commerce and other electronic or electronically-facilitated transactions.
The invention also relates to secure chains of handling and control for both information content and information employed to regulate the use of such content and consequences of such use. It also relates to systems and techniques that manage, including meter and/or limit and/or otherwise monitor use of electronically stored and/or disseminated information. The invention particularly relates to transactions, conduct and arrangements that make use of, including consequences of use of, such systems and/or techniques.
The invention also relates to distributed and other operating systems, environments and architectures. It also generally relates to secure architectures, including, for example, tamper-resistant hardware-based processors, that can be used to establish security at each node of a distributed system.
Telecommunications, financial transactions, government processes, business operations, entertainment, and personal business productivity all now depend on electronic appliances. Millions of these electronic appliances have been electronically connected together. These interconnected electronic appliances comprise what is increasingly called the “information highway.” Many businesses, academicians, and government leaders are concerned about how to protect the rights of citizens and organizations who use this information (also “electronic” or “digital”) highway.
Today, virtually anything that can be represented by words, numbers, graphics, or system of commands and instructions can be formatted into electronic digital information. Television, cable, satellite transmissions, and on-line services transmitted over telephone lines, compete to distribute digital information and entertainment to homes and businesses. The owners and marketers of this content include software developers, motion picture and recording companies, publishers of books, magazines, and newspapers, and information database providers. The popularization of on-line services has also enabled the individual personal computer user to participate as a content provider. It is estimated that the worldwide market for electronic information in 1992 was approximately $40 billion and is expected to grow to $200 billion by 1997, according to Microsoft Corporation. The present invention can materially enhance the revenue of content providers, lower the distribution costs and the costs for content, better support advertising and usage information gathering, and better satisfy the needs of electronic information users. These improvements can lead to a significant increase in the amount and variety of electronic information and the methods by which such information is distributed.
The inability of conventional products to be shaped to the needs of electronic information providers and users is sharply in contrast to the present invention. Despite the attention devoted by a cross-section of America's largest telecommunications, computer, entertainment and information provider companies to some of the problems addressed by the present invention, only the present invention provides commercially secure, effective solutions for configurable, general purpose electronic commerce transaction/distribution control systems.
Controlling Electronic Content
The present invention provides a new kind of “virtual distribution environment” (called “VDE” in this document) that secures, administers, and audits electronic information use. VDE also features fundamentally important capabilities for managing content that travels “across” the “information highway.” These capabilities comprise a rights protection solution that serves all electronic community members. These members include content creators and distributors, financial service providers, end-users, and others. VDE is the first general purpose, configurable, transaction control/rights protection solution for users of computers, other electronic appliances, networks, and the information highway.
A fundamental problem for electronic content providers is extending their ability to control the use of proprietary information. Content providers often need to limit use to authorized activities and amounts. Participants in a business model involving, for example, provision of movies and advertising on optical discs may include actors, directors, script and other writers, musicians, studios, publishers, distributors, retailers, advertisers, credit card services, and content end-users. These participants need the ability to embody their range of agreements and requirements, including use limitations, into an “extended” agreement comprising an overall electronic business model. This extended agreement is represented by electronic content control information that can automatically enforce agreed upon rights and obligations. Under VDE, such an extended agreement may comprise an electronic contract involving all business model participants. Such an agreement may alternatively, or in addition, be made up of electronic agreements between subsets of the business model participants. Through the use of VDE, electronic commerce can function in the same way as traditional commerce—that is commercial relationships regarding products and services can be shaped through the negotiation of one or more agreements between a variety of parties.
Commercial content providers are concerned with ensuring proper compensation for the use of their electronic information. Electronic digital information, for example a CD recording, can today be copied relatively easily and inexpensively. Similarly, unauthorized copying and use of software programs deprives rightful owners of billions of dollars in annual revenue according to the International Intellectual Property Alliance. Content providers and distributors have devised a number of limited function rights protection mechanisms to protect their rights. Authorization passwords and protocols, license servers, “lock/unlock” distribution methods, and non-electronic contractual limitations imposed on users of shrink-wrapped software are a few of the more prevalent content protection schemes. In a commercial context, these efforts are inefficient and limited solutions.
Providers of “electronic currency” have also created protections for their type of content. These systems are not sufficiently adaptable, efficient, nor flexible enough to support the generalized use of electronic currency. Furthermore, they do not provide sophisticated auditing and control configuration capabilities. This means that current electronic currency tools lack the sophistication needed for many real-world financial business models. VDE provides means for anonymous currency and for “conditionally” anonymous currency, wherein currency related activities remain anonymous except under special circumstances.
VDE Control Capabilities
VDE allows the owners and distributors of electronic digital information to reliably bill for, and securely control, audit, and budget the use of, electronic information. It can reliably detect and monitor the use of commercial information products. VDE uses a wide variety of different electronic information delivery means: including, for example, digital networks, digital broadcast, and physical storage media such as optical and magnetic disks. VDE can be used by major network providers, hardware manufacturers, owners of electronic information, providers of such information, and clearinghouses that gather usage information regarding, and bill for the use of, electronic information.
VDE provides comprehensive and configurable transaction management, metering and monitoring technology. It can change how electronic information products are protected, marketed, packaged, and distributed. When used, VDE should result in higher revenues for information providers and greater user satisfaction and value. Use of VDE will normally result in lower usage costs, decreased transaction costs, more efficient access to electronic information, re-usability of rights protection and other transaction management implementations, greatly improved flexibility in the use of secured information, and greater standardization of tools and processes for electronic transaction management. VDE can be used to create an adaptable environment that fulfills the needs of electronic information owners, distributors, and users; financial clearinghouses; and usage information analyzers and resellers.
Rights and Control Information
In general, the present invention can be used to protect the rights of parties who have:
Protecting the rights of electronic community members involves a broad range of technologies VDE combines these technologies in a way that creates a “distributed” electronic rights protection “environment.” This environment secures and protects transactions and other processes important for rights protection. VDE, for example, provides the ability to prevent, or impede, interference with and/or observation of, important rights related transactions and processes. VDE, in its preferred embodiment, uses special purpose tamper resistant Secure Processing Units (SPUs) to help provide a high level of security for VDE processes and information storage and communication.
The rights protection problems solved by the present invention are electronic versions of basic societal issues. These issues include protecting property rights, protecting privacy rights, properly compensating people and organizations for their work and risk, protecting money and credit, and generally protecting the security of information VDE employs a system that uses a common set of processes to manage rights issues in an efficient, trusted, and cost-effective way.
VDE can be used to protect the rights of parties who create electronic content such as, for example: records, games, movies, newspapers, electronic books and reference materials, personal electronic mail, and confidential records and communications. The invention can also be used to protect the rights of parties who provide electronic products, such as publishers and distributors; the rights of parties who provide electronic credit and currency to pay for use of products, for example, credit clearinghouses and banks; the rights to privacy of parties who use electronic content (such as consumers, business people, governments); and the privacy rights of parties described by electronic information, such as privacy rights related to information contained in a medical record, tax record, or personnel record.
In general, the present invention can protect the rights of parties who have:
VDE Functional Properties
VDE is a cost-effective and efficient rights protection solution that provides a unified, consistent system for securing and managing transaction processing. VDE can:
In addition, VDE:
VDE economically and efficiently fulfills the rights protection needs of electronic community members. Users of VDE will not require additional rights protection systems for different information highway products and rights problems—nor will they be required to install and learn a new system for each new information highway application.
VDE provides a unified solution that allows all content creators, providers, and users to employ the same electronic rights protection solution. Under authorized circumstances, the participants can freely exchange content and associated content control sets. This means that a user of VDE may, if allowed, use the same electronic system to work with different kinds of content having different sets of content control information. The content and control information supplied by one group can be used by people who normally use content and control information supplied by a different group. VDE can allow content to be exchanged “universally” and users of an implementation of the present invention can interact electronically without fear of incompatibilities in content control, violation of rights, or the need to get, install, or learn a new content control system.
The VDE securely administers transactions that specify protection of rights. It can protect electronic rights including, for example:
The VDE can enable a very broad variety of electronically enforced commercial and societal agreements. These agreements can include electronically implemented contracts, licenses, laws, regulations, and tax collection.
Contrast with Traditional Solutions
Traditional content control mechanisms often require users to purchase more electronic information than the user needs or desires. For example, infrequent users of shrink-wrapped software are required to purchase a program at the same price as frequent users, even though they may receive much less value from their less frequent use. Traditional systems do not scale cost according to the extent or character of usage and traditional systems can not attract potential customers who find that a fixed price is too high. Systems using traditional mechanisms are also not normally particularly secure. For example, shrink-wrapping does not prevent the constant illegal pirating of software once removed from either its physical or electronic package.
Traditional electronic information rights protection systems are often inflexible and inefficient and may cause a content provider to choose costly distribution channels that increase a product's price. In general these mechanisms restrict product pricing, configuration, and marketing flexibility. These compromises are the result of techniques for controlling information which cannot accommodate both different content models and content models which reflect the many, varied requirements, such as content delivery strategies, of the model participants. This can limit a provider's ability to deliver sufficient overall value to justify a given product's cost in the eyes of many potential users. VDE allows content providers and distributors to create applications and distribution networks that reflect content providers' and users' preferred business models. It offers users a uniquely cost effective and feature rich system that supports the ways providers want to distribute information and the ways users want to use such information. VDE supports content control models that ensure rights and allow content delivery strategies to be shaped for maximum commercial results.
Chain of Handling and Control
VDE can protect a collection of rights belonging to various parties having in rights in, or to, electronic information. This information may be at one location or dispersed across (and/or moving between) multiple locations. The information may pass through a “chain” of distributors and a “chain” of users. Usage information may also be reported through one or more “chains” of parties. In general, VDE enables parties that (a) have rights in electronic information, and/or (b) act as direct or indirect agents for parties who have rights in electronic information, to ensure that the moving, accessing, modifying, or otherwise using of information can be securely controlled by rules regarding how, when, where, and by whom such activities can be performed.
VDE Applications and Software
VDE is a secure system for regulating electronic conduct and commerce. Regulation is ensured by control information put in place by one or more parties. These parties may include content providers, electronic hardware manufacturers, financial service providers, or electronic “infrastructure” companies such as cable or telecommunications companies. The control information implements “Rights Applications.” Rights applications “run on” the “base software” of the preferred embodiment. This base software serves as a secure, flexible, general purpose foundation that can accommodate many different rights applications, that is, many different business models and their respective participant requirements.
A rights application under VDE is made up of special purpose pieces, each of which can correspond to one or more basic electronic processes needed for a rights protection environment. These processes can be combined together like building blocks to create electronic agreements that can protect the rights, and may enforce fulfillment of the obligations, of electronic information users and providers. One or more providers of electronic information can easily combine selected building blocks to create a rights application that is unique to a specific content distribution model. A group of these pieces can represent the capabilities needed to fulfill the agreement(s) between users and providers. These, pieces accommodate many requirements of electronic commerce including:
For electronic commerce, a rights application, under the preferred embodiment of the present invention, can provide electronic enforcement of the business agreements between all participants. Since different groups of components can be put together for different applications, the present invention can provide electronic control information for a wide variety of different products and markets. This means the present invention can provide a “unified,” efficient, secure, and cost-effective system for electronic commerce and data security. This allows VDE to serve as a single standard for electronic rights protection, data security, and electronic currency and banking.
In a VDE, the separation between a rights application and its foundation permits the efficient selection of sets of control information that are appropriate for each of many different types of applications and uses. These control sets can reflect both rights of electronic community members, as well as obligations (such as providing a history of one's use of a product or paying taxes on one's electronic purchases). VDE flexibility allows its users to electronically implement and enforce common social and commercial ethics and practices. By providing a unified control system, the present invention supports a vast range of possible transaction related interests and concerns of individuals, communities, businesses, and governments. Due to its open design, VDE allows (normally under securely controlled circumstances) applications using technology independently created by users to be “added” to the system and used in conjunction with the foundation of the invention. In sum, VDE provides a system that can fairly reflect and enforce agreements among parties. It is a broad ranging and systematic solution that answers the pressing need for a secure, cost-effective, and fair electronic environment.
The preferred embodiment of the present invention includes various tools that enable system designers to directly insert VDE capabilities into their products. These tools include an Application Programmer's Interface (“API”) and a Rights Permissioning and Management Language (“RPML”). The RPML provides comprehensive and detailed control over the use of the invention's features. VDE also includes certain user interface subsystems for satisfying the needs of content providers, distributors, and users.
Information distributed using VDE may take many forms. It may, for example, be “distributed” for use on an individual's own computer, that is the present invention can be used to provide security for locally stored data. Alternatively, VDE may be used with information that is dispersed by authors and/or publishers to one or more recipients. This information may take many forms including: movies, audio recordings, games, electronic catalog shopping, multimedia, training materials, E-mail and personal documents, object oriented libraries, software programming resources, and reference/record keeping information resources (such as business, medical, legal, scientific, governmental, and consumer databases).
Electronic rights protection provided by the present invention will also provide an important foundation for trusted and efficient home and commercial banking, electronic credit processes, electronic purchasing, true or conditionally anonymous electronic cash, and EDI (Electronic Data Interchange). VDE provides important enhancements for improving data security in organizations by providing “smart” transaction management features that can be far more effective than key and password based “go/no go” technology.
VDE normally employs an integration of cryptographic and other security technologies (e.g. encryption, digital signatures, etc.), with other technologies including: component, distributed, and event driven operating system technology, and related communications, object container, database, smart agent, smart card, and semiconductor design technologies.
A. VDE Solves Important Problems and Fills Critical Needs
The world is moving towards an integration of electronic information appliances. This interconnection of appliances provides a foundation for much greater electronic interaction and the evolution of electronic commerce. A variety of capabilities are required to implement an electronic commerce environment. VDE is the first system that provides many of these capabilities and therefore solves fundamental problems related to electronic dissemination of information
VDE allows electronic arrangements to be created involving two or more parties. These agreements can themselves comprise a collection of agreements between participants in a commercial value chain and/or a data security chain model for handling, auditing, reporting, and payment. It can provide efficient, reusable, modifiable, and consistent means for secure electronic content: distribution, usage control, usage payment, usage auditing, and usage reporting. Content may, for example, include:
VDE enables an electronic commerce marketplace that supports differing, competitive business partnerships, agreements, and evolving overall business models.
The features of VDE allow it to function as the first trusted electronic information control environment that can conform to, and support, the bulk of conventional electronic commerce and data security requirements. In particular, VDE enables the participants in a business value chain model to create an electronic version of traditional business agreement terms and conditions and further enables these participants to shape’ and evolve their electronic commerce models as they believe appropriate to their business requirements.
VDE offers an architecture that avoids reflecting specific distribution biases, administrative and control perspectives, and content types. Instead, VDE provides a broad-spectrum, fundamentally configurable and portable, electronic transaction control, distributing, usage, auditing, reporting, and payment operating environment. VDE is not limited to being an application or application specific toolset that covers only a limited subset of electronic interaction activities and participants. Rather, VDE supports systems by which such applications can be created, modified, and/or reused. As a result, the present invention answers pressing, unsolved needs by offering a system that supports a standardized control environment which facilitates interoperability of electronic appliances, interoperability of content containers, and efficient creation of electronic commerce applications and models through the use of a programmable, secure electronic transactions management foundation and reusable and extensible executable components. VDE can support a single electronic “world” within which most forms of electronic transaction activities can be managed.
To answer the developing needs of rights owners and content providers and to provide a system that can accommodate the requirements and agreements of all parties that may be involved in electronic business models (creators, distributors, administrators, users, credit providers, etc.), VDE supplies an efficient, largely transparent, low cost and sufficiently secure system (supporting both hardware/software and software only models). VDE provides the widely varying secure control and administration capabilities required for:
1. Different types of electronic content,
2. Differing electronic content delivery schemes,
3. Differing electronic content usage schemes,
4. Different content usage platforms, and
5. Differing content marketing and model strategies.
VDE may be combined with, or integrated into, many separate computers and/or other electronic appliances. These appliances typically include a secure subsystem that can enable control of content use such as displaying, encrypting, decrypting, printing, copying, saving, extracting, embedding, distributing, auditing usage, etc. The secure subsystem in the preferred embodiment comprises one or more “protected processing environments”, one or more secure databases, and secure “component assemblies” and other items and processes that need to be kept secured. VDE can, for example, securely control electronic currency, payments, and/or credit management (including electronic credit and/or currency receipt, disbursement, encumbering, and/or allocation) using such a “secure subsystem.”
VDE provides a secure, distributed electronic transaction management system for controlling the distribution and/or other usage of electronically provided and/or stored information. VDE controls auditing and reporting of electronic content and/or appliance usage. Users of VDE may include content creators who apply content usage, usage reporting, and/or usage payment related control information to electronic content and/or appliances for users such as end-user organizations, individuals, and content and/or appliance distributors. VDE also securely supports the payment of money owed (including money owed for content and/or appliance usage) by one or more parties to one or more other parties, in the form of electronic credit and/or currency.
Electronic appliances under control of VDE represent VDE ‘nodes’ that securely process and control; distributed electronic information and/or appliance usage, control information formulation, and related transactions. VDE can securely manage the integration of control information provided by two or more parties. As a result, VDE can construct an electronic agreement between VDE participants that represent a “negotiation” between, the control requirements of, two or more parties and enacts terms and conditions of a resulting agreement. VDE ensures the rights of each party to an electronic agreement regarding a wide range of electronic activities related to electronic information and/or appliance usage.
Through use of VDE's control system, traditional content providers and users can create electronic relationships that reflect traditional, non-electronic relationships. They can shape and modify commercial relationships to accommodate the evolving needs of, and agreements among, themselves. VDE does not require electronic content providers and users to modify their business practices and personal preferences to conform to a metering and control application program that supports limited, largely fixed functionality. Furthermore, VDE permits participants to develop business models not feasible with non-electronic commerce, for example, involving detailed reporting of content usage information, large numbers of distinct transactions at hitherto infeasibly low price points, “pass-along” control information that is enforced without involvement or advance knowledge of the participants, etc.
The present invention allows content providers and users to formulate their transaction environment to accommodate:
VDE's transaction management capabilities can enforce:
VDE can support “real” commerce in an electronic form, that is the progressive creation of commercial relationships that form, over time, a network of interrelated agreements representing a value chain business model. This is achieved in part by enabling content control information to develop through the interaction of (negotiation between) securely created and independently submitted sets of content and/or appliance control information. Different sets of content and/or appliance control information can be submitted by different parties in an electronic business value chain enabled by the present invention. These parties create control information sets through the use of their respective VDE installations. Independently, securely deliverable, component based control information allows efficient interaction among control information sets supplied by different parties.
VDE permits multiple, separate electronic arrangements to be formed between subsets of parties in a VDE supported electronic value chain model. These multiple agreements together comprise a VDE value chain “extended” agreement. VDE allows such constituent electronic agreements, and therefore overall VDE extended agreements, to evolve and reshape over time as additional VDE participants become involved in VDE content and/or appliance control information handling. VDE electronic agreements may also be extended as new control information is submitted by existing participants. With VDE, electronic commerce participants are free to structure and restructure their electronic commerce business activities and relationships. As a result, the present invention allows a competitive electronic commerce marketplace to develop since the use of VDE enables different, widely varying business models using the same or shared content.
A significant facet of the present invention's ability to broadly support electronic commerce is its ability to securely manage independently delivered VDE component objects containing control information (normally in the form of VDE objects containing one or more methods, data, or load module VDE components). This independently delivered control information can be integrated with senior and other pre-existing content control information to securely form derived control information using the negotiation mechanisms of the present invention. All requirements specified by this derived control information must be satisfied before VDE controlled content can be accessed or otherwise used. This means that, for example, all load modules and any mediating data which are listed by the derived control information as required must be available and securely perform their required function. In combination with other aspects of the present invention, securely, independently delivered control components allow electronic commerce participants to freely stipulate their business requirements and trade offs. As a result, much as with traditional, non-electronic commerce, the present invention allows electronic commerce (through a progressive stipulation of various control requirements by VDE participants) to evolve into forms of business that are the most efficient, competitive and useful.
VDE provides capabilities that rationalize the support of electronic commerce and electronic transaction management. This rationalization stems from the reusability of control structures and user interfaces for a wide variety of transaction management related activities. As a result, content usage control, data security, information auditing, and electronic financial activities, can be supported with tools that are reusable, convenient, consistent, and familiar. In addition, a rational approach—a transaction/distribution control standard—allows all participants in VDE the same foundation set of hardware control and security, authoring, administration, and management tools to support widely varying types of information, business market model, and/or personal objectives.
Employing VDE as a general purpose electronic transaction/distribution control system allows users to maintain a single transaction management control arrangement on each of their computers, networks, communication nodes, and/or other electronic appliances. Such a general purpose system can serve the needs of many electronic transaction management applications without requiring distinct, different installations for different purposes. As a result, users of VDE can avoid the confusion and expense and other inefficiencies of different, limited purpose transaction control applications for each different content and/or business model. For example, VDE allows content creators to use the same VDE foundation control arrangement for both content authoring and for licensing content from other content creators for inclusion into their products or for other use. Clearinghouses, distributors, content creators, and other VDE users can all interact, both with the applications running on their VDE installations, and with each other, in an entirely consistent manner, using and reusing (largely transparently) the same distributed tools, mechanisms, and consistent user interfaces, regardless of the type of VDE activity.
VDE prevents many forms of unauthorized use of electronic information, by controlling and auditing (and other administration of use) electronically stored and/or disseminated information. This includes, for example, commercially distributed content, electronic currency, electronic credit, business transactions (such as EDI), confidential communications, and the like. VDE can further be used to enable commercially provided electronic content to be made available to users in user defined portions, rather than constraining the user to use portions of content that were “predetermined” by a content creator and/or other provider for billing purposes.
VDE, for example, can employ:
VDE may be used to migrate most non-electronic, traditional information delivery models (including entertainment, reference materials, catalog shopping, etc.) into an adequately secure digital distribution and usage management and payment context. The distribution and financial pathways managed by a VDE arrangement may include:
financial and/or other clearinghouse(s),
and/or government agencies.
These distribution and financial pathways may also include:
Normally, participants in a VDE arrangement will employ the same secure VDE foundation. Alternate embodiments support VDE arrangements employing differing VDE foundations. Such alternate embodiments may employ procedures to ensure certain interoperability requirements are met.
Secure VDE hardware (also known as SPUs for Secure Processing Units), or VDE installations that use software to substitute for, or complement, said hardware (provided by Host Processing Environments (HPEs)), operate in conjunction with secure communications, systems integration software, and distributed software control information and support structures, to achieve the electronic contract/rights protection environment of the present invention. Together, these VDE components comprise a secure, virtual, distributed content and/or appliance control, auditing (and other administration), reporting, and payment environment. In some embodiments and where commercially acceptable, certain VDE participants, such as clearinghouses that normally maintain sufficiently physically secure non-VDE processing environments, may be allowed to employ HPEs rather VDE hardware elements and interoperate, for example, with VDE end-users and content providers. VDE components together comprise a configurable, consistent, secure and “trusted” architecture for distributed, asynchronous control of electronic content and/or appliance usage. VDE supports a “universe wide” environment for electronic content delivery, broad dissemination, usage reporting, and usage related payment activities.
VDE provides generalized configurability. This results, in part, from decomposition of generalized requirements for supporting electronic commerce and data security into a broad range of constituent “atomic” and higher level components (such as load modules, data elements, and methods) that may be variously aggregated together to form control methods for electronic commerce applications, commercial electronic agreements, and data security arrangements. VDE provides a secure operating environment employing VDE foundation elements along with secure independently deliverable VDE components that enable electronic commerce models and relationships to develop. VDE specifically supports the unfolding of distribution models in which content providers, over time, can expressly agree to, or allow, subsequent content providers and/or users to participate in shaping the control information for, and consequences of use of electronic content and/or appliances. A very broad range of the functional attributes important for supporting simple to very complex electronic commerce and data security activities are supported by capabilities of the present invention. As a result, VDE supports most types of electronic information and/or appliance: usage control (including distribution), security, usage auditing, reporting, other administration, and payment arrangements.
VDE, in its preferred embodiment, employs object software technology and uses object technology to form “containers” for delivery of information that is (at least in part) encrypted or otherwise secured. These containers may contain electronic content products or other electronic information and some or all of their associated permissions (control) information. These container objects may be distributed along pathways involving content providers and/or content users. They may be securely moved among nodes of a Virtual Distribution Environment (VDE) arrangement, which nodes operate VDE foundation software and execute control methods to enact electronic information usage control and/or administration models. The containers delivered through use of the preferred embodiment of the present invention may be employed both for distributing VDE control instructions (information) and/or to encapsulate and electronically distribute content that has been at least partially secured.
Content providers who employ the present invention may include, for example, software application and game publishers, database publishers, cable, television, and radio broadcasters, electronic shopping vendors, and distributors of information in electronic document, book, periodical, e-mail and/or other forms. Corporations, government agencies, and/or individual “end-users” who act as storers of, and/or distributors of, electronic information, may also be VDE content providers (in a restricted model, a user provides content only to himself and employs VDE to secure his own confidential information against unauthorized use by other parties). Electronic information may include proprietary and/or confidential information for personal or internal organization use, as well as information, such as software applications, documents, entertainment materials, and/or reference information, which may be provided to other parties. Distribution may be by, for example, physical media delivery, broadcast and/or telecommunication means, and in the form of “static” files and/or streams of data. VDE may also be used, for example, for multi-site “real-time” interaction such as teleconferencing, interactive games, or on-line bulletin boards, where restrictions on, and/or auditing of, the use of all or portions of communicated information is enforced.
VDE provides important mechanisms for both enforcing commercial agreements and enabling the protection of privacy rights. VDE can securely deliver information from one party to another concerning the use of commercially distributed electronic content. Even if parties are separated by several “steps” in a chain (pathway) of handling for such content usage information, such information is protected by VDE through encryption and/or other secure processing. Because of that protection, the accuracy of such information is guaranteed by VDE, and the information can be trusted by all parties to whom it is delivered. Furthermore, VDE guarantees that all parties can trust that such information cannot be received by anyone other than the intended, authorized, party(ies) because it is encrypted such that only an authorized party, or her agents, can decrypt it. Such information may also be derived through a secure VDE process at a previous pathway-of-handling location to produce secure VDE reporting information that is then communicated securely to its intended recipient's VDE secure subsystem. Because VDE can deliver such information securely, parties to an electronic agreement need not trust the accuracy of commercial usage and/or other information delivered through means other than those under control of VDE.
VDE participants in a commercial value, chain can be “commercially” confident (that is, sufficiently confident for commercial purposes) that the direct (constituent) and/or “extended” electronic agreements they entered into through the use of VDE can be enforced reliably. These agreements may have both “dynamic” transaction management related, aspects, such as content usage control information enforced through budgeting, metering, and/or reporting of electronic information and/or appliance use, and/or they may include “static” electronic assertions, such as an end-user using the system to assert his or her agreement to pay for services, not to pass to unauthorized parties electronic information derived from usage of content or systems, and/or agreeing to observe copyright laws. Not only can electronically reported transaction related information be trusted under the present invention, but payment may be automated by, the passing of payment tokens through a pathway of payment (which may or may not be the same as a pathway for reporting). Such payment can be contained within a VDE container created automatically by a VDE installation in response to control information (located, in the preferred embodiment, in one or more permissions records) stipulating the “withdrawal” of credit or electronic currency (such as tokens) from an electronic account (for example, an account securely maintained by a user's VDE installation secure subsystem) based upon usage of VDE controlled electronic content and/or appliances (such as governments, financial credit providers, and users).
VDE allows the needs of electronic commerce participants to be served and it can bind such participants together in a universe wide, trusted commercial network that can be secure enough to support very large amounts of commerce. VDE's security and metering secure subsystem core will be present at all physical locations where VDE related content is (a) assigned usage related control information (rules and mediating data), and/or (b) used. This core can perform security and auditing functions (including metering) that operate within a “virtual black box,” a collection of distributed, very secure VDE related hardware instances that are interconnected by secured information exchange (for example, telecommunication) processes and distributed database means. VDE further includes highly configurable transaction operating system technology, one or more associated libraries of load modules along with affiliated data, VDE related administration, data preparation, and analysis applications, as well as system software designed to enable VDE integration into host environments and applications. VDE's usage control information, for example, provide for property content and/or appliance related: usage authorization, usage auditing (which may include audit reduction), usage billing, usage payment, privacy filtering, reporting, and security related communication and encryption techniques.
VDE extensively employs methods in the form of software objects to augment configurability, portability, and security of the VDE environment. It also employs a software object architecture for VDE content containers that carries protected content and may also carry both freely available information (e.g, summary, table of contents) and secured content control information which ensures the performance of control information. Content control information governs content usage according to criteria set by holders of rights to an object's contents and/or according to parties who otherwise have rights associated with distributing such content (such as governments, financial credit providers, and users).
In part, security is enhanced by object methods employed by the present invention because the encryption schemes used to protect an object can efficiently be further used to protect the associated content control information (software control information and relevant data) from modification. Said object techniques also enhance portability between various computer and/or other appliance environments because electronic information in the form of content can be inserted along with (for example, in the same object container as) content control information (for said content) to produce a “published” object. As a result, various portions of said control information may be specifically adapted for different environments, such as for diverse computer platforms and operating systems, and said various portions may all be carried by a VDE container.
An objective of VDE is supporting a transaction/distribution control standard. Development of such a standard has many obstacles, given the security requirements and related hardware and communications issues, widely differing environments, information types, types of information usage, business and/or data security goals, varieties of participants, and properties of delivered information. A significant feature of VDE accommodates the many, varying distribution and other transaction variables by, in part, decomposing electronic commerce and data security functions into generalized capability modules executable within a secure hardware SPU and/or corresponding software subsystem and further allowing extensive flexibility in assembling, modifying, and/or replacing, such modules (e.g. load modules and/or methods) in applications run on a VDE installation foundation. This configurability and reconfigurability allows electronic commerce and data security participants to reflect their priorities and requirements through a process of iteratively shaping an evolving extended electronic agreement (electronic control model). This shaping can occur as content control information passes from one VDE participant to another and to the extent allowed by “in place” content control information. This process allows users of VDE to recast existing control information and/or add new control information as necessary (including the elimination of no longer required elements).
VDE supports trusted (sufficiently secure) electronic information distribution and usage control models for both commercial electronic content distribution and data security applications. It can be configured to meet the diverse requirements of a network of interrelated participants that may include content creators, content distributors, client administrators, end users, and/or clearinghouses and/or other content usage information users. These parties may constitute a network of participants involved in simple to complex electronic content dissemination, usage control, usage reporting, and/or usage payment. Disseminated content may include both originally provided and VDE generated information (such as content usage information) and content control information may persist through both chains (one or more pathways) of content and content control information handling, as well as the direct usage of content. The configurability provided by the present invention is particularly critical for supporting electronic commerce, that is enabling businesses to create relationships and evolve strategies that offer, competitive value. Electronic commerce tools that are not inherently configurable and interoperable will ultimately fail to produce products (and services) that meet both basic requirements and evolving needs of most commerce applications.
VDE's fundamental configurability will allow a broad range of competitive electronic commerce business models to flourish. It allows business models to be shaped to maximize revenues sources, end-user product value, and operating efficiencies. VDE can be employed to support multiple, differing models, take advantage of new revenue opportunities, and deliver product configurations most desired by users. Electronic commerce technologies that do not, as the present invention does:
will result in products that are often intrinsically more costly and less appealing and therefore less competitive in the marketplace.
Some of the key factors contributing to the configurability intrinsic to the present invention include:
Because of the breadth of issues resolved by the present invention, it can provide the emerging “electronic highway” with a single transaction/distribution control system that can, for a very broad range of commercial and data security models, ensure against unauthorized use of confidential and/or proprietary information and commercial electronic transactions. VDE's electronic transaction management mechanisms can enforce the electronic rights and agreements of all parties participating in widely varying business and data security models, and this can be efficiently achieved through a single VDE implementation within each VDE participant's electronic appliance. VDE supports widely varying business and/or data security models that can involve a broad range of participants at various “levels” of VDE content and/or content control information pathways of handling. Different content control and/or auditing models and agreements may be available on the same VDE installation. These models and agreements may control content in relationship to, for example, VDE installations and/or users in general; certain specific users, installations, classes and/or other groupings of installations and/or users; as well as to electronic content generally on a given installation, to specific properties, property portions, classes and/or other groupings of content.
Distribution using VDE may package both the electronic content and control information into the same VDE container, and/or may involve the delivery to an end-user site of different pieces of the same VDE managed property from plural separate remote locations and/or in plural separate VDE content containers and/or employing plural different delivery means. Content control information may be partially or fully delivered separately from its associated content to a user VDE installation in one or more VDE administrative objects. Portions of said control information may be delivered from one or more sources. Control information may also be available for use by access from a user's VDE installation secure sub-system to one or more remote VDE secure sub-systems and/or VDE compatible, certified secure remote locations. VDE control processes such as metering, budgeting, decrypting and/or fingerprinting, may as relates to a certain user content usage activity, be performed in a user's local VDE installation secure subsystem, or said processes may be divided amongst plural secure subsystems which may be located in the same user VDE installations and/or in a network server and in the user installation. For example, a local VDE installation may perform decryption and save any, or all of, usage metering information related to content and/or electronic appliance usage at such user installation could be performed at the server employing secure (e.g., encrypted) communications between said secure subsystems. Said server location may also be used for near real time, frequent, or more periodic secure receipt of content usage information from said user installation, with, for example, metered information being maintained only temporarily at a local user installation.
Delivery means for VDE managed content may include electronic data storage means such as optical disks for delivering one portion of said information and broadcasting and/or telecommunicating means for other portions of said information. Electronic data storage means may include magnetic media, optical media, combined magneto-optical systems, flash RAM memory, bubble memory, and/or other memory storage means such as huge capacity optical storage systems employing holographic, frequency, and/or polarity data storage techniques. Data storage means may also employ layered disc techniques, such as the use of generally transparent and/or translucent materials that pass light through layers of data carrying discs which themselves are physically packaged together as one thicker disc. Data carrying locations on such discs may be, at least in part, opaque.
VDE supports a general purpose foundation for secure transaction management, including usage control, auditing, reporting, and/or payment. This general purpose foundation is called “VDE Functions” (“VDEFs”). VDE also supports a collection of “atomic” application elements (e.g., load modules) that can be selectively aggregated together to form various VDEF capabilities called control methods and which serve as VDEF applications and operating system functions. When a host operating environment of an electronic appliance includes VDEF capabilities, it is called a “Rights Operating System” (ROS) VDEF load modules, associated data, and methods form a body of information that for the purposes of the present invention are called “control information.” VDEF control information may be specifically associated with one or more pieces of electronic content and/or it may be employed as a general component of the operating system capabilities of a VDE installation.
VDEF transaction control elements reflect and enact content specific and/or more generalized administrative (for example, general operating system) control information. VDEF capabilities which can generally take the form of applications (application models) that have more or less configurability which can be shaped by VDE participants, through the use, for example, of VDE templates, to employ specific capabilities, along, for example, with capability parameter data to reflect the elements of one or more express electronic agreements between VDE participants in regards to the use of electronic content such as commercially distributed products. These control capabilities manage the use of, and/or auditing of use of, electronic content, as well as reporting information based upon content use, and any payment for said use. VDEF capabilities may “evolve” to reflect the requirements of one or more successive parties who receive or otherwise contribute to a given set of control information. Frequently, for a VDE application for a given content model (such as distribution of entertainment on CD-ROM, content delivery from an Internet repository, or electronic catalog shopping and advertising, or some combination of the above) participants would be able to securely select from amongst available, alternative control methods and apply related parameter data, wherein such selection of control method and/or submission of data would constitute their “contribution” of control information. Alternatively, or in addition, certain control methods that have been expressly certified as securely interoperable and compatible with said application may be independently submitted by a participant as part of such a contribution. In the most general example, a generally certified load module (certified for a given VDE arrangement and/or content class) may be used with many or any VDE application that operates in nodes of said arrangement. These parties, to the extent they are allowed, can independently and securely add, delete, and/or otherwise modify the specification of load modules and methods, as well as add, delete or otherwise modify related information.
Normally the party who creates a VDE content container defines the general nature of the VDEF capabilities that will and/or may apply to certain electronic information. A VDE content container is an object that contains both content (for example, commercially distributed electronic information products such as computer software programs, movies, electronic publications or reference materials, etc.) and certain control information related to the use of the object's content. A creating party may make a VDE container available to other parties. Control information delivered by, and/or otherwise available for use with, VDE content containers comprise (for commercial content distribution purposes) VDEF control capabilities (and any associated parameter data) for electronic content. These capabilities may constitute one or more “proposed” electronic agreements (and/or agreement functions available for selection and/or use with parameter data) that manage the use and/or the consequences of use of such content and which can enact the terms and conditions of agreements involving multiple parties and their various rights and obligations.
A VDE electronic agreement may be explicit, through a user interface acceptance by one or more parties, for example by a “junior” party who has received control information from a “senior” party, or it may be a process amongst equal parties who individually assert their agreement. Agreement may also result from an automated electronic process during which terms and conditions are “evaluated” by certain VDE participant control information that assesses whether certain other electronic terms and conditions attached to content and/or submitted by another party are acceptable (do not violate acceptable control information criteria). Such an evaluation process may be quite simple, for example a comparison to ensure compatibility between a portion of, or all senior, control terms and conditions in a table of terms and conditions and the submitted control information of a subsequent participant in a pathway of content control information handling, or it may be a more elaborate process that evaluates the potential outcome of, and/or implements a negotiation process between, two or more sets of control information submitted by two or more parties. VDE also accommodates a semi-automated process during which one or more VDE participants directly, through user interface means, resolve “disagreements” between control information sets by accepting and/or proposing certain control information that may be acceptable to control information representing one or more other parties interests and/or responds to certain user interface queries for selection of certain alternative choices and/or for certain parameter information, the responses being adopted if acceptable to applicable senior control information.
When another party (other than the first applier of rules), perhaps through a negotiation process, accepts, and/or adds to and/or otherwise modifies, “in place” content control information, a VDE agreement between two or more parties related to the use of such electronic content may be created (so long as any modifications are consistent with senior control information). Acceptance of terms and conditions related to certain electronic content may be direct and express, or it may be implicit as a result of use of content (depending, for example, on legal requirements, previous exposure to such terms and conditions, and requirements of in place control information).
VDEF capabilities may be employed, and a VDE agreement may be entered into, by a plurality of parties without the VDEF capabilities being directly associated with the controlling of certain, specific electronic information. For example, certain one or more VDEF capabilities may be present at a VDE installation, and certain VDE agreements may have been entered into during the registration process for a content distribution application, to be used by such installation for securely controlling VDE content usage, auditing, reporting and/or payment. Similarly, a specific VDE participant may enter into a VDE user agreement with a VDE content or electronic appliance provider when the user and/or her appliance register with such provider as a VDE installation and/or user. In such events, VDEF in place control information available to the user VDE installation may require that certain VDEF methods are employed, for example in a certain sequence, in order to be able to use all and/or certain classes, of electronic content and/or VDE applications.
VDE ensures that certain prerequisites necessary for a given transaction to occur are met. This includes the secure execution of any required load modules and the availability of any required, associated data. For example, required load modules and data (e.g. in the form of a method) might specify that sufficient credit from an authorized source must be confirmed as available. It might further require certain one or more load modules execute as processes at an appropriate time to ensure that such credit will be used in order to pay for user use of the content. A certain content provider might, for example, require metering the number of copies made for distribution to employees of a given software program (a portion of the program might be maintained in encrypted form and require the presence of a VDE installation to run). This would require the execution of a metering method for copying of the property each time a copy was made for another employee. This same provider might also charge fees based on the total number of different properties licensed from them by the user and a metering history of their licensing of properties might be required to maintain this information.
VDE provides organization, community, and/or universe wide secure environments whose integrity is assured by processes securely controlled in VDE participant user installations (nodes). VDE installations, in the preferred embodiment, may include both software and tamper resistant hardware semiconductor elements. Such a semiconductor arrangement comprises, at least in part, special purpose circuitry that has been designed to protect against tampering with, or unauthorized observation of, the information and functions used in performing the VDE's control functions. The special purpose secure circuitry provided by the present invention includes at least one of a dedicated semiconductor arrangement known as a Secure Processing Unit (SPU) and/or a standard microprocessor, microcontroller, and/or other processing logic that accommodates the requirements of the present invention and functions as an SPU. VDE's secure hardware may be found incorporated into, for example, a fax/modem chip or chip pack, I/O controller, video display controller, and/or other available digital processing arrangements. It is anticipated that portions of the present invention's VDE secure hardware capabilities may ultimately be standard design elements of central processing units (CPUs) for computers and various other electronic devices.
Designing VDE capabilities into one or more standard microprocessor, microcontroller and/or other digital processing components may materially reduce VDE related hardware costs by employing the same hardware resources for both the transaction management uses contemplated by the present invention and for other, host electronic appliance functions. This means that a VDE SPU can employ (share) circuitry elements of a “standard” CPU. For example, if a “standard” processor can operate in protected mode and can execute VDE related instructions as a protected activity, then such an embodiment may provide sufficient hardware security for a variety of applications and the expense of a special purpose processor might be avoided. Under one preferred embodiment of the present invention, certain memory (e.g., RAM, ROM, NVRAM) is maintained during VDE related instruction processing in a protected mode (for example, as supported by protected mode microprocessors). This memory is located in the same package as the processing logic (e.g. processor). Desirably, the packaging and memory of such a processor would be designed using security techniques that enhance its resistance to tampering.
The degree of overall security of the VDE system is primarily dependent on the degree of tamper resistance and concealment of VDE control process execution and related data storage activities. Employing special purpose semiconductor packaging techniques can significantly contribute to the degree of security. Concealment and tamper-resistance in semiconductor memory (e.g., RAM, ROM, NVRAM) can be achieved, in part, by employing such memory within an SPU package, by encrypting data before it is sent to external memory (such as an external RAM package) and decrypting encrypted data within the CPU/RAM package before it is executed. This process is used for important VDE related data when such data is stored on unprotected media, for example, standard host storage, such as random access memory, mass storage, etc. In that event, a VDE SPU would encrypt data that results from a secure VDE execution before such data was stored in external memory.
Summary of Some Important Features Provided by VDE in Accordance with the Present Invention
VDE employs a variety of capabilities that serve as a foundation for a general purpose, sufficiently secure distributed electronic commerce solution. VDE enables an electronic commerce marketplace that supports divergent, competitive business partnerships, agreements, and evolving overall business models. For example, VDE includes features that:
VDE supports as many simultaneous predefined increment types as may be practical for a given type of content and business model.
Use of bitmap meters (including “regular” and “wide” bitmap meters) to record usage and/or purchase of information, in conjunction with other elements of the preferred embodiment of the present invention, uniquely supports efficient maintenance of usage history for: (a) rental, (b) flat fee licensing or purchase, (c) licensing or purchase discounts based upon historical usage variables, and (d) reporting to users in a manner enabling users to determine whether a certain item was acquired, or acquired within a certain time period (without requiring the use of conventional database mechanisms, which are highly inefficient for these applications). Bitmap meter methods record activities associated with electronic appliances, properties, objects, or portions thereof, and/or administrative activities that are independent of specific properties, objects, etc., performed by a user and/or electronic appliance such that a content and/or appliance provider and/or controller of an administrative activity can determine whether a certain activity has occurred at some point, or during a certain period, in the past (for example, certain use of a commercial electronic content product and/or appliance). Such determinations can then be used as part of pricing and/or control strategies of a content and/or appliance provider, and/or controller of an administrative activity. For example, the content provider may choose to charge only once for access to a portion of a property, regardless of the number of times that portion of the property is accessed by a user.
Generally, the extraction features of the present invention allow users to aggregate and/or disseminate and/or otherwise use protected electronic content information extracted from content container sources while maintaining secure VDE capabilities thus preserving the rights of providers in said content information after various content usage processes.
VDE control information (e.g., methods) that collectively control use of VDE managed properties (database, document, individual commercial product), are either shipped with the content itself (for example, in a content container) and/or one or more portions of such control information is shipped to distributors and/or other users in separably deliverable “administrative objects.” A subset of the methods for a property may in part be delivered with each property while one or more other subsets of methods can be delivered separately to a user or otherwise made available for use (such as being available remotely by telecommunication means). Required methods (methods listed as required for property and/or appliance use) must be available as specified if VDE controlled content (such as intellectual property distributed within a VDE content container) is to be used. Methods that control content may apply to a plurality of VDE container objects, such as a class or other grouping of such objects. Methods may also be required by certain users or classes of users and/or VDE installations and/or classes of installations for such parties to use one or more specific, or classes of, objects.
A feature of VDE provided by the present invention is that certain one or more methods can be specified as required in order for a VDE installation and/or user to be able to use certain and/or all content. For example, a distributor of a certain type of content might be allowed by “senior” participants (by content creators, for example) to require a method which prohibits end-users from electronically saving decrypted content, a provider of credit for VDE transactions might require an audit method that records the time of an electronic purchase, and/or a user might require a method that summarizes usage information for reporting to a clearinghouse (e.g. billing information) in a way that does not convey confidential, personal information regarding detailed usage behavior.
A further feature of VDE provided by the present invention is that creators, distributors, and users of content can select from among a set of predefined methods (if available) to control container content usage and distribution functions and/or they may have the right to provide new customized methods to control at least certain usage functions (such “new” methods may be required to be certified for trustedness and interoperability to the VDE installation and/or for of a group of VDE applications). As a result, VDE provides a very high degree of configurability with respect to how the distribution and other usage of each property or object (or one or more portions of objects or properties as desired and/or applicable) will be controlled. Each VDE participant in a VDE pathway of content control information may set methods for some or all of the content in a VDE container, so long as such control information does not conflict with senior control information already in place with respect to:
(1) certain or all VDE managed content,
(2) certain one or more VDE users and/or groupings of users,
(3) certain one or more VDE nodes and/or groupings of nodes, and/or
(4) certain one or more VDE applications and/or arrangements.
For example, a content creator's VDE control information for certain content can take precedence over other submitted VDE participant control information and, for example, if allowed by senior control information, a content distributor's control information may itself take precedence over a client administrator's control information, which may take precedence over an end-user's control information. A path of distribution participant's ability to set such electronic content control information can be limited to certain control information (for example, method mediating data such as pricing and/or sales dates) or it may be limited only to the extent that one or more of the participant's proposed control information conflicts with control information set by senior control information submitted previously by participants in a chain of handling of the property, or managed in said participant's VDE secure subsystem.
VDE control information may, in part or in full, (a) represent control information directly put in place by VDE content control information pathway participants, and/or (b) comprise control information put in place by such a participant on behalf of a party who does not directly handle electronic content (or electronic appliance) permissions records information (for example control information inserted by a participant on behalf of a financial clearinghouse or government agency). Such control information methods (and/or load modules and/or mediating data and/or component assemblies) may also be put in place by either an electronic automated, or a semi-automated and human assisted, control information (control set) negotiating process that assesses whether the use of one or more pieces of submitted control information will be integrated into and/or replace existing control information (and/or chooses between alternative control information based upon interaction with in-place control information) and how such control information may be used.
Control information may be provided by a party who does not directly participate in the handling of electronic content (and/or appliance) and/or control information for such content (and/or appliance). Such control information may be provided in secure form using VDE installation secure sub-system managed communications (including, for example, authenticating the deliverer of at least in part encrypted control information) between such not directly participating one or more parties' VDE installation secure subsystems, and a pathway of VDE content control information participant's VDE installation secure subsystem. This control information may relate to, for example, the right to access credit supplied by a financial services provider, the enforcement of regulations or laws enacted by a government agency, or the requirements of a customer of VDE managed content usage information (reflecting usage of content by one or more parties other than such customer) relating to the creation, handling and/or manner of reporting of usage information received by such customer. Such control information may, for example, enforce societal requirements such as laws related to electronic commerce.
VDE content control information may apply differently to different pathway of content and/or control information handling participants. Furthermore, permissions records rights may be added, altered, and/or removed by a VDE participant if they are allowed to take such action. Rights of VDE participants may be defined in relation to specific parties and/or categories of parties and/or other groups of parties in a chain of handling of content and/or content control information (e.g., permissions records). Modifications to control information that may be made by a given, eligible party or parties, may be limited in the number of modifications, and/or degree of modification, they may make.
At least one secure subsystem in electronic appliances of creators, distributors, auditors, clearinghouses, client administrators, and end-users (understanding that two or more of the above classifications may describe a single user) provides a “sufficiently” secure (for the intended applications) environment for:
Normally, most usage, audit, reporting, payment, and distribution control methods are themselves at least in part encrypted and are executed by the secure subsystem of a VDE installation. Thus, for example, billing and metering records can be securely generated and updated, and encryption and decryption keys are securely utilized, within a secure subsystem. Since VDE also employs secure (e.g. encrypted and authenticated) communications when passing information between the participant location (nodes) secure subsystems of a VDE arrangement, important components of a VDE electronic agreement can be reliably enforced with sufficient security (sufficiently trusted) for the intended commercial purposes. A VDE electronic agreement for a value chain can be composed, at least in part, of one or more subagreements between one or more subsets of the value chain participants. These subagreements are comprised of one or more electronic contract “compliance” elements (methods including associated parameter data) that ensure the protection of the rights of VDE participants.
The degree of trustedness of a VDE arrangement will be primarily based on whether hardware SPUs are employed at participant location secure subsystems and the effectiveness of the SPU hardware security architecture, software security techniques when an SPU is emulated in software, and the encryption algorithm(s) and keys that are employed for securing content, control information, communications, and access to VDE node (VDE installation) secure subsystems. Physical facility and user identity authentication security procedures may be used instead of hardware SPUs at certain nodes, such as at an established financial clearinghouse, where such procedures may provide sufficient security for trusted interoperability with a VDE arrangement employing hardware SPUs at user nodes.
The updating of property management files at each location of a VDE arrangement, to accommodate new or modified control information, is performed in the VDE secure subsystem and under the control of secure management file updating programs executed by the protected subsystem. Since all secure communications are at least in part encrypted and the processing inside the secure subsystem is concealed from outside observation and interference, the present invention ensures that content control information can be enforced. As a result, the creator and/or distributor and/or client administrator and/or other contributor of secure control information for each property (for example, an end-user restricting the kind of audit information he or she will allow to be reported and/or a financial clearinghouse establishing certain criteria for use of its credit for payment for use of distributed content) can be confident that their contributed and accepted control information will be enforced (within the security limitations of a given VDE security implementation design). This control information can determine, for example:
Seniority of contributed control information, including resolution of conflicts between content control information submitted by multiple parties, is normally established by:
An important feature of VDE is that it can be used to assure the administration of, and adequacy of security and rights protection for, electronic agreements implemented through the use of the present invention. Such agreements may involve one or more of:
VDE supports commercially secure “extended” value chain electronic agreements. VDE can be configured to support the various underlying agreements between parties that comprise this extended agreement. These agreements can define important electronic commerce considerations including:
VDE agreements may define the electronic commerce relationship of two or more parties of a value chain, but such agreements may, at times, not directly obligate or otherwise directly involve other VDE value chain participants. For example, an electronic agreement between a content creator and a distributor may establish both the price to the distributor for a creator's content (such as for a property distributed in a VDE container object) and the number of copies of this object that this distributor may distribute to end-users over a given period of time. In a second agreement, a value chain end-user may be involved in a three party agreement in which the end-user agrees to certain requirements for using the distributed product such as accepting distributor charges for content use and agreeing to observe the copyright rights of the creator. A third agreement might exist between the distributor and a financial clearinghouse that allows the distributor to employ the clearinghouse's credit for payment for the product if the end-user has a separate (fourth) agreement directly with the clearinghouse extending credit to the end-user. A fifth, evolving agreement may develop between all value chain participants as content control information passes along its chain of handling. This evolving agreement can establish the rights of all parties to content usage information, including, for example, the nature of information to be received by each party and the pathway of handling of content usage information and related procedures. A sixth agreement in this example, may involve all parties to the agreement and establishes certain general assumptions, such as security techniques and degree of trustedness (for example, commercial integrity of the system may require each VDE installation secure subsystem to electronically warrant that their VDE node meets certain interoperability requirements). In the above example, these six agreements could comprise agreements of an extended agreement for this commercial value chain instance.
VDE agreements support evolving (“living”) electronic agreement arrangements that can be modified by current and/or new participants through very simple to sophisticated “negotiations” between newly proposed content control information interacting with control information already in place and/or by negotiation between concurrently proposed content control information submitted by a plurality of parties. A given model may be asynchronously and progressively modified over time in accordance with existing senior rules and such modification may be applied to all, to classes of, and/or to specific content, and/or to classes and/or specific users and/or user nodes. A given piece of content may be subject to different control information at different times or places of handling, depending on the evolution of its content control information (and/or on differing, applicable VDE installation content control information). The evolution of control information can occur during the passing along of one or more VDE control information containing objects, that is control information may be modified at one or more points along a chain of control information handling, so long as such modification is allowed. As a result, VDE managed content may have different control information applied at both different “locations” in a chain of content handling and at similar locations in differing chains of the handling of such content. Such different application of control information may also result from content control information specifying that a certain party or group of parties shall be subject to content control information that differs from another party or group of parties. For example, content control information for a given piece of content may be stipulated as senior information and therefore not changeable, might be put in place by a content creator and might stipulate that national distributors of a given piece of their content may be permitted to make 100,000 copies per calendar quarter, so long as such copies are provided to boni fide end-users, but may pass only a single copy of such content to a local retailers and the control information limits such a retailer to making no more than 1,000 copies per month for retail sales to end-users. In addition, for example, an end-user of such content might be limited by the same content control information to making three copies of such content, one for each of three different computers he or she uses (one desktop computer at work, one for a desktop computer at home, and one for a portable computer).
Electronic agreements supported by the preferred embodiment of the present invention can vary from very simple to very elaborate. They can support widely diverse information management models that provide for electronic information security, usage administration, and communication and may support:
An important part of VDE provided by the present invention is the core secure transaction control arrangement, herein called an SPU (or SPUs), that typically must be present in each user's computer, other electronic appliance, or network. SPUs provide a trusted environment for generating decryption keys, encrypting and decrypting information, managing the secure communication of keys and other information between electronic appliances (i.e. between VDE installations and/or between plural VDE instances within a single VDE installation), securely accumulating and managing audit trail, reporting, and budget information in secure and/or non-secure non-volatile memory, maintaining a secure database of control information management instructions, and providing a secure environment for performing certain other control and administrative functions.
A hardware SPU (rather than a software emulation) within a VDE node is necessary if a highly trusted environment for performing certain VDE activities is required. Such a trusted environment may be created through the use of certain control software, one or more tamper resistant hardware modules such as a semiconductor or semiconductor chipset (including, for example, a tamper resistant hardware electronic appliance peripheral device), for use within, and/or operatively connected to, an electronic appliance. With the present invention, the trustedness of a hardware SPU can be enhanced by enclosing some or all of its hardware elements within tamper resistant packaging and/or by employing other tamper resisting techniques (e.g. microfusing and/or thin wire detection techniques). A trusted environment of the present invention implemented, in part, through the use of tamper resistant semiconductor design, contains control logic, such as a microprocessor, that securely executes VDE processes.
A VDE node's hardware SPU is a core component of a VDE secure subsystem and may employ some or all of an electronic appliance's primary control logic, such as a microcontroller, microcomputer or other CPU arrangement. This primary control logic may be otherwise employed for non VDE purposes such as the control of some or all of an electronic appliance's non-VDE functions. When operating in a hardware SPU mode, said primary control logic must be sufficiently secure so as to protect and conceal important VDE processes. For example, a hardware SPU may employ a host electronic appliance microcomputer operating in protected mode while performing VDE related activities, thus allowing portions of VDE processes to execute with a certain degree of security. This alternate embodiment is in contrast to the preferred embodiment wherein a trusted environment is created using a combination of one or more tamper resistant semiconductors that are not part of said primary control logic. In either embodiment, certain control information (software and parameter data) must be securely maintained within the SPU, and further control information can be stored externally and securely (e.g. in encrypted and tagged form) and loaded into said hardware SPU when needed. In many cases, and in particular with microcomputers, the preferred embodiment approach of employing special purpose secure hardware for executing said VDE processes, rather than using said primary control logic, may be more secure and efficient. The level of security and tamper resistance required for trusted SPU hardware processes depends on the commercial requirements of particular markets or market niches, and may vary widely.
These and other features and advantages provided by the present invention(s) may be better and more completely understood by referring to the following detailed description of presently preferred example embodiments in connection with the drawings, of which:
FIGS. 49, 49A-49F show an example OPEN method;
FIGS. 50, 50A-50F show an example of a READ method;
FIGS. 51, 51A-51F show an example of a WRITE method;
Almost any sort of transaction you can think of can be supported by virtual distribution environment 100. A few of many examples of transactions that can be supported by virtual distribution environment 100 include:
Virtual distribution environment 100 is “virtual” because it does not require many of the physical “things” that used to be necessary to protect rights, ensure reliable and predictable distribution, and ensure proper compensation to content creators and distributors. For example, in the past, information was distributed on records or disks that were difficult to copy. In the past, private or secret content was distributed in sealed envelopes or locked briefcases delivered by courier. To ensure appropriate compensation, consumers received goods and services only after they handed cash over to a seller. Although information utility 200 may deliver information by transferring physical “things” such as electronic storage media, the virtual distribution environment 100 facilitates a completely electronic “chain of handling and control.”
VDE Flexibility Supports Transactions
Information utility 200 flexibly supports many different kinds of information transactions. Different VDE participants may define and/or participate in different parts of a transaction. Information utility 200 may assist with delivering information about a transaction, or it may be one of the transaction participants.
For example, the video production studio 204 in the upper right-hand corner of
Even if a consumer has a copy of a video program, she cannot watch or copy the program unless she has “rules and controls” that authorize use of the program. She can use the program only as permitted by the “rules and controls.”
For example, video production studio 204 might release a half-hour exercise video in the hope that as many viewers as possible will view it. The video production studio 204 wishes to receive $2.00 per viewing. Video production studio 204 may, through information utility 200, make the exercise video available in “protected” form to all consumers 206, 208, 210. Video production studio 204 may also provide “rules and controls” for the video. These “rules and controls” may specify for example:
Information utility 200 allows even a small video production studio to market videos to consumers and receive compensation for its efforts. Moreover, the videos can, with appropriate payment to the video production studio, be made available to other video 10 publishers who may add value and/or act as repackagers or redistributors.
Use rights distributed by publishing house 214 may, for example, permit office 210 to make and distribute copies of the content to its employees. Office 210 may act as a redistributor by extending a “chain of handling and control” to its employees. The office 210 may add or modify “rules and controls” (consistent with the “rules and controls” it receives from publishing house 214) to provide office-internal control information and mechanisms. For example, office 210 may set a maximum usage budget for each individual user and/or group within the office, or it may permit only specified employees and/or groups to access certain information.
Example of What's Inside Information Utility 200
“Information utility” 200 in
Information utility 200 may include a “transaction processor” 200 b that processes transactions (to transfer electronic funds, for example) based on requests from participants and/or report receiver 200 e. It may also include a “usage analyst” 200 c that analyzes reported usage information. A “report creator” 200 d may create reports based on usage for example, and may provide these reports to outside participants and/or to participants within information utility 200. A “report receiver” 200 e may receive reports such as usage reports from content users. A “permissioning agent” 200 f may distribute “rules and controls” granting usage or distribution permissions based on a profile of a consumer's credit worthiness, for example. An administrator 200 h may provide information that keeps the virtual distribution environment 100 operating properly. A content and message storage 200 g may store information for use by participants within or outside of information utility 200.
Example of Distributing “Content” Using a “Chain of Handling and Control”
As explained above, virtual distribution environment 100 can be used to manage almost any sort of transaction. One type of important transaction that virtual distribution environment 100 may be used to manage is the distribution or communication of “content” or other important information.
Arrow 104 shows the content creator 102 sending the “rules and controls” associated with the content to a VDE rights distributor 106 (“distributor”) over an electronic highway 108 (or by some other path such as an optical disk sent by a delivery service such as U.S. mail). The content can be distributed over the same or different path used to send the “rules and controls.”
The distributor 106 generates her own “rules and controls” that relate to usage of the content. The usage-related “rules and controls” may, for example, specify what a user can and can't do with the content and how much it costs to use the content. These usage-related “rules and controls” must be consistent with the “rules and controls” specified by content creator 102.
Arrow 110 shows the distributor 106 distributing rights to use the content by sending the content's “rules and controls” to a content user 112 such as a consumer. The content user 112 uses the content in accordance with the usage-related “rules and controls.”
The distributor 106 and the content creator 102 may be the same person, or they may be different people. For example, a musical performing group may act as both content creator 102 and distributor 106 by creating and distributing its own musical recordings. As another example, a publishing house may act as a distributor 106 to distribute rights to use works created by an author content creator 102. Content creators 102 may use a distributor 106 to efficiently manage the financial end of content distribution.
The “financial clearinghouse” 116 shown in
More about “Rules and Controls”
The virtual distribution environment 100 prevents use of protected information except as permitted by the “rules and controls” (control information). For example, the “rules and controls” shown in
Every VDE participant in “chain of handling and control” is normally subject to “rules and controls.” “Rules and controls” define the respective rights and obligations of each of the various VDE participants. “Rules and controls” provide information and mechanisms that may establish interdependencies and relationships between the participants. “Rules and controls” are flexible, and permit “virtual distribution environment” 100 to support most “traditional” business transactions. For example:
“Rules and controls” may self limit if and how they may be changed. Often, “rules and controls” specified by one VDE participant cannot be changed by another VDE participant. For example, a content user 112 generally can't change “rules and controls” specified by a distributor 106 that require the user to pay for content usage at a certain rate. “Rules and controls” may “persist” as they pass through a “chain of handling and control,” and may be “inherited” as they are passed down from one VDE participant to the next.
Depending upon their needs, VDE participants can specify that their “rules and controls” can be changed under conditions specified by the same or other “rules and controls.” For example, “rules and controls” specified by the content creator 102 may permit the distributor 106 to “mark up” the usage price just as retail stores “mark up” the wholesale price of goods.
“Rules and controls” can be used to protect the content user's privacy by limiting the information that is reported to other VDE participants. As one example, “rules and controls” can cause content usage information to be reported anonymously without revealing content user identity, or it can reveal only certain information to certain participants (for example, information derived from usage) with appropriate permission, if required. This ability to securely control what information is revealed and what VDE participant(s) it is revealed to allows the privacy rights of all VDE participants to be protected.
“Rules and Contents” can be Separately Delivered
As mentioned above, virtual distribution environment 100 “associates” content with corresponding “rules and controls,” and prevents the content from being used or accessed unless a set of corresponding “rules and controls” is available. The distributor 106 doesn't need to deliver content to control the content's distribution. The preferred embodiment can securely protect content by protecting corresponding, usage enabling “rules and controls” against unauthorized distribution and use.
In some examples, “rules and controls” may travel with the content they apply to. Virtual distribution environment 100 also allows “rules and controls” to be delivered separately from content. Since no one can use or access protected content without “permission” from corresponding “rules and controls,” the distributor 106 can control use of content that has already been (or will in the future be), delivered. “Rules and controls” may be delivered over a path different from the one used for content delivery. “Rules and controls” may also be delivered at some other time. The content creator 102 might deliver content to content user 112 over the electronic highway 108, or could make the content available to anyone on the highway. Content may be used at the time it is delivered, or it may be stored for later use or reuse.
The virtual distribution environment 100 also allows payment and reporting means to be delivered separately. For example, the content user 112 may have a virtual “credit card” that extends credit (up to a certain limit) to pay for usage of any content. A “credit transaction” can take place at the user's site without requiring any “online” connection or further authorization. This invention can be used to help securely protect the virtual “credit card” against unauthorized use.
“Rules and Contents” Define Processes
The “events process” 402 detects things that happen (“events”) and determines which of those “events” need action by the other “processes.” The “events” may include, for example, a request to use content or generate a usage permission. Some events may need additional processing, and others may not. Whether an “event” needs more processing depends on the “rules and controls” corresponding to the content. For example, a user who lacks permission will not have her request satisfied (“No Go”). As another example, each user request to turn to a new page of an electronic book may be satisfied (“Go”), but it may not be necessary to meter, bill or budget those requests. A user who has purchased a copy of a novel may be permitted to open and read the novel as many times as she wants to without any further metering, billing or budgeting. In this simple example, the “event process” 402 may request metering, billing and/or budgeting processes the first time the user asks to open the protected novel (so the purchase price can be charged to the user), and treat all later requests to open the same novel as “insignificant events.” Other content (for example, searching an electronic telephone directory) may require the user to pay a fee for each access.
“Meter” process 404 keeps track of events, and may report usage to distributor 106 and/or other appropriate VDE participant(s).
(a) type of usage to charge for,
(b) what kind of unit to base charges on,
(c) how much to charge per unit,
(d) when to report, and
(e) how to pay.
These factors may be specified by the “rules and controls” that control the meter process.
Billing process 406 determines how much to charge for events. It records and reports payment information.
Budget process 408 limits how much content usage is permitted. For example, budget process 408 may limit the number of times content may be accessed or copied, or it may limit the number of pages or other amount of content that can be used based on, for example, the number of dollars available in a credit account. Budget process 408 records and reports financial and other transaction information associated with such limits.
Content may be supplied to the user once these processes have been successfully performed.
Containers and “Objects”
Container 302 may contain information content 304 in electronic (such as “digital”) form. Information content 304 could be the text of a novel, a picture, sound such as a musical performance or a reading, a movie or other video, computer software, or just about any other kind of electronic information you can think of. Other types of “objects” 300 (such as “administrative objects”) may contain “administrative” or other information instead of or in addition to information content 304.
(a) a “permissions record” 808,
(b) “budgets” 308; and
(c) “other methods” 1000.
“Budgets” 308 shown in
“Other methods” 1000 define basic operations used by “rules and controls.” Such “methods” 1000 may include, for example, how usage is to be “metered,” if and how content 304 and other information is to be scrambled and descrambled, and other processes associated with handling and controlling information content 304. For example, methods 1000 may record the identity of anyone who opens the electronic container 302, and can also control how information content is to be charged based on “metering.” Methods 1000 may apply to one or several different information contents 304 and associated containers 302, as well as to all or specific portions of information content 304.
Secure Processing Unit (SPU)
The “VDE participants” may each have an “electronic appliance.” The appliance may be or contain a computer. The appliances may communicate over the electronic highway 108.
SPU 500 is enclosed within and protected by a “tamper resistant security barrier” 502. Security barrier 502 separates the secure environment 503 from the rest of the world. It prevents information and processes within the secure environment 503 from being observed, interfered with and leaving except under appropriate secure conditions. Barrier 502 also controls external access to secure resources, processes and information within SPU 500. In one example, tamper resistant security barrier 502 is formed by security features such as “encryption,” and hardware that detects tampering and/or destroys sensitive information within secure environment 503 when tampering is detected.
SPU 500 in this example is an integrated circuit (“IC”) “chip” 504 including “hardware” 506 and “firmware” 508. SPU 500 connects to the rest of the electronic appliance through an “appliance link” 510. SPU “firmware” 508 in this example is “software” such as a “computer program(s)” “embedded” within chip 504. Firmware 508 makes the hardware 506 work. Hardware 506 preferably contains a processor to perform instructions specified by firmware 508. “Hardware” 506 also contains long-term and short-term memories to store information securely so it can't be tampered with. SPU 500 may also have a protected clock/calendar used for timing events. The SPU hardware 506 in this example may include special purpose electronic circuits that are specially designed to perform certain processes (such as “encryption” and “decryption”) rapidly and efficiently.
The particular context in which SPU 500 is being used will determine how much processing capabilities SPU 500 should have SPU hardware 506, in this example, provides at least enough processing capabilities to support the secure parts of processes shown in
VDE Electronic Appliance and “Rights Operating System”
a T.V. “set top” control box
a sound system
a video reproduction system
a video game player
a “smart” credit card.
Electronic appliance 600 in this example may include a keyboard or keypad 612, a voice recognizer 613, and a display 614. A human user can input commands through keyboard 612 and/or voice recognizer 613, and may view information on display 614. Appliance 600 may communicate with the outside world through any of the connections/devices normally used within an electronic appliance. The connections/devices shown along the bottom of the drawing are examples:
a “modem” 618 or other telecommunications link;
a CD ROM disk 620 or other storage medium or device;
a printer 622;
broadcast reception 624;
a document scanner 626; and
a “cable” 628 connecting the appliance with a “network.”
Virtual distribution environment 100 provides a “rights operating system” 602 that manages appliance 600 and SPU 500 by controlling their hardware resources. The operating system 602 may also support at least one “application” 608. Generally, “application” 608 is hardware and/or software specific to the context of appliance 600. For example, if appliance 600 is a personal computer, then “application” 608 could be a program loaded by the user, for instance, a word processor, a communications system or a sound recorder. If appliance 600 is a television controller box, then application 608 might be hardware or software that allows a user to order videos on demand and perform other functions such as fast forward and rewind. In this example, operating system 602 provides a standardized, well defined, generalized “interface” that could support and work with many different “applications” 608.
Operating system 602 in this example provides “rights and, auditing operating system functions” 604 and “other operating system functions” 606. The “rights and auditing operating system functions” 604 securely handle tasks that relate to virtual distribution environment 100. SPU 500 provides or supports many of the security functions of the “rights and auditing operating system functions” 402. The “other operating system functions” 606 handle general appliance functions. Overall operating system 602 may be designed from the beginning to include the “rights and auditing operating system functions” 604 plus the “other operating system functions” 606, or the “rights and auditing operating system functions” may be an add-on to a preexisting operating system providing the “other operating system functions.”
“Rights operating system” 602 in this example can work with many different types of appliances 600. For example, it can work with large mainframe computers, “minicomputers” and “microcomputers” such as personal computers and portable computing devices. It can also work in control boxes on the top of television sets, small portable “pagers,” desktop radios, stereo sound systems, telephones, telephone switches, or any other electronic appliance. This ability to work on big appliances as well as little appliances is called “scalable.” A “scalable” operating system 602 means that there can be a standardized interface across many different appliances performing a wide variety of tasks.
The “rights operating system functions” 604 are “services-based” in this example. For example, “rights operating system functions” 604 handle summary requests from application 608 rather than requiring the application to always make more detailed “subrequests” or otherwise get involved with the underlying complexities involved in satisfying a summary request. For example, application 608 may simply ask to read specified information; “rights operating system functions” 604 can then decide whether the desired information is VDE-protected content and, if it is, perform processes needed to make the information available. This feature is called “transparency.” “Transparency” makes tasks easy for the application 608. “Rights operating system functions” 604 can support applications 608 that “know” nothing about virtual distribution environment 100. Applications 608 that are “aware” of virtual distribution environment 100 may be able to make more detailed use of virtual distribution environment 100.
In this example, “rights operating system functions” 604 are “event driven.” Rather than repeatedly examining the state of electronic appliance 600 to determine whether a condition has arisen, the “rights operating system functions” 604 may respond directly to “events” or “happenings” within appliance 600.
In this example, some of the services performed by “rights operating system functions” 604 may be extended based on additional “components” delivered to operating system 602. “Rights operating system functions” 604 can collect together and use “components” sent by different participants at different times. The “components” help to make the operating system 602 “scalable.” Some components can change how services work on little appliances versus how they work on big appliances (e.g., multi-user). Other components are designed to work with specific applications or classes of applications (e.g., some types of meters and some types of budgets).
Electronic Appliance 600
An electronic appliance 600 provided by the preferred embodiment may, for example, be any electronic apparatus that contains one or more microprocessors and/or microcontrollers and/or other devices which perform logical and/or mathematical calculations. This may include computers; computer terminals; device controllers for use with computers; peripheral devices for use with computers; digital display devices; televisions; video and audio/video projection systems; channel selectors and/or decoders for use with broadcast and/or cable transmissions; remote control devices; video and/or audio recorders; media players including compact disc players, videodisc players and tape players; audio and/or video amplifiers; virtual reality machines; electronic game players; multimedia players; radios; telephones; videophones; facsimile machines; robots; numerically controlled machines including machine tools and the like; and other devices containing one or more microcomputers and/or microcontrollers and/or other CPUs, including those not yet in existence.
In the example shown, I/O controller 660 is connected to secondary storage device 652, a keyboard/display 612,614, a communications controller 666, and a backup storage device 668. Backup storage device 668 may, for example, store information on mass media such as a tape 670, a floppy disk, a removable memory card, etc. Communications controller 666 may allow electronic appliance 600 to communicate with other electronic appliances via network 672 or other telecommunications links. Different electronic appliances 600 may interoperate even if they use different CPUs and different instances of ROS 602, so long as they typically use compatible communication protocols and/or security methods. In this example, I/O controller 660 permits CPU 654 and SPU 500 to read from and write to secondary storage 662, keyboard/display 612, 614, communications controller 666, and backup storage device 668.
Secondary storage 662 may comprise the same one or more non-secure secondary storage devices (such as a magnetic disk and a CD-ROM drive as one example) that electronic appliance 600 uses for general secondary storage functions. In some implementations, part or all of secondary storage 652 may comprise a secondary storage device(s) that is physically enclosed within a secure enclosure. However, since it may not be practical or cost-effective to physically secure secondary storage 652 in many implementations, secondary storage 652 may be used to store information in a secure manner by encrypting information before storing it in secondary storage 652. If information is encrypted before it is stored, physical access to secondary storage 652 or its contents does not readily reveal or compromise the information.
Secondary storage 652 in this example stores code and data used by CPU 654 and/or SPU 500 to control the overall operation of electronic appliance 600. For example,
Secure Processing Unit 500
Each VDE node or other electronic appliance 600 in the preferred embodiment may include one or more SPUs 500. SPUs 500 may be used to perform all secure processing for VDE 100. For example, SPU 500 is used for decrypting (or otherwise unsecuring) VDE protected objects 300. It is also used for managing encrypted and/or otherwise secured communication (such as by employing authentication and/or error-correction validation of information). SPU 500 may also perform secure data management processes including governing usage of, auditing of, and where appropriate, payment for VDE objects 300 (through the use of prepayments, credits, real-time electronic debits from bank accounts and/or VDE node currency token deposit accounts). SPU 500 may perform other transactions related to such VDE objects 300.
SPU Physical Packaging and Security Barrier 502
As shown in
It is possible to remove the plastic package of an IC chip and gain access to the “die.” It is also possible to analyze and “reverse engineer” the “die” itself (e.g., using various types of logic analyzers and microprobes to collect and analyze signals on the die while the circuitry is operating, using acid etching or other techniques to remove semiconductor layers to expose other layers, viewing and photographing the die using an electron microscope, etc.) Although no system or circuit is absolutely impervious to such attacks, SPU barrier 502 may include additional hardware protections that make successful attacks exceedingly costly and time consuming. For example, ion implantation and/or other fabrication techniques may be used to make it very difficult to visually discern SPU die conductive pathways, and SPU internal circuitry may be fabricated in such a way that it “self-destructs” when exposed to air and/or light. SPU 500 may store secret information in internal memory that loses its contents when power is lost. Circuitry may be incorporated within SPU 500 that detects microprobing or other tampering, and self-destructs (or destroys other parts of the SPU) when tampering is detected. These and other hardware-based physical security techniques contribute to tamper-resistant hardware security barrier 502.
To increase the security of security barrier 502 even further, it is possible to encase or include SPU 500 in one or more further physical enclosures such as, for example: epoxy or other “potting compound”; further module enclosures including additional self-destruct, self-disabling or other features activated when tampering is detected, further modules providing additional security protections such as requiring password or other authentication to operate; and the like. In addition, further layers of metal may be added to the die to complicate acid etching, micro probing, and the like; circuitry designed to “zeroize” memory may be included as an aspect of self-destruct processes; the plastic package itself may be designed to resist chemical as well as physical “attacks”; and memories internal to SPU 500 may have specialized addressing and refresh circuitry that “shuffles” the location of bits to complicate efforts to electrically determine the value of memory locations. These and other techniques may contribute to the security of barrier 502.
In some electronic appliances 600, SPU 500 may be integrated together with the device microcontroller or equivalent or with a device I/O or communications microcontroller into a common chip (or chip set) 505. For example, in one preferred embodiment, SPU 500 may be integrated together with one or more other CPU(s) (e.g., a CPU 654 of an electronic appliance) in a single component or package. The other CPU(s) 654 may be any centrally controlling logic arrangement, such as for example, a microprocessor, other microcontroller, and/or array or other parallel processor. This integrated configuration may result in lower overall cost, smaller overall size, and potentially faster interaction between an SPU 500 and a CPU 654. Integration may also provide wider distribution if an integrated SPU/CPU component is a standard feature of a widely distributed microprocessor line. Merging an SPU 500 into a main CPU 654 of an electronic appliance 600 (or into another appliance or appliance peripheral microcomputer or other microcontroller) may substantially reduce the overhead cost of implementing VDE 100. Integration considerations may include cost of implementation, cost of manufacture, desired degree of security, and value of compactness.
SPU 500 may also be integrated with devices other than CPUs. For example, for video and multimedia applications, some performance and/or security advantages (depending on overall design) could result from integrating an SPU 500 into a video controller chip or chipset. SPU 500 can also be integrated directly into a network communications chip or chipset or the like. Certain performance advantages in high speed communications applications may also result from integrating an SPU 500 with a modem chip or chipset. This may facilitate incorporation of an SPU 500 into communication appliances such as stand-alone fax machines. SPU 500 may also be integrated into other, peripheral devices, such as CD-ROM devices, set-top cable devices, game devices, and a wide variety of other electronic appliances that use, allow access to, perform transactions related to, or consume, distributed information.
SPU 500 Internal Architecture
The following section discusses each of these SPU components in more detail.
Microprocessor 520 is the “brain” of SPU 500. In this example, it executes a sequence of steps specified by code stored (at least temporarily) within ROM 532 and/or RAM 534. Microprocessor 520 in the preferred embodiment comprises a dedicated central processing, arrangement (e.g., a RISC and/or CISC processor unit, a microcontroller, and/or other central processing means or, less desirably in most applications, process specific dedicated control logic) for executing instructions stored in the ROM 532 and/or other memory. Microprocessor 520 may be separate elements of a circuitry layout, or may be separate packages within a secure SPU 500.
In the preferred embodiment, microprocessor 520 normally handles the most security sensitive aspects of the operation of electronic appliance 600. For example, microprocessor 520 may manage VDE decrypting, encrypting, certain content and/or appliance usage control information, keeping track of usage of VDE secured content, and other VDE usage control related functions.
Stored in each SPU 500 and/or electronic appliance secondary memory 652 may be, for example, an instance of ROS 602 software, application programs 608, objects 300 containing VDE controlled property content and related information, and management database 610 that stores both information associated with objects and VDE control information. ROS 602 includes software intended for execution by SPU microprocessor 520 for, in part, controlling usage of VDE related objects 300 by electronic appliance 600. As will be explained, these SPU programs include “load modules” for performing basic control functions. These various programs and associated data are executed and manipulated primarily by microprocessor 520.
Real Time Clock (RTC) 528
In the preferred embodiment, SPU 500 includes a real time clock circuit (“RTC”) 528 that serves as a reliable, tamper resistant time base for the SPU. RTC 528 keeps track of time of day and date (e.g., month, day and year) in the preferred embodiment, and thus may comprise a combination calendar and clock. A reliable time base is important for implementing time based usage metering methods, “time aged decryption keys,” and other time based SPU functions.
The RTC 528 must receive power in order to operate. Optimally, the RTC 528 power source could comprise a small battery located within SPU 500 or other secure enclosure. However, the RTC 528 may employ a power source such as an externally located battery that is external to the SPU 500. Such an externally located battery may provide relatively uninterrupted power to RTC 528, and may also maintain as non-volatile at least a portion of the otherwise volatile RAM 534 within SPU 500.
In one implementation, electronic appliance power supply 659 is also used to power SPU 500. Using any external power supply as the only power source for RTC 528 may significantly reduce the usefulness of time based security techniques unless, at minimum, SPU 500 recognizes any interruption (or any material interruption) of the supply of external power, records such interruption, and responds as may be appropriate such as disabling the ability of the SPU 500 to perform certain or all VDE processes. Recognizing a power interruption may, for example, be accomplished by employing a circuit which is activated by power failure. The power failure sensing circuit may power another circuit that includes associated logic for recording one or more power fail events. Capacitor discharge circuitry may provide the necessary temporary power to operate this logic. In addition or alternatively, SPU 500 may from time to time compare an output of RTC 528 to a clock output of a host electronic appliance 600, if available. In the event a discrepancy is detected, SPU 500 may respond as appropriate, including recording the discrepancy and/or disabling at least some portion of processes performed by SPU 500 under at least some circumstances.
If a power failure and/or RTC 528 discrepancy and/or other event indicates the possibility of tampering, SPU 500 may automatically destroy, or render inaccessible without privileged intervention, one or more portions of sensitive information it stores, such as execution related information and/or encryption key related information. To provide further SPU operation, such destroyed information would have to be replaced by a VDE clearinghouse, administrator and/or distributor, as may be appropriate. This may be achieved by remotely downloading update and/or replacement data and/or code. In the event of a disabling and/or destruction of processes and/or information as described above, the electronic appliance 600 may require a secure VDE communication with an administrator, clearinghouse, and/or distributor as appropriate in order to reinitialize the RTC 528. Some or all secure SPU 500 processes may not operate until then.
It may be desirable to provide a mechanism for setting and/or synchronizing RTC 528. In the preferred embodiment, when communication occurs between VDE electronic appliance 600 and another VDE appliance, an output of RTC 528 may be compared to a controlled RTC 528 output time under control of the party authorized to be “senior” and controlling. In the event of a discrepancy, appropriate action may be taken, including resetting the RTC 528 of the “junior” controlled participant in the communication.
SPU Encrypt/Decrypt Engine 522
In the preferred embodiment, SPU encrypt/decrypt engine 522 provides special purpose hardware (e.g., a hardware state machine) for rapidly and efficiently encrypting and/or decrypting data. In some implementations, the encrypt/decrypt functions may be performed instead by microprocessor 520 under software control, but providing special purpose encrypt/decrypt hardware engine 522 will, in general, provide increased performance. Microprocessor 520 may, if desired, comprise a combination of processor circuitry and dedicated encryption/decryption logic that may be integrated together in the same circuitry layout so as to, for example, optimally share one or more circuit elements.
Generally, it is preferable that a computationally efficient but highly secure “bulk” encryption/decryption technique should be used to protect most of the data and objects handled by SPU 500. It is preferable that an extremely secure encryption/decryption technique be used as an aspect of authenticating the identity of electronic appliances 600 that are establishing a communication channel and securing any transferred permission, method, and administrative information. In the preferred embodiment, the encrypt/decrypt engine 522 includes both a symmetric key encryption/decryption circuit (e.g., DES, Skipjack/Clipper, IDEA, RC-2, RC-4, etc.) and an antisymmetric (asymmetric) or Public Key (“PK”) encryption/decryption circuit. The public/private key encryption/decryption circuit is used principally as an aspect of secure communications between an SPU 500 and VDE administrators, or other electronic appliances 600, that is between VDE secure subsystems. A symmetric encryption/decryption circuit may be used for “bulk” encrypting and decrypting most data stored in secondary storage 662 of electronic appliance 600 in which SPU 500 resides. The symmetric key encryption/decryption circuit may also be used for encrypting and decrypting content stored within VDE objects 300.
DES or public/private key methods may be used for all encryption functions. In alternate embodiments, encryption and decryption methods other than the DES and public/private key methods could be used for the various encryption related functions. For instance, other types of symmetric encryption/decryption techniques in which the same key is used for encryption and decryption could be used in place of DES encryption and decryption. The preferred embodiment can support a plurality of decryption/encryption techniques using multiple dedicated circuits within encrypt/decrypt engine 522 and/or the processing arrangement within SPU 500.
Pattern Matching Engine 524
Optional pattern matching engine 524 may provide special purpose hardware for performing pattern matching functions. One of the functions SPU 500 may perform is to validate/authenticate VDE objects 300 and other items. Validation/authentication often involves comparing long data strings to determine whether they compare in a predetermined way. In addition, certain forms of usage (such as logical and/or physical (contiguous) relatedness of accessed elements) may require searching potentially long strings of data for certain bit patterns or other significant pattern related metrics. Although pattern matching can be performed by SPU microprocessor 520 under software control, providing special purpose hardware pattern matching engine 524 may speed up the pattern matching process.
Compression/Decompression Engine 546
An optional compression/decompression engine 546 may be provided within an SPU 500 to, for example, compress and/or decompress content stored in, or released from, VDE objects 300. Compression/decompression engine 546 may implement one or more compression algorithms using hardware circuitry to improve the performance of compression/decompression operations that would otherwise be performed by software operating on microprocessor 520, or outside SPU 500. Decompression is important in the release of data such as video and audio that is usually compressed before distribution and whose decompression speed is important. In some cases, information that is useful for usage monitoring purposes (such as record separators or other delimiters) is “bidden” under a compression layer that must be removed before this information can be detected and used inside SPU 500.
Random Number Generator 542
Optional random number generator 542 may provide specialized hardware circuitry for generating random values (e.g., from inherently unpredictable physical processes such as quantum noise). Such random values are particularly useful for constructing encryption keys or unique identifiers, and for initializing the generation of pseudo-random sequences. Random number generator 542 may produce values of any convenient length, including as small as a single bit per use. A random number of arbitrary size may be constructed by concatenating values produced by random number generator 542. A cryptographically strong pseudo-random sequence may be generated from a random key and seed generated with random number generator 542 and repeated encryption either with the encrypt/decrypt engine 522 or cryptographic algorithms in SPU 500. Such sequences may be used, for example, in private headers to frustrate efforts to determine an encryption key through cryptoanalysis.
Arithmetic Accelerator 544
An optional arithmetic accelerator 544 may be provided within an SPU 500 in the form of hardware circuitry that can rapidly perform mathematical calculations such as multiplication and exponentiation involving large numbers. These calculations can, for example, be requested by microprocessor 520 or encrypt/decrypt engine 522, to assist in the computations required for certain asymmetric encryption/decryption operations. Such arithmetic accelerators are well-known to those skilled in the art. In some implementations, a separate arithmetic accelerator 544 may be omitted and any necessary calculations may be performed by microprocessor 520 under software control.
DMA Controller 526
DMA controller 526 controls information transfers over address/data bus 536 without requiring microprocessor 520 to process each individual data transfer. Typically, microprocessor 520 may write to DMA controller 526 target and destination addresses and the number of bytes to transfer, and DMA controller 526 may then automatically transfer a block of data between components of SPU 500 (e.g., from ROM 532 to RAM 534, between encrypt/decrypt engine 522 and RAM 534, between bus interface unit 530 and RAM 534, etc.). DMA controller 526 may have multiple channels to handle multiple transfers simultaneously. In some implementations, a separate DMA controller 526 may be omitted, and any necessary data movements may be performed by 20 microprocessor 520 under software control.
Bus Interface Unit (BIU) 530
Bus interface unit (BIU) 530 communicates information between SPU 500 and the outside world across the security barrier 502. BIU 530 shown in
Memory Management Unit 540
Memory Management Unit (MMU) 540, if present, provides hardware support for memory management and virtual memory management functions. It may also provide heightened security by enforcing hardware compartmentalization of the secure execution space (e.g., to prevent a less trusted task from modifying a more trusted task). More details are provided below in connection with a discussion of the architecture of a Secure Processing Environment (“SPE”) 503 supported by SPU 500.
MMU 540 may also provide hardware-level support functions related to memory management such as, for example, address mapping.
SPU Memory Architecture
In the preferred embodiment, SPU 500 uses three general kinds of memory:
The internal ROM 532 and RAM 534 within SPU 500 provide a secure operating environment and execution space. Because of cost limitations, chip fabrication size, complexity and other limitations, it may not be possible to provide sufficient memory within SPU 500 to store all information that an SPU needs to process in a secure manner. Due to the practical limits on the amount of ROM 532 and RAM 534 that may be included within SPU 500, SPU 500 may store information in memory external to it, and move this information into and out of its secure internal memory space on an as needed basis. In these cases, secure processing steps performed by an SPU typically must be segmented into small, securely packaged elements that may be “paged in” and “paged out” of the limited available internal memory space. Memory external to an SPU 500 may not be secure. Since the external memory may not be secure, SPU 500 may encrypt and cryptographically seal code and other information before storing it in external memory. Similarly, SPU 500 must typically decrypt code and other information obtained from external memory in encrypted form before processing (e.g., executing) based on it. In the preferred embodiment, there are two general approaches used to address potential memory limitations in a SPU 500. In the first case, the small, securely packaged elements represent information contained in secure database 610. In the second case, such elements may represent protected (e g, encrypted) virtual memory pages. Although virtual memory pages may correspond to information elements stored in secure database 610, this is not required in this example of a SPU memory architecture.
The following is a more detailed discussion of each of these three SPU memory resources.
SPU Internal ROM
SPU 500 read only memory (ROM) 532 or comparable purpose device provides secure internal non-volatile storage for certain programs and other information. For example, ROM 532 may store “kernel” programs such as SPU control firmware 508 and, if desired, encryption key information and certain fundamental “load modules.” The “kernel” programs, load module information, and encryption key information enable the control of certain basic functions of the SPU 500. Those components that are at least in part dependent on device configuration (e.g., POST, memory allocation, and a dispatcher) may be loaded in ROM 532 along with additional load modules that have been determined to be required for specific installations or applications.
In the preferred embodiment, ROM 532 may comprise a combination of a masked ROM 532 a and an EEPROM and/or equivalent “flash” memory 532 b. EEPROM or flash memory 532 b is used to store items that need to be updated and/or initialized, such as for example, certain encryption keys. An additional benefit of providing EEPROM and/or flash memory 532 b is the ability to optimize any load modules and library functions persistently stored within SPU 500 based on typical usage at a specific site. Although these items could also be stored in NVRAM 534 b, EEPROM and/or flash memory 532 b may be more cost effective.
Masked ROM 532 a may cost less than flash and/or EEPROM 532 b, and can be used to store permanent portions of SPU software/firmware. Such permanent portions may include, for example, code that interfaces to hardware elements such as the RTC 528, encryption/decryption engine 522, interrupt handlers, key generators, etc. Some of the operating system, library calls, libraries, and many of the core services provided by SPU 500 may also be in masked ROM 532 a. In addition, some of the more commonly used executables are also good candidates for inclusion in masked ROM 532 a. Items that need to be updated or that need to disappear when power is removed from SPU 500 should not be stored in masked ROM 532 a.
Under some circumstances, RAM 534 a and/or NVRAM 534 b (NVRAM 534 b may, for example, be constantly powered conventional RAM) may perform at least part of the role of ROM 532.
SPU Internal RAM
SPU 500 general purpose RAM 534 provides, among other things, secure execution space for secure processes. In the preferred embodiment, RAM 534 is comprised of different types of RAM such as a combination of high-speed RAM 534 a and an NVRAM (“non-volatile RAM”) 534 b. RAM 534 a may be volatile, while NVRAM 534 b is preferably battery backed or otherwise arranged so as to be non-volatile (i.e., it does not lose its contents when power is turned off).
High-speed RAM 534 a stores active code to be executed and associated data structures.
NVRAM 534 b preferably contains certain keys and summary values that are preloaded as part of an initialization process in which SPU 500 communicates with a VDE administrator, and may also store changeable or changing information associated with the operation of SPU 500. For security reasons, certain highly sensitive information (e.g., certain load modules and certain encryption key related information such as internally generated private keys) needs to be loaded into or generated internally by SPU 500 from time to time but, once loaded or generated internally, should never leave the SPU. In this preferred embodiment, the SPU 500 non-volatile random access memory (NVRAM) 534 b may be used for securely storing such highly sensitive information. NVRAM 534 b is also used by SPU 500 to store data that may change frequently but which preferably should not be lost in a power down or power fail mode.
NVRAM 534 b is preferably a flash memory array, but may in addition or alternatively be electrically erasable programmable read only memory (EEPROM), static RAM (SRAM), bubble memory, three dimensional holographic or other electro-optical memory, or the like, or any other writable (e.g., randomly accessible) non-volatile memory of sufficient speed and cost-effectiveness.
SPU External Memory
The SPU 500 can store certain information on memory devices external to the SPU. If available, electronic appliance 600 memory can also be used to support any device external portions of SPU 500 software. Certain advantages may be gained by allowing the SPU 500 to use external memory. As one example, memory internal to SPU 500 may be reduced in size by using non-volatile read/write memory in the host electronic appliance 600 such as a non-volatile portion of RAM 656 and/or ROM 658.
Such external memory may be used to store SPU programs, data and/or other information. For example, a VDE control program may be, at least in part, loaded into the memory and communicated to and decrypted within SPU 500 prior to execution. Such control programs may be re-encrypted and communicated back to external memory where they may be stored for later execution by SPU 500. “Kernel” programs and/or some or all of the non-kernel “load modules” may be stored by SPU 500 in memory external to it. Since a secure database 610 may be relatively large, SPU 500 can store some or all of secure database 610 in external memory and call portions into the SPU 500 as needed.
As mentioned above, memory external to SPU 500 may not be secure. Therefore, when security is required, SPU 500 must encrypt secure information before writing it to external memory, and decrypt secure information read from external memory before using it. Inasmuch as the encryption layer relies on secure processes and information (e.g., encryption algorithms and keys) present within SPU 500, the encryption layer effectively “extends” the SPU security barrier 502 to protect information the SPU 500 stores in memory external to it.
SPU 500 can use a wide variety of different types of external memory. For example, external memory may comprise electronic appliance secondary storage 652 such as a disk; external EEPROM or flash memory 658; and/or external RAM 656. External RAM 656 may comprise an external nonvolatile (e.g., constantly powered) RAM and/or cache RAM.
Using external RAM local to SPU 500 can significantly improve access times to information stored externally to an SPU. For example, external RAM may be used:
Dual ported external RAM can be particularly effective in improving SPU 500 performance, since it can decrease the data movement overhead of the SPU bus interface unit 530 and SPU microprocessor 520.
Using external flash memory local to SPU 500 can be used to significantly improve access times to virtually all data structures. Since most available flash storage devices have limited write lifetimes, flash storage needs to take into account the number of writes that will occur during the lifetime of the flash memory. Hence, flash storage of frequently written temporary items is not recommended. If external RAM is non-volatile, then transfer to flash (or hard disk) may not be necessary.
External memory used by SPU 500 may include two categories:
For some VDE implementations, sharing memory (e.g., electronic appliance RAM 656, ROM 658 and/or secondary storage 652) with CPU 654 or other elements of an electronic appliance 600 may be the most cost effective way to store VDE secure database management files 610 and information that needs to be stored external to SPU 500. A host system hard disk secondary memory 652 used for general purpose file storage can, for example, also be used to store VDE management files 610. SPU 500 may be given exclusive access to the external-memory (e.g., over a local bus high speed connection provided by BIU 530). Both dedicated and shared external memory may be provided.
The hardware configuration of an example of electronic appliance 600 has been described above. The following section describes an example of the software architecture of electronic appliance 600 provided by the preferred embodiment, including the structure and operation of preferred embodiment “Rights Operating System” (“ROS”) 602.
Rights Operating System 602
Rights Operating System (“ROS”) 602 in the preferred embodiment is a compact, secure, event-driven, services-based, “component” oriented, distributed multiprocessing operating system environment that integrates VDE information security control information, components and protocols with traditional operating system concepts. Like traditional operating systems, ROS 602 provided by the preferred embodiment is a piece of software that manages hardware resources of a computer system and extends management functions to input and/or output devices, including communications devices. Also like traditional operating systems, preferred embodiment ROS 602 provides a coherent set of basic functions and abstraction layers for hiding the differences between, and many of the detailed complexities of, particular hardware implementations. In addition to these characteristics found in many or most operating systems, ROS 602 provides secure VDE transaction management and other advantageous features not found in other operating systems. The following is a non-exhaustive list of some of the advantageous features provided by ROS 602 in the preferred embodiment:
Standardized Interface Provides Coherent Set of Basic Functions
Component Based Architecture
An “operating system” provides a control mechanism for organizing computer system resources that allows programmers to create applications for computer systems more easily. An operating system does this by providing commonly used functions, and by helping to ensure compatibility between different computer hardware and architectures (which may, for example, be manufactured by different vendors). Operating systems also enable computer “peripheral device” manufacturers to far more easily supply compatible equipment to computer manufacturers users.
Computer systems are usually made up of several different hardware components. These hardware components include, for example:
Most computer systems also include input/output devices such as keyboards, mice, video systems, printers, scanners and communications devices.
To organize the CPU's execution capabilities with available RAM, ROM and secondary storage devices, and to provide commonly used functions for use by programmers, a piece of software called an “operating system” is usually included with the other components. Typically, this piece of software is designed to begin executing after power is applied to the computer system and hardware diagnostics are completed. Thereafter, all use of the CPU, main memory and secondary memory devices is normally managed by this “operating system” software. Most computer operating systems, also typically include a mechanism for extending their management functions to I/O and other peripheral devices, including commonly used functions associated with these devices.
By managing the CPU, memory and peripheral devices through the operating system, a coherent set of basic functions and abstraction layers for hiding hardware details allows programmers to more easily create sophisticated applications. In addition, managing the computer's hardware resources with an operating system allows many differences in design and equipment requirements between different manufacturers to be hidden. Furthermore, applications can be more easily shared with other computer users who have the same operating system, with significantly less work to support different manufacturers' base hardware and peripheral devices.
ROS 602 is an Operating System Providing Significant Advantages
ROS 602 is an “operating system.” It manages the resources of electronic appliance 600, and provides a commonly used set of functions for programmers writing applications 608 for the electronic appliance. ROS 602 in the preferred embodiment manages the hardware (e.g., CPU(s), memory(ies), secure RTC(s), and encrypt/decrypt engines) within SPU 500. ROS may also manage the hardware (e.g., CPU(s) and memory(ies)) within one or more general purpose processors within electronic appliance 600. ROS 602 also manages other electronic appliance hardware resources, such as peripheral devices attached to an electronic appliance. For example, referring to
ROS 602 supports multiple Processors. ROS 602 in the preferred embodiment supports any number of local and/or remote processors. Supported processors may include at least two types: one or more electronic appliance processors 654, and/or one or more SPUs 500. A host processor CPU 654 may provide storage, database, and communications services SPU 500 may provide cryptographic and secured process execution services. Diverse control and execution structures supported by ROS 602 may require that processing of control information occur within a controllable execution space—this controllable execution space may be provided by SPU 500. Additional host and/or SPU processors may increase efficiencies and/or capabilities ROS 602 may access, coordinate and/or manage further processors remote to an electronic appliance 600 (e.g., via network or other communications link) to provide additional processor resources and/or capabilities.
ROS 602 is services based. The ROS services provided using a host processor 654 and/or a secure processor (SPU 500) are linked in the preferred embodiment using a “Remote Procedure Call” (“RPC”) internal processing request structure. Cooperating processors may request interprocess services using a RPC mechanism, which is minimally time dependent and can be distributed over cooperating processors on a network of hosts. The multi-processor architecture provided by ROS 602 is easily extensible to support any number of host or security processors. This extensibility supports high levels of scalability. Services also allow functions to be implemented differently on different equipment. For example, a small appliance that typically has low levels of usage by one user may implement a database service using very different techniques than a very large appliance with high levels of usage by many users. This is another aspect of scalability.
ROS 602 provides a distributed processing environment. For example, it permits information and control structures to automatically, securely pass between sites as required to fulfill a user's requests. Communications between VDE nodes under the distributed processing features of ROS 602 may include interprocess service requests as discussed above. ROS 602 supports conditional and/or state dependent execution of controlled processors within any VDE node. The location that the process executes and the control structures used may be locally resident, remotely accessible, or carried along by the process to support execution on a remote system.
ROS 602 provides distribution of control information, including for example the distribution of control structures required to permit “agents” to operate in remote environments. Thus, ROS 602 provides facilities for passing execution and/or information control as part of emerging requirements for “agent” processes.
If desired, ROS 602 may independently distribute control information over very low bandwidth connections that may or may not be “real time” connections. ROS 602 provided by the preferred embodiment is “network friendly,” and can be implemented with any level of networking protocol. Some examples include e-mail and direct connection at approximately “Layer 5” of the ISO model.
The ROS 602 distribution process (and the associated auditing of distributed information) is a controlled event that itself uses such control structures. This “reflective” distributed processing mechanism permits ROS 602 to securely distribute rights and permissions in a controlled manner, and effectively restrict the characteristics of use of information content. The controlled delegation of rights in a distributed environment and the secure processing techniques used by ROS 602 to support this approach provide significant advantages.
Certain control mechanisms within ROS 602 are “reciprocal.” Reciprocal control mechanisms place one or more control components at one or more locations that interact with one or more components at the same or other locations in a controlled way. For example, a usage control associated with object content at a user's location may have a reciprocal control at a distributor's location that governs distribution of the, usage control, auditing of the usage control, and logic to process user requests associated with the usage control. A usage control at a user's location (in addition to controlling one or more aspects of usage) may prepare audits for a distributor and format requests associated with the usage control for processing by a distributor. Processes at either end of a reciprocal control may be further controlled by other processes (e.g., a distributor may be limited by a budget for the number of usage control mechanisms they may produce). Reciprocal control mechanisms may extend over many sites and many levels (e.g., a creator to a distributor to a user) and may take any relationship into account (e.g., creator/distributor, distributor/user, user/user, user/creator, user/creator/distributor, etc.) Reciprocal control mechanisms have many uses in VDE 100 in representing relationships and agreements in a distributed environment.
ROS 602 is scalable. Many portions of ROS 602 control structures and kernel(s) are easily portable to various host platforms without recompilation. Any control structure may be distributed (or redistributed) if a granting authority permits this type of activity. The executable references within ROS 602 are portable within a target platform. Different instances of ROS 602 may execute the references using different resources. For example, one instance of ROS 602 may perform a task using an SPU 500, while another instance of ROS 602 might perform the same task using a host processing environment running in protected memory that is emulating an SPU in software. ROS 602 control information is similarly portable; in many cases the event processing structures may be passed between machines and host platforms as easily as between cooperative processors in a single computer. Appliances with different levels of usage and/or resources available for ROS 602 functions may implement those functions in very different ways. Some services may be omitted entirely if insufficient resources exist. As described elsewhere, ROS 602 “knows” what services are available, and how to proceed based on any given event. Not all events may be processable if resources are missing or inadequate.
ROS 602 is component based. Much of the functionality provided by ROS 602 in the preferred embodiment may be based on “components” that can be securely, independently deliverable, replaceable and capable of being modified (e.g., under appropriately secure conditions and authorizations). Moreover, the “components” may themselves be made of independently deliverable elements. ROS 602 may assemble these elements together (using a construct provided by the preferred embodiment called a “channel”) at execution time. For example, a “load module” for execution by SPU 500 may reference one or more “method cores,” method parameters and other associated data structures that ROS 602 may collect and assemble together to perform a task such as billing or metering. Different users may have different combinations of elements, and some of the elements may be customizable by users with appropriate authorization. This increases flexibility, allows elements to be reused, and has other advantages.
ROS 602 is highly secure. ROS 602 provides mechanisms to protect information control structures from exposure by end users and conduit hosts. ROS 602 can protect information, VDE control structures and control executables using strong encryption and validation mechanisms. These encryption and validation mechanisms are designed to make them highly resistant to undetected tampering. ROS 602 encrypts information stored on secondary storage device(s) 652 to inhibit tampering. ROS 602 also separately encrypts and validates its various components. ROS 602 correlates control and data structure components to prevent unauthorized use of elements. These features permit ROS 602 to independently distribute elements, and also allows integration of VDE functions 604 with non-secure “other” OS functions 606.
ROS 602 provided by the preferred embodiment extends conventional capabilities such as, for example, Access Control List (ACL) structures, to user and process defined events, including state transitions. ROS 602 may provide full control information over pre-defined and user-defined application events. These control mechanisms include “go/no-go” permissions, and also include optional event-specific executables that permit complete flexibility in the processing and/or controlling of events. This structure permits events to be individually controlled so that, for example, metering and budgeting may be provided using independent executables. For example, ROS 602 extends ACL structures to control arbitrary granularity of information. Traditional operating systems provide static “go-no go” control mechanisms at a file or resource level; ROS 602 extends the control concept in a general way from the largest to the smallest sub-element using a flexible control structure. ROS 602 can, for example, control the printing of a single paragraph out of a document file.
ROS 602 provided by the preferred embodiment permits secure modification and update of control information governing each component. The control information may be provided in a template format such as method options to an end-user. An end-user may then customize the actual control information used within guidelines provided by a distributor or content creator. Modification and update of existing control structures is preferably also a controllable event subject to auditing and control information.
ROS 602 provided by the preferred embodiment validates control structures and secured executables prior to use. This validation provides assurance that control structures and executables have not been tampered with by end-users. The validation also permits ROS 602 to securely implement components that include fragments of files and other operating system structures. ROS 602 provided by the preferred embodiment integrates security considerations at the operating system I/O level (which is below the access level), and provides “on-the-fly” decryption of information at release time. These features permit non-secure storage of ROS 602 secured components and information using an OS layer “on top of” traditional operating system platforms.
ROS 602 is highly integratable with host platforms as an additional operating system layer. Thus, ROS 602 may be created by “adding on” to existing operating systems. This involves hooking VDE “add ons” to the host operating system at the device driver and network interface levels. Alternatively, ROS 602 may comprise a wholly new operating system that integrates both VDE functions and other operating system functions.
Indeed, there are at least three general approaches to integrating VDE functions into a new operating system, potentially based on an existing operating system, to create a Rights Operating System 602 including:
(1) Redesign the operating system based on VDE transaction management requirements;
(2) Compile VDE API functions into an existing operating systems; and
(3) Integrate a VDE Interpreter into an existing operating system.
The first approach could be most effectively applied when a new operating system is being designed, or if a significant upgrade to an existing operating system is planned. The transaction management and security requirements provided by the VDE functions could be added to the design requirements list for the design of a new operating system that provides, in an optimally efficient manner, an integration of “traditional” operating system capabilities and VDE capabilities. For example, the engineers responsible for the design of the new version or instance of an operating system would include the requirements of VDE metering/transaction management in addition to other requirements (if any) that they use to form their design approach, specifications, and actual implementations. This approach could lead to a “seamless” integration of VDE functions and capabilities by threading metering/transaction management functionality throughout the system design and implementation.
The second approach would involve taking an existing set of API (Application Programmer Interface) functions, and incorporating references in the operating system code to VDE function calls. This is similar to the way that the current Windows operating system is integrated with DOS, wherein DOS serves as both the launch point and as a significant portion of the kernel underpinning of the Windows operating system. This approach would be also provide a high degree of “seamless” integration (although not quite as “seamless” as the first approach). The benefits of this approach include the possibility that the incorporation of metering/transaction management functionality into the new version or instance of an operating system may be accomplished with lower cost (by making use of the existing code embodied in an API, and also using the design’ implications of the API functional approach to influence the design of the elements into which the metering/transaction management functionality is incorporated).
The third approach is distinct from the first two in that it does not incorporate VDE functionality associated with metering/transaction management and data security directly into the operating system code, but instead adds a new generalized capability to the operating system for executing metering/transaction management functionality. In this case, an interpreter including metering/transaction management functions would be integrated with other operating system code in a “stand alone” mode. This interpreter might take scripts or other inputs to determine what metering/transaction management functions should be performed, and in what order and under which circumstances or conditions they should be performed.
Instead of (or in addition to) integrating VDE functions into/with an electronic appliance operating system, it would be possible to provide certain VDE functionality available as an application running on a conventional operating system.
ROS Software Architecture
HPE(s) 655 and SPE(s) 503 are self-contained computing and processing environments that may include their own operating system kernel 688 including code and data processing resources. A given electronic appliance 600 may include any number of SPE(s) 503 and/or any number of HPE(s) 655. HPE(s) 655 and SPE(s) 503 may process information in a secure way, and provide secure processing support for ROS 602. For example, they may each perform secure processing based on one or more VDE component assemblies 690, and they may each offer secure processing services to OS kernel 680.
In the preferred embodiment, SPE 503 is a secure processing environment provided at least in part by an SPU 500. Thus, SPU 500 provides the hardware tamper-resistant barrier 503 surrounding SPE 503. SPE 503 provided by the preferred embodiment is preferably:
In the preferred embodiment, HPE 655 is a secure processing environment supported by a processor other than an SPU, such as for example an electronic appliance CPU 654 general-purpose microprocessor or other processing system or device. In the preferred embodiment, HPE 655 may be considered to “emulate” an SPU 500 in the sense that it may use software to provide some or all of the processing resources provided in hardware and/or firmware by an SPU. HPE 655 in one preferred embodiment of the present invention is full-featured and fully compatible with SPE 503—that is, HPE 655 can handle each and every service call SPE 503 can handle such that the SPE and the HPE are “plug compatible” from an outside interface standpoint (with the exception that the HPE may not provide as much security as the SPE).
HPEs 655 may be provided in two types: secure and not secure. For example, it may be desirable to provide non-secure versions of HPE 655 to allow electronic appliance 600 to efficiently run non-sensitive VDE tasks using the full resources of a fast general purpose processor or computer. Such non-secure versions of HPE 655 may run under supervision of an instance of ROS 602 that also includes an SPE 503. In this way, ROS 602 may run all secure processes within SPE 503, and only use HPE 655 for processes that do not require security but that may require (or run more efficiently) under potentially greater resources provided by a general purpose computer or processor supporting HPE 655. Non-secure and secure HPE 655 may operate together with a secure SPE 503.
HPEs 655 may (as shown in
The software-based tamper resistant barrier 674 provided by HPE 655 may be provided, for example, by: introducing time checks and/or code modifications to complicate the process of stepping through code comprising a portion of kernel 688 a and/or a portion of component assemblies 690 using a debugger; using a map of defects on a storage device (e.g., a hard disk, memory card, etc.) to form internal test values to impede moving and/or copying HPE 655 to other electronic appliances 600; using kernel code that contains false branches and other complications in flow of control to disguise internal processes to some degree from disassembly or other efforts to discover details of processes; using “self-generating” code (based on the output of a co-sine transform, for example) such that detailed and/or complete instruction sequences are not stored explicitly on storage devices and/or in active memory but rather are generated as needed; using code that “shuffles” memory locations used for data values based on operational parameters to complicate efforts to manipulate such values; using any software and/or hardware memory management resources of electronic appliance 600 to “protect” the operation of HPE 655 from other processes, functions, etc. Although such a software-based tamper resistant barrier 674 may provide a fair degree of security, it typically will not be as secure as the hardware-based tamper resistant barrier 502 provided (at least in part) by SPU 500. Because security may be better/more effectively enforced with the assistance of hardware security features such as those provided by SPU 500 (and because of other factors such as increased performance provided by special purpose circuitry within SPU 500), at least one SPE 503 is preferred for many or most higher security applications. However, in applications where lesser security can be tolerated and/or the cost of an SPU 500 cannot be tolerated, the SPE 503 may be omitted and all secure processing may instead be performed by one or more secure HPEs 655 executing on general-purpose CPUs 654. Some VDE processes may not be allowed to proceed on reduced-security electronic appliances of this type if insufficient security is provided for the particular process involved.
Only those processes that execute completely within SPEs 503 (and in some cases, HPEs 655) may be considered to be truly secure. Memory and other resources external to SPE 503 and HPEs 655 used to store and/or process code and/or data to be used in secure processes should only receive and handle that information in encrypted form unless SPE 503/HPE 655 can protect secure process code and/or data from non-secure processes.
OS “core” 679 in the preferred embodiment includes a kernel 680, an RPC manager 732, and an “object switch” 734. API 682, HPE 655 and SPE 503 may communicate “event” messages with one another via OS “core” 679. They may also communicate messages directly with one another without messages going through OS “core” 679.
Kernel 680 may manage the hardware of an electronic appliance 600. For example, it may provide appropriate drivers and hardware managers for interacting with input/output and/or peripheral devices such as keyboard 612, display 614, other devices such as a “mouse” pointing device and speech recognizer 613, modem 618, printer 622, and an adapter for network 672. Kernel 680 may also be responsible for initially loading the remainder of ROS 602, and may manage the various ROS tasks (and associated underlying hardware resources) during execution. OS kernel 680 may also manage and access secure database 610 and file system 687. OS kernel 680 also provides execution services for applications 608 a(1), 608 a(2), etc. and other applications.
RPC manager 732 performs messaging routing and resource management/integration for ROS 680. It receives and routes “calls” from/to API 682, HPE 655 and SPE 503, for example.
Object switch 734 may manage construction, deconstruction and other manipulation of VDE objects 300.
User Notification/Exception Interface 686 in the preferred embodiment (which may be considered part of API 682 or another application coupled to the API) provides “pop up” windows/displays on display 614. This allows ROS 602 to communicate directly with a user without having to pass information to be communicated through applications 608. For applications that are not “VDE aware,” user notification/exception interface 686 may provide communications between ROS 602 and the user.
API 682 in the preferred embodiment provides a standardized, documented software interface to applications 608. In part, API 682 may translate operating system “calls” generated by applications 608 into Remote Procedure Calls (“RPCs”) specifying “events.” RPC manager 732 may route these RPCs to kernel 680 or elsewhere (e.g., to HPE(s) 655 and/or SPE(s) 503, or to remote electronic appliances 600, processors, or VDE participants) for processing. The API 682 may also service RPC requests by passing them to applications 608 that register to receive and process specific requests.
API 682 provides an “Applications Programming Interface” that is preferably standardized and documented. It provides a concise set of function calls an application program can use to access services provided by ROS 602. In at least one preferred example, API 682 will include two parts: an application program interface to VDE functions 604; and an application program interface to other OS functions 606. These parts may be interwoven into the same software, or they may be provided as two or more discrete pieces of software (for example).
Some applications, such as application 608 a(1) shown in
Other applications, such as application 608 b shown in
This “translation” feature of redirector 684 provides “transparency.” It allows VDE functions to be provided to the application 608(b) in a “transparent” way without requiring the application to become involved in the complexity and details associated with generating the one or more calls to VDE functions 604. This aspect of the “transparency” features of ROS 602 has at least two important advantages:
Since the second advantage (reducing complexity) makes it easier for an application creator to produce applications, even “VDE aware” applications 608 a(2) may be designed so that some calls invoking VDE functions 604 are requested at the level of an “other OS functions” call and then “translated” by redirector 684 into a VDE function call (in this sense, redirector 684 may be considered a part of API 682).
Referring again to
Secure ROS Components and Component Assemblies
As discussed above, ROS 602 in the preferred embodiment is a component-based architecture. ROS VDE functions 604 may be based on segmented, independently loadable executable “component assemblies” 690. These component assemblies 690 are independently securely deliverable. The component assemblies 690 provided by the preferred embodiment comprise code and data elements that are themselves independently deliverable. Thus, each component assembly 690 provided by the preferred embodiment is comprised of independently securely deliverable elements which may be communicated using VDE secure communication techniques, between VDE secure subsystems.
These component assemblies 690 are the basic functional unit provided by ROS 602. The component assemblies 690 are executed to perform operating system or application tasks. Thus, some component assemblies 690 may be considered to be part of the ROS operating system 602, while other component assemblies may be considered to be “applications” that run under the support of the operating system. As with any system incorporating “applications” and “operating systems, “the boundary between these aspects of an overall system can be ambiguous. For example, commonly used “application” functions (such as determining the structure and/or other attributes of a content container) may be incorporated into an operating system. Furthermore, “operating system” functions (such as task management, or memory allocation) may be modified and/or replaced by an application. A common thread in the preferred embodiment's ROS 602 is that component assemblies 690 provide functions needed for a user to fulfill her intended activities, some of which may be “application-like” and some of which may be “operating system-like.”
Components 690 are preferably designed to be easily separable and individually loadable. ROS 602 assembles these elements together into an executable component assembly 690 prior to loading and executing the component assembly (e.g., in a secure operating environment such as SPE 503 and/or HPE 655). ROS 602 provides an element identification and referencing mechanism that includes information necessary to automatically assemble elements into a component assembly 690 in a secure manner prior to, and/or during, execution.
ROS 602 application structures and control parameters used to form component assemblies 690 can be provided by different parties. Because the components forming component assemblies 690 are independently securely deliverable, they may be delivered at different times and/or by different parties (“delivery” may take place within a local VDE secure subsystem, that is submission through the use of such a secure subsystem of control information by a chain of content control information handling participant for the preparation of a modified control information set constitutes independent, secure delivery). For example, a content creator can produce a ROS 602 application that defines the circumstances required for licensing content contained within a VDE object 300. This application may reference structures provided by other parties. Such references might, for example, take the form of a control path that uses content creator structures to meter user activities; and structures created/owned by a financial provider to handle financial parts of a content distribution transaction (e.g., defining a credit budget that must be present in a control structure to establish creditworthiness, audit processes which must be performed by the licensee, etc.). As another example, a distributor may give one user more favorable pricing than another user by delivering different data elements defining pricing to different users. This attribute of supporting multiple party securely, independently deliverable control information is fundamental to enabling electronic commerce, that is, defining of a content and/or appliance control information set that represents the requirements of a collection of independent parties such as content creators, other content providers, financial service providers, and/or users.
In the preferred embodiment, ROS 602 assembles securely independently deliverable elements into a component assembly 690 based in part on context parameters (e.g., object, user). Thus, for example, ROS 602 may securely assemble different elements together to form different component assemblies 690 for different users performing the same task on the same VDE object 300. Similarly, ROS 602 may assemble differing element sets which may include, that is reuse, one or more of the same components to form different component assemblies 690 for the same user performing the same task on different VDE objects 300.
The component assembly organization provided by ROS 602 is recursive” in that a component assembly 690 may comprise one or more component “subassemblies” that are themselves independently loadable and executable component assemblies 690. These component “subassemblies” may, in turn, be made of one or more component “sub-sub-assemblies.” In the general case, a component assembly 690 may include N levels of component subassemblies.
Thus, for example, a component assembly 690(k) that may includes a component subassembly 690(k+1). Component subassembly 690(k+1), in turn, may include a component sub-sub-assembly 690(3), . . . and so on to N-level subassembly 690(k+N). The ability of ROS 602 to build component assemblies 690 out of other component assemblies provides great advantages in terms of, for example, code/data reusability, and the ability to allow different parties to manage different parts of an overall component.
Each component assembly 690 in the preferred embodiment is made of distinct components.
ROS 602 generates component assemblies 690 in a secure manner. As shown graphically in
In the preferred embodiment, ROS 602 assembles component assemblies 690 based on the following types of elements:
Permissions Records (“PERC”s) 808;
Method “Cores” 1000;
Load Modules 1100;
Data Elements (e.g., User Data Elements (“UDEs”) 1200 and Method Data Elements (“MDEs”) 1202); and
Other component assemblies 690.
Briefly, a PERC 808 provided by the preferred embodiment s a record corresponding to a VDE object 300 that identifies to ROS 602, among other things, the elements ROS is to assemble together to form a component assembly 690. Thus PERC 808 in effect contains a “list of assembly instructions” or a “plan” specifying what elements ROS 602 is to assemble together into a component assembly and how the elements are to be connected together. PERC 808 may itself contain data or other elements that are to become part of the component assembly 690.
The PERC 808 may reference one or more method “cores” 1000′. A method core 1000′ may define a basic “method” 1000 (e.g., “control,” “billing,” “metering,” etc.)
In the preferred embodiment, a “method” 1000 is a collection of basic instructions, and information related to basic instructions, that provides context, data, requirements, and/or relationships for use in performing, and/or preparing to perform, basic instructions in relation to the operation of one or more electronic appliances 600. Basic instructions may be comprised of, for example:
Information relating to said basic instructions may comprise, for example, data associated intrinsically With basic instructions such as for example, an identifier for the combined basic instructions and intrinsic data, addresses, constants, and/or the like. The information may also, for example, include one or more of the following:
Such information associated with a method may be stored, in part or whole, separately from basic instructions and intrinsic data. When these components are stored separately, a method may nevertheless include and encompass the other information and one or more sets of basic instructions and intrinsic data (the latter being included because of said other information's reference to one or more sets of basic instructions and intrinsic data), whether or not said one or more sets of basic instructions and intrinsic data are accessible at any given point in time.
Method core 1000′ may be parameterized by an “event code” to permit it to respond to different events in different ways. For example, a METER method may respond to a “use” event by storing usage information in a meter data structure. The same METER method may respond to an “administrative” event by reporting the meter data structure to a VDE clearinghouse or other VDE participant.
In the preferred embodiment, method core 1000′ may “contain,” either explicitly or by reference, one or more “load modules” 1100 and one or more data elements (UDEs 1200, MDEs 1202). In the preferred embodiment, a “load module” 1100 is a portion of a method that reflects basic instructions and intrinsic data. Load modules 1100 in the preferred embodiment contain executable code, and may also contain data elements (“DTDs” 1108) associated with the executable code. In the preferred embodiment, load modules 1100 supply the program instructions that are actually “executed” by hardware to perform the process defined by the method. Load modules 1100 may contain or reference other load modules.
Load modules 1100 in the preferred embodiment are modular and “code pure” so that individual load modules may be renterable and reusable. In order for components 690 to be dynamically updatable, they may be individually addressable within a global public name space. In view of these design goals, load modules 1100 are preferably small, code (and code-like) pure modules that are individually named and addressable. A single method may provide different load modules 1100 that perform the same or similar functions on different platforms, thereby making the method scalable and/or portable across a wide range of different electronic appliances.
UDEs 1200 and MDEs 1202 may store data for input to or output from executable component assembly 690 (or data describing such inputs and/or outputs). In the preferred embodiment, UDEs 1200 may be user dependent, whereas MDEs 202 may be user independent.
The component assembly example 690(k) shown in
One of the load modules 1100 b shown in this example is itself comprised of plural load modules 1100 c, 1100 d. Some of the load modules (e.g., 1100 a, 1100 d) in this example include one or more “DTD” data elements 1108 (e g, 1108 a, 1108 b) “DTD” data elements 1108 may be used, for example, to inform load module 1100 a of the data elements included in MDE 1202 and/or UDEs 1200 a, 1200 b. Furthermore, DTDs 1108 may be used as an aspect of forming a portion of an application used to inform a user as to the information required and/or manipulated by one or more load modules 1100, or other component elements. Such an application program may also include functions for creating and/or manipulating UDE(s) 1200, MDE(s) 1202, or other component elements, subassemblies, etc.
Components within component assemblies 690 may be “reused” to form different component assemblies. As mentioned above,
As mentioned above, ROS 602 provides several layers of security to ensure the security of component assemblies 690. One important security layer involves ensuring that certain component assemblies 690 are formed, loaded and executed only in secure execution space such as provided within an SPU 500. Components 690 and/or elements comprising them may be stored on external media encrypted using local SPU 500 generated and/or distributor provided keys.
ROS 602 also provides a tagging and sequencing scheme that may be used within the loadable component assemblies 690 to detect tampering by substitution. Each element comprising a component assembly 690 may be loaded into an SPU 500, decrypted using encrypt/decrypt engine 522, and then tested/compared to ensure that the proper element has been loaded. Several independent comparisons may be used to ensure here has been no unauthorized substitution. For example, the public and private copies of the element ID may be compared to ensure that they are the same, thereby preventing gross substitution of elements. In addition, a validation/correlation tag stored under the encrypted layer of the loadable element may be compared to make sure it matches one or more tags provided by a requesting process. This prevents unauthorized use of information. As a third protection, a device assigned tag (e.g., a sequence number) stored under an encryption layer of a loadable element may be checked to make sure it matches a corresponding tag value expected by SPU 500. This prevents substitution of older elements. Validation/correlation tags are typically passed only in secure wrappers to prevent plaintext exposure of this information outside of SPU 500.
The secure component based architecture of ROS 602 has important advantages. For example, it accommodates limited resource execution environments such as provided by a lower cost SPU 500. It also provides an extremely high level of configurability. In fact, ROS 602 will accommodate an almost unlimited diversity of content types, content provider objectives, transaction types and client requirements. In addition, the ability to dynamically assemble independently deliverable components at execution time based on particular objects and users provides a high degree of flexibility, and facilitates or enables a distributed database, processing, and execution environment.
One aspect of an advantage of the component-based architecture provided by ROS 602 relates to the ability to “stage” functionality and capabilities over time. As designed, implementation of ROS 602 is a finite task. Aspects of its wealth of functionality can remain unexploited until market realities dictate the implementation of corresponding VDE application functionality. As a result, initial product implementation investment and complexity may be limited. The process of “surfacing” the full range of capabilities provided by ROS 602 in terms of authoring, administrative, and artificial intelligence applications may take place over tune Moreover, already-designed functionality of ROS 602 may be changed or enhanced at any time to adapt to changing needs or requirements.
More Detailed Discussion of Rights Operating System 602 Architecture
As mentioned above, three basic components of ROS 602 are a kernel 680, a Remote Procedure Call (RPC) manager 732 and an object switch 734. These components, and the way they interact with other portions of ROS 602, will be discussed below.
Kernel 680 manages the basic hardware resources of electronic appliance 600, and controls the basic tasking provided by ROS 602. Kernel 680 in the preferred embodiment may include a memory manager 680 a, a task manager 680 b, and an 110 manager 680 c. Task manager 680 b may initiate and/or manage initiation of executable tasks and schedule them to be executed by a processor on which ROS 602 runs (e.g., CPU 654 shown in
RPC Manager 782
ROS 602 in a preferred embodiment is designed around a “services based” Remote Procedure Call architecture/interface. All functions performed by ROS 602 may use this common interface to request services and share information. For example, SPE(s) 503 provide processing for one or more RPC based services. In addition to supporting SPUs 500, the RPC interface permits the dynamic integration of external services and provides an array of configuration options using existing operating system components. ROS 602 also communicates with external services through the RPC interface to seamlessly provide distributed and/or remote processing. In smaller scale instances of ROS 602, a simpler message passing IPC protocol may be used to conserve resources. This may limit the configurability of ROS 602 services, but this possible limitation may be acceptable in some electronic appliances.
The RPC structure allows services to be called/requested without the calling process having to know or specify where the service is physically provided, what system or device will service the request, or how the service request will be fulfilled. This feature supports families of services that may be scaled and/or customized for specific applications. Service requests can be forwarded and serviced by different processors and/or different sites as easily as they can be forwarded and serviced by a local service system. Since the same RPC interface is used by ROS 602 in the preferred embodiment to request services within and outside of the operating system, a request for distributed and/or remote processing incurs substantially no additional operating system overhead. Remote processing is easily and simply integrated as part of the same service calls used by ROS 602 for requesting local-based services. In addition, the use of a standard RPC interface (“RSI”) allows ROS 602 to be modularized, with the different modules presenting a standardized interface to the remainder of the operating system. Such modularization and standardized interfacing permits different vendors/operating system programmers to create different portions of the operating system independently, and also allows the functionality of ROS 602 to be flexibly updated and/or changed based on different requirements and/or platforms.
RPC manager 732 manages the RPC interface. It receives service requests in the form of one or more “Remote Procedure Calls” (RPCs) from a service requestor, and routes the service requests to a service provider(s) that can service the request. For example, when rights operating system 602 receives a request from a user application via user API 682, RPC manager 732 may route the service request to an appropriate service through the PC service interface” (“RSI”). The RSI is an interface between RPC manager 732, service requesters, and a resource that will accept and service requests.
The RPC interface (RSI) is used for several major. ROS 602 subsystems in the preferred embodiment.
RPC services provided by ROS 602 in the preferred embodiment are divided into subservices, i.e., individual instances a specific service each of which may be tracked individually by RPC manager 732. This mechanism permits multiple instances of a specific service on higher throughput systems while maintaining a common interface across a spectrum of implementations. The subservice concept extends to supporting multiple processors, multiple SPEs 503, multiple HPEs 655, and multiple communications services.
The preferred embodiment ROS 602 provides the following RPC based service providers/requestors (each of which have an RPC interface or “RSI” that communicates with RPC manager 732):
The types of services provided by HFE 655, SPE 503, User Notification 686, API 742 and Redirector 684 have already been described above. Here is a brief description of the type(s) of services provided by OS resources 744, 752, 754, 756 and 776:
Object switch 734 handles, controls and communicates (both locally and remotely) VDE objects 300. In the preferred embodiment, the object switch may include the following elements:
Stream router 758 routes to/from “real time” and “time independent” data streams handled respectively by real time stream interface(s) 760 and tune dependent stream interface(s) 762. Intercept 692 intercepts 110 requests that involve real-time information streams such as, for example, real time feed 694. The routing performed by stream router 758 may be determined by routing tables 766. Buffering/storage 768 provides temporary store-and-forward, buffering and related services. Container manager 764 may (typically in conjunction with VSPE 503) perform processes on VDE objects 300 such as constructing, deconstructing, and locating portions of objects.
Object switch 734 communicates through an Object Switch Interface (“OSI”) with other parts of ROS 602. The Object Switch Interface may resemble, for example, the interface for a Unix socket in the preferred embodiment. Each of the “OSI” interfaces shown in
ROS 602 includes the following object switch service providers/resources (each of which can communicate with the object switch 734 through an “OSI”):
In the preferred embodiment, communications manager 776 may include a network manager 780 and a mail gateway (manager) 782. Mail gateway 782 may include one or more mail filters 784 to, for example, automatically route VDE related electronic mail between object switch 734 and the outside world electronic mail services. External Services Manager 772 may interface to communications manager 776 through a Service Transport Layer 786. Service Transport Layer 786 a may enable External Services Manager 772 to communicate with external computers and systems using various protocols managed using the service transport layer 786.
The characteristics of and interfaces to the various subsystems of ROS 680 shown in
RPC Manager 732 and its RPC Services Interface
As discussed above, the basic system services provided by ROS 602 are invoked by using an RPC service interface (RSI). This RPC service interface provides a generic, standardized interface for different services systems and subsystems provided by ROS 602.
RPC Manager 732 routes RPCs requesting services to an appropriate RPC service interface. In the preferred embodiment, upon receiving an RPC call, RPC manager 732 determines one or more service managers that are to service the request. RPC manager 732 then routes a service request to the appropriate service(s) (via a RSI associated with a service) for action by the appropriate service manager(s).
For example, if a SPE 503 is to service a request, the RPC Manager 732 routes the request to RSI 736 a, which passes the request on to SPE device driver 736 for forwarding to the SPE. Similarly, if HPE 655 is to service the request, RPC Manager 732 routes the request to RSI 738 a for forwarding to a HPE. In one preferred embodiment, SPE 503 and HPE 655 may perform essentially the same services so that RSIs 736 a, 738 a are different instances of the same RSI. Once a service request has been received by SPE 503 (or HPE 655), the SPE (or HPE) typically dispatches the request internally using its own internal RPC manager (as will be discussed shortly). Processes within SPEs 503 and HPEs 655 can also generate RPC requests. These requests may be processed internally by a SPE/HPE, or if not internally serviceable, passed out of the SPE/HPE for dispatch by RPC Manager 732.
Remote (and local) procedure calls may be dispatched by a RPC Manager 732 using an “RPC Services Table.” An RPC Services Table describes where requests for specific services are to be routed for processing Each row of an RPC Services Table in the preferred embodiment contains a services ID, the location of the service, and an address to which control will be passed to service a request. An RPC Services Table may also include control information, that indicates which instance of the RPC dispatcher controls the service. Both RPC Manager 732 and any attached SPEs 503 and HPEs 655 may have symmetric copies of the RPC Services Table. If an RPC service is not found in the RPC services tables, it is either rejected or passed to external services manager 772 for remote servicing.
Assuming RPC manager 732 finds a row corresponding to the request in an RPC Services Table, it may dispatch the request to an appropriate RSI. The receiving RSI accepts a request from the RPC manager 732 (which may have looked up the request in an RPC service table), and processes that request in accordance with internal priorities associated with the specific service.
In the preferred embodiment, RPC Service Interface(s) supported by RPC Manager 732 may be standardized and published to support add-on service modules developed by third party vendors, and to facilitate scalability by making it easier to program ROS 602. The preferred embodiment RSI closely follows the DOS and Unix device driver models for block devices so that common code may be developed for many platforms with minimum effort. An example of one possible set of common entry points are listed below in the table.
In the preferred embodiment, services (and the associated RSIs they present to RPC manager 732) may be activated during boot by an installation boot process that issues an RPC LOAD. This process reads an RPC Services Table from a configuration file, loads the service module if it is run time loadable (as opposed to being a kernel linked device driver), and then calls the LOAD entry point for the service. A successful return from the LOAD entry point will indicate that the service has properly loaded and is ready to accept requests.
RPC LOAD Call Example: SVC_LOAD (Long Service_Id)
This LOAD interface call is called by the RPC manager 732 during rights operating system 602 initialization. It permits a service manager to load any dynamically loadable components and to’ initialize any device and memory required by the service. The service number that the service is loaded as is passed in as service_id parameter. In the preferred embodiment, the service returns 0 is the initialization process was completed successfully or an error number if some error occurred.
Once a service has been loaded, it may not be fully functional for all subservices. Some subservices (e.g., communications based services) may require the establishment of additional connections, or they may require additional modules to be loaded. If the service is defined as “mountable,” a RPC manager 732 will call the MOUNT subservice entry point with the requested subservice ID prior to opening an instance of a subservice.
RPC MOUNT Call Example:
SVC_MOUNT (long service_id, long subservice_id, BYTE *buffer)
This MOUNT interface call instructs a service to make a specific subservice ready. This may include services related to networking, communications, other system services, or external resources. The service_id and subservice_id parameters may be specific to the specific service being requested. The buffer parameter is a memory address that references a control structure appropriate to a specific service.
Once a service is loaded and “mounted,” specific instances of a service may be “opened” for use. “Opening” an instance of a service may allocate memory to store control and status information. For example, in a BSD socket based network connection, a LOAD call will initialize the software and protocol control tables, a MOUNT call will specify networks and hardware resources, and an OPEN will actually open a socket to a remote installation.
Some services, such as commercial database manager 730 that underlies the secure database service, may not be “mountable.” In this case, a LOAD call will make a connection to a database manager 730 and ensure that records are readable; An OPEN call may create instances of internal cache manager 746 for various classes of records.
RPC OPEN Call Example:
This OPEN interface call instructs a service to open a specific subservice. The service_id and subservice_id parameters are specific to the specific service being requested, and the buffer parameter is a memory address that references a control structure appropriate to a specific service.
The optional receive parameter is the address of a notification callback function that is called by a service whenever a message is ready for the service to retrieve it. One call to this address is made for each incoming message received. If the caller passes a NULL to the interface, the software will not generate a callback for each message.
Close, Unmount and Unload
The converse of the OPEN, MOUNT, and LOAD calls are CLOSE, UNMOUNT, and UNLOAD. These interface calls release any allocated resources back to ROS 602 (e.g., memory manager 680 a).
RPC CLOSE Call Example: SVC_CLOSE (Long Svc_Handle)
This LOAD interface call closes an open service “handle.” A service “handle” describes a service and subservice that a user wants to close. The call returns 0 if the CLOSE request succeeds (and the handle is no longer valid) or an error number.
RPC UNLOAD Call Example: SVC_UNLOAD (Void)
This UNLOAD interface call is called by a RPC manager 732 during shutdown or resource reallocation of rights operating system 602. It permits a service to close any open connections, flush buffers, and to release any operating system resources that it may have allocated. The service returns 0.
RPC UNMOUNT Call Example: SVC_UNMOUNT (Long Service_Id, Long Subservice_Id)
This UNMOUNT interface call instructs a service to deactivate a specific subservice. The service_id and subservice_id parameters are specific to the specific áervice being requested, and must have been previously mounted using the SVC_MOUNT( ) request. The call releases all system resources associated with the subservice before it returns.
Read and Write
The READ and WRITE calls provide a basic mechanism for sending information to and receiving responses from a mounted and opened service. For example, a service has requests written to it in the form of an RPC request, and makes its response available to be read by RPC Manager 732 as they become available.
RPC READ Call Example:
SVC_READ (long svc_handle, long request_id, BYTE *buffer, long size)
This READ call reads a message response from a service. The svc_handle and request_id parameters uniquely identify a request. The results of a request will be stored in the user specified buffer up to size bytes. If the buffer is too small, the first size bytes of the message will be stored in the buffer and an error will be returned.
If a message response was returned to the caller's buffer correctly, the function will return 0. Otherwise, an error message will be returned.
RPC WRITE Call Example:
SVC_write (long service_id, long subservice_id, BYTE *buffer, long size, int (*receive) (long requestid)
This WRITE call writes a message to a service and subservice specified by the service_id/subservice_id parameter pair. The message is stored in buffer (and usually conforms to the VDE RPC message format) and is size bytes long. The function returns the request id for the message (if it was accepted, for sending) or an error number. If a user specifies the receive callback functions, all messages regarding a request will be sent to the request specific callback; routine instead of the generalized message callback.
The IOCTL (“Input/Output ConTroL”) call provides a mechanism for querying the status of and controlling a loaded service. Each service type will respond to specific general IOCTL requests, all required class IOCTL requests, and service specific IOCTL requests.
RPC IOCTL Call Example: ROLSVC_IOCTL (Lông Service_Id, Long Subservice_Id,
int command, BYTE *buffer)
This IOCTL function provides a generalized control interface for a RSI. A user specifies the service_id parameter and an optional subservice_id parameter that they wish to control They specify the control command parameter(s), and a buffer into/from which the command parameters may be written/read. An example of a list of commands and the appropriate buffer structures are given below.
Now that a generic RPC Service Interface provided by the preferred embodiment has been described, the following description relates to particular examples of services provided by ROS 602.
SPE Device Driver 736
SPE device driver 736 provides an interface between ROS 602 and SPE 503. Since SPE 503 in the preferred embodiment runs within the confines of an SPU 500, one aspect of this device driver 736 is to provide low level communications services with the SPU 500 hardware. Another aspect of SPE device driver 736 is to provide an RPC service interface (RSI) 736 a particular to SPE 503 (this same RSI may be used to communicate with HPE 655 through HPE device driver 738).
SPE RSI 736 a and driver 736 isolates calling processes within ROS 602 (or external to the ROS) from the detailed service provided by the SPE 503 by providing a set of basic interface points providing a concise function set. This has several advantages. For example, it permits a full line of scaled SPUs 500 that all provide common functionality to the outside world but which may differ in detailed internal structure and architecture. SPU 500 characteristics such as the amount of memory resident in the device, processor speed, and the number of services supported within SPU 500 may be the decision of the specific SPU manufacturer, and in any event may differ from one SPU configuration to another. To maintain compatibility, SPE device driver 736 and the RSI 736 a it provides conform to a basic common RPC interface standard that “hides” differences between detailed configurations of SPUs 500 and/or the SPEs 503 they may support.
To provide for such compatibility, SPE RSI 736 a in the preferred embodiment follows a simple block based standard. In the preferred embodiment, an SPE RSI 736 a may be modeled after the packet interfaces for network Ethernet cards. This standard closely models the block mode interface characteristics of SPUs 500 in the preferred embodiment.
An SPE RSI 736 a allows RPC calls from RPC manager 732 to access specific services provided by an SPE 736. To do this, SPE RSI 736 a provides a set of “service notification address interfaces.” These provide interfaces to individual services provided by SPE 503 to the outside world. Any calling process within ROS 602 may access these SPE-provided services by directing an RPC call to SPE RSI 736 a and specifying a corresponding “service notification address” in an RPC call. The specified service notification “address” causes SPE 503 to internally route an RPC call to a particular service within an SPE. The following is a listing of one example of a SPE service breakdown for which individual service notification addresses may be provided:
Channel Services Manager
Authentication Manager/Secure Communications Manager
Secure Database Manager
The Channel Services Manager is the principal service provider and access point to SPE 503 for the rest of ROS 602. Event processing, as will be discussed later, is primarily managed (from the point of view of processes outside SPE 503) by this service. The Authentication Manager/Secure Communications Manager may provide login/logout services for users of ROS 602, and provide a direct service for managing communications (typically encrypted or otherwise protected) related to component assemblies 690, VDE objects 300, etc. Requests for display of information (e.g., value remaining in a financial budget) may be provided by a direct service request to a Secure Database Manager inside SPE 503 The instances of Authentication Manager/Secure Communications Manager and Secure Database Manager, if available at all, may provide only a subset of the information and/or capabilities available to processes operating inside SPE 503. As stated above, most (potentially all) service requests entering SPE are routed to a Channel Services Manager for processing. As will be discussed in more detail later on, most control structures and event processing logic is associated with component assemblies 690 under the management of a Channel Services Manager.
The SPE 503 must be accessed through its associated SPE driver 736 in this example. Generally, calls to SPE driver 736 are made in response to RPC calls. In this example, SPE driver RSI 736 a may translate RPC calls directed to control or ascertain information about SPE driver 736 into driver calls SPE driver RSI 736 a in conjunction with driver 736 may pass RPC calls directed to SPE 503 through to the SPE.
The following table shows one example of SPE device driver 736 calls:
The following are more detailed examples of each of the SPE driver calls set forth in the table above.
Example of an “SPE Information” Driver Call: SPE_Info (Void)
This function returns a pointer to an SPE_INFO data structure that defines the SPE device driver 736 a. This data structure may provide certain information about SPE device driver 736, RSI 736 a and/or SPU 500. An example of a SPE_INFO structure is described below:
SPE_initialize_interface (int (fcn*receiver) (void))
A receiver function passed in by way of a parameter will be called for all packets received from SPE 503 unless their destination service is over-ridden using the set_notify( ) call. A receiver function allows ROS 602 to specify a format for packet communication between RPC manager 732 and SPE 503.
This function returns “0” in the preferred embodiment if the initialization of the interface succeeds and non-zero if it fails. If the function fails, it will return a code that describes the reason for the failure as the value of the function.
Example of an SPE “Terminate Interface” Driver Call:
In the preferred embodiment, this function shuts down SPE Driver 736, clears all notification addresses, and terminates all outstanding requests between an SPE and an ROS RPC manager, 732. It also resets an SPE 503 (e.g., by a warm reboot of SPU 500) after all requests are resolved.
Termination of driver 736 should be performed by ROS 602 when the operating system is starting to shut down. It may also be necessary to issue this call if an SPE 503 and ROS 602 get so far out of synchronization that all processing in an SPE must be reset to a known state.
Example of an SPE “Reset Interface” Driver Call:
This function resets driver 736, terminates all outstanding requests between SPE 503 and an ROS RPC manager 732, and clears all statistics counts. It does not reset the SPU 500, but simply restores driver 736 to a known stable state.
Example of an SPE “Get Statistics” Driver Call: SPE_Get_Stats (Long Service_Id)
This function returns statistics for a specific service notification interface or for the SPE driver 736 in general. It returns a pointer to a static buffer that contains these statistics or NULL if statistics are unavailable (either because an interface is not initialized or because a receiver address was not specified). An example of the SPE_STATS structure may have the following definition:
If a user specifies a service ID, statistics associated with packets sent by that service are returned. If a user specified 0 as the parameter, the total packet statistics for the interface are returned.
Example of an SPE “Clear Statistic” Driver Call: SPE_Clear_Stats (Long Service_d)
This function clears statistics associated with the SPE service_id specified. If no service_id is specified (i.e., the caller passes in 0), global statistics will be cleared. The function returns 0 if statistics are successfully cleared or an error number if an error occurs.
Example of an SPE “Set Notification Address” Driver Call:
SPE_set_notify (long service_id, int (fcn*receiver) (void))
This function sets a notification address (receiver) for a specified service. If the notification address is set to NULL, SPE device driver 736 will send notifications for packets to the specified service to the default notification address.
Example of a SPE “Get Notification Address” Driver Call:
SPE_get_notify (long service_id)
This function returns a notification address associated with the named service or NULL if no specific notification address has been specified.
Example of an “SPE Send Packet” Driver Call:
send_pkt (BYTE*buffer, long size, int (far*receive) (void))
This function sends a packet stored in buffer of “length” size. It returns 0 if the packet is sent successfully, or returns an error code associated with the failure.
Redirector Service Manager 684
The redirector 684 is a piece of systems integration software used principally when ROS 602 is provided by “adding on” to a pre-existing operating system or when “transparent” operation is desired for some VDE functions, as described earlier. In one embodiment the kernel 680, part of communications manager 776, file system 687, and part of API service 742 may be part of a pre-existing operating system such as DOS, Windows, UNIX, Macintosh System, OS9, PSOS, OS/2, or other operating system platform. The remainder of ROS 602 subsystems shown in
In a scenario of this type of integration, ROS 602 will continue to be supported by a preexisting OS kernel 680, but may supplement (or even substitute) many of its functions by providing additional add-on pieces such as, for example, a virtual memory manager.
Also in this integration scenario, an add-on portion of API service 742 that integrates readily with a preexisting API service is provided to support VDE function calls. A pre-existing API service integrated with an add-on portion supports an enhanced set of operating system calls including both calls to VDE functions 604 and calls to functions 606 other than VDE functions (see
ROS 602 may use a standard communications manager 776 provided by the preexisting operating system, or it may provide “add ons” and/or substitutions to it that may be readily integrated into it. Redirector 684 may provide this integration function.
This leaves a requirement for ROS 602 to integrate with a preexisting file system 687. Redirector 684 provides this integration function.
In this integration scenario, file system 687 of the preexisting operating system is used for all accesses to secondary storage. However, VDE objects 300 may be stored on secondary storage in the form of external object repository 728, file system 687, or remotely accessible through communications manager 776. When object switch 734 wants to access external object repository 728, it makes a request to the object repository manager 770 that then routes the request to object repository 728 or to redirector 692 (which in turn accesses the object in file system 687).
Generally, redirector 684 maps VDE object repository 728 content into preexisting calls to file system 687. The redirector 684 provides preexisting OS level information about a VDE object 300, including mapping the object into a preexisting OS's name space. This permits seamless access to VDE protected content using “normal” file system 687 access techniques provided by a preexisting operating system.
In the integration scenarios discussed above, each preexisting target OS file system 687 has different interface requirements by which the redirector mechanism 684 may be “hooked.” In general, since all commercially viable operating systems today provide support for network based volumes, file systems, and other devices (e.g., printers, modems, etc.), the redirector 684 may use low level network and file access “hooks” to integrate with a preexisting operating system. “Add-ons” for supporting VDE functions 602 may use these existing hooks to integrate with a preexisting operating system.
User Notification Service Manager 740
User Notification Service Manager 740 and associated user notification exception interface (“pop up”) 686 provides ROS 602 with an enhanced ability to communicate with a user of electronic appliance 600. Not all applications 608 may be designed to respond to messaging from ROS 602 passed through API 682, and it may in any event be important or desirable to give ROS 602 the ability to communicate with a user no matter what state an application is in. User notification services manager 740 and interface 686 provides ROS 602 with a mechanism to communicate directly with a user, instead of or in addition to passing a return call through API 682 and an application 608. This is similar, for example, to the ability of the Windows operating system to display a user message in a “dialog box” that displays “on top of” a running application irrespective of the state of the application.
The User Notification 686 block in the preferred embodiment may be implemented as application code. The implementation of interface 740 a is preferably built over notification service manager 740, which may be implemented as part of API service manager 742. Notification services manager 740 in the preferred embodiment provides notification support to dispatch specific notifications to an appropriate user process via the appropriate API return, or by another path. This mechanism permits notifications to be routed to any authorized process—not just back to a process that specified a notification mechanism.
API Service Manager 742
The preferred embodiment API Service Manager 742 is implemented as a service interface to the RPC service manager 732 All user API requests are built on top of this basic interface. The API Service Manager 742 preferably provides a service instance for each running user application 608.
Most RPC calls to ROS functions supported by API Service Manager 742 in the preferred embodiment may map directly to service calls with some additional parameter checking. This mechanism permits developers to create their own extended API libraries with additional or changed functionality.
In the scenario discussed above in which ROS 602 is formed by integrating “add on” with a preexisting operating system, the API service 742 code may be shared (eg, resident in a host environment like a Windows DLL), or it may be directly linked with an applications's code—depending on an application programmer's implementation decision, and/or the type of electronic appliance 600. The Notification Service Manager 740 may be implemented within API 682. These components interface with Notification Service component 686 to provide a transition between system and user space.
Secure Database Service Manager (“SDSM”) 744
There are at least two ways that may be used for managing secure database 600:
a commercial database approach, and
a site record number approach.
Which way is chosen may be based on the number of records that a VDE site stores in the secure database 610.
The commercial database approach uses a commercial database to store securely wrappered records in a commercial database. This way may be preferred when there are a large number of records that are stored in the secure database 610. This way provides high speed access, efficient updates, and easy integration to host systems at the cost of resource usage (most commercial database managers use many system resources).
The site record number approach uses a “site record number” (“SRN”) to locate records in the system This scheme is preferred when the number of records stored in the secure database 610 is small and is not expected to change extensively over time. This way provides efficient resources use with limited update capabilities. SRNs permit further grouping of similar data records to speed access and increase performance.
Since VDE 100 is highly scalable, different electronic appliances 600 may suggest one way more than the other. For example, in limited environments like a set top, PDA, or other low end electronic appliance, the SRN scheme may be preferred because it limits the amount of resources (memory and processor) required. When VDE is deployed on more capable electronic appliances 600 such as desktop computers, servers and at clearinghouses, the commercial database scheme may be more desirable because it provides high performance in environments where resources are not limited.
One difference between the database records in the two approaches is whether the records are specified using a full VDE ID or SRN. To translate between the two schemes, a SRN reference may be replaced with a VDE ID database reference wherever it occurs Similarly, VDE IDs that are used as indices or references to other items may be replaced by the appropriate SRN value.
In the preferred embodiment, a commercially available database manager 730 is used to maintain secure database 610. ROS 602 interacts with commercial database manager 730 through a database driver 750 and a database interface 748. The database interface 748 between ROS 602 and external, third party database vendors' commercial database manager 730 may be an open standard to permit any database vendor to implement a VDE compliant database driver 750 for their products.
ROS 602 may encrypt each secure database 610 record so that a VDE-provided security layer is “on top of” the commercial database structure. In other words, SPE 736 may write secure records in sizes and formats that may be stored within a database record structure supported by commercial database manager 730. Commercial database manager 730 may then be used to organize, store, and retrieve the records In some embodiments, it may be desirable to use a proprietary and/or newly created database manager in place of commercial database manager 730. However, the use of commercial database manager 730 may provide certain advantages such as, for example, an ability to use already existing database management product(s).
The Secure Database Services Manager (“SDSM”) 744 makes calls to an underlying commercial database manager 730 to obtain, modify, and store records in secure database 610. In the preferred embodiment, “SDSM” 744 provides a layer “on top of” the structure of commercial database manager 730. For example, all VDE-secure information is sent to commercial database manager 730 in encrypted form. SDSM 744 in conjunction with cache manager 746 and database interface 748 may provide record management, caching (using cache manager 746), and related services (on top of) commercial database systems 730 and/or record managers. Database Interface 748 and cache manager 746 in the preferred embodiment do not present their own RSI, but rather the RPC Manager 732 communicates to them through the Secure Database Manager RSI 744 a.
Name Services Manager 752
The Name Services Manager 752 supports three subservices: user name services, host name services, and services name services. User name services provides mapping and lookup between user name and user ID numbers, and may also support other aspects of user-based resource and information security. Host name services provides mapping and lookup between the names (and other information, such as for example address, communications connection/routing information, etc) of other processing resources (eg, other host electronic appliances) and VDE node Ids. Services name service provides a mapping and lookup between services names and other pertinent information such as connection information (eg, remotely available service routing and contact information) and service IDs.
Name Services Manager 752 in the preferred embodiment is connected to External Services Manager 772 so that it may provide external service routing information directly to the external services manager. Name services manager 752 is also connected to secure database manager 744 to permit the name services manager 752 to access name services records stored within secure database 610.
External Services Manager 772 & Services Transport Layer 786
The External Services Manager 772 provides protocol support capabilities to interface to external service providers. External services manager 772 may, for example, obtain external service routing information from name services manager 752, and then initiate contact to a particular external service (e.g., another VDE electronic appliance 600, a financial clearinghouse, etc.) through communications manager 776. External services manager 772 uses a service transport layer 786 to supply communications protocols and other information necessary to provide communications.
There are several important examples of the use of External Services Manager 772. Some VDE objects may have some or all of their content stored at an Object Repository 728 on an electronic appliance 600 other than the one operated by a user who has, or wishes to obtain, some usage rights to such VDE objects. In this case, External Services Manager 772 may manage a connection to the electronic appliance 600 where the VDE objects desired (or their content) is stored. In addition, file system 687 may be a network file system (e.g., Netware, LANtastic, NFS, etc.) that allows access to VDE objects using redirecter 684. Object switch 734 also supports this capability.
If External Services Manager 772 is used to access VDE objects, many different techniques are possible. For example, the VDE objects may be formatted for use with the World Wide Web protocols (HTML, HTTP, and TJRL) by including relevant headers, content tags, host ID to URL conversion (e.g., using Name Services Manager 752) and an HTTP-aware instance of Services Transport Layer 786.
In other examples, External Services Manager 772 may be used to locate, connect to, and utilize remote event processing services, smart agent execution services (both to provide these services and locate them), certification services for Public Keys, remote Name Services; and other remote functions either supported by ROS 602 RPCs (e.g., have RSIs), or using protocols supported by Services Transport Layer 786.
Outgoing Administrative Object Manager 754
Outgoing administrative object manager 754 receives administrative objects from object switch 734, object repository manager 770 or other source for transmission to another VDE electronic appliance. Outgoing administrative object manager 754 takes care of sending the outgoing object to its proper destination. Outgoing administrative object manager 754 may obtain routing information from name services manager 752, and may use communications service 776 to send the object. Outgoing administrative object manager 754 typically maintains records (in concert with SPE 503) in secure database 610 (e.g., shipping table 444) that reflect when objects have been successfully transmitted, when an object should be transmitted, and other information related to transmission of objects.
Incoming Administrative Object Manager 756
Incoming administrative object manager 756 receives administrative objects from other VDE electronic appliances 600 via communications manager 776. It may route the object to object repository manager 770, object switch 734 or other destination. Incoming administrative object manager 756 typically maintains records (in concert with SPE 503) in secure database 610 (e.g., receiving table 446) that record which objects have been received, objects expected for receipt, and other information related to received and/or expected objects.
Object Repository Manager 770
Object repository manager 770 is a form of database or file manager. It manages the storage of VDE objects 300 in object repository 728, in a database, or in the file system 687. Object repository manager 770 may also provide the ability to browse and/or search information related to objects (such as summaries of content, abstracts, reviewers' commentary, schedules, promotional materials, etc.), for example, by using INFORMATION methods associated with VDE objects 300.
Object Submittal Manager 774
Object submittal manager 774 in the preferred embodiment provides an interface between an application 608 and object switch 734, and thus may be considered in some respects part of API 682. For example, it may allow a user application to create new VDE objects 300. It may also allow incoming/outgoing administrative object managers 756, 754 to create VDE objects 300 (administrative objects).
In one of its roles or instances, object submittal manager 774 provides a user interface 774 a that allows the user to create an object configuration file 1240 specifying certain characteristics of a VDE object 300 to be created. This user interface 774 a may, for example, allow the user to specify that she wants to create an object, allow the user to designate the content the object will contain, and allow the user to specify certain other aspects of the information to be contained within the object (e.g., rules and control information, identifying information, etc.).
Part of the object definition task 1220 in the preferred embodiment may be to analyze the content or other information to be placed within an object. Object definition user interface 774 a may issue calls to object switch 734 to analyze “content” or other information that is to be included within the object to be created in order to define or organize the content into “atomic elements” specified by the user. As explained elsewhere herein, such “atomic element” organizations might, for example, break up the content into paragraphs, pages or other subdivisions specified by the user, and might be explicit (e.g., inserting a control character between each “atomic element”) or implicit. Object switch 734 may receive static and dynamic content (e.g., by way of time independent stream interface 762 and real time stream interface 760), and is capable of accessing and retrieving stored content or other information stored within file system 687.
The result of object definition 1240 may be an object configuration file 1240 specifying certain parameters relating to the object to be created. Such parameters may include, for example, map tables, key management specifications, and event method parameters. The object construction stage 1230 may take the object configuration file 1240 and the information or content to be included within the new object as input, construct an object based on these inputs, and store the object within object repository 728.
Object construction stage 1230 may use information in object configuration file 1240 to assemble or modify a container. This process typically involves communicating a series of events to SPE 503 to create one or more PERCs 808, public headers, private headers, and to encrypt content, all for storage in the new object 300 (or within secure database 610 within records associated with the new object).
The object configuration file 1240 may be passed to container manager 764 within object switch 734. Container manager 734 is responsible for constructing an object 300 based on the object configuration file 1240 and further user input. The user may interact with the object construction 1230 through another instance 774(2) of object submittal manager 774. In this further user interaction provided by object submittal manager 774, the user may specify permissions, rules and/or control information to be applied to or associated with the new object 300. To specify permissions, rules and control information, object submittal manager 774 and/or container manager 764 within object switch 734 generally will, as mentioned above, need to issue calls to SPE 503 (e.g., through gateway 734) to cause the SPE to obtain appropriate information from secure database 610, generate appropriate database items, and store the database items into the secure database 610 and/or provide them in encrypted, protected form to the object switch for incorporation into the object. Such information provided by SPE 503 may include, in addition to encrypted content or other information, one or more PERCs 808, one or more method cores 1000′, one or more load modules 1100, one or more data structures such as UDEs 1200 and/or MDEs 1202, along with various key blocks, tags, public and private headers, and error correction information.
The container manager 764 may, in cooperation with SPE 503, construct an object container 302 based at least in part on parameters about new object content or other information as specified by object configuration file 1240. Container manager 764 may then insert into the container 302 the content or other information (as encrypted by SPE 503) to be included in the new object. Container manager 764 may also insert appropriate permissions, rules and/or control information into the container 302 (this permissions, rules and/or control information may be defined at least in part by user interaction through object submittal manager 774, and may be processed at least in part by SPE 503 to create secure data control structures). Container manager 764 may then write the new object to object repository 687, and the user or the electronic appliance may “register” the new object by including appropriate information within secure database 610.
Communications Subsystem 776
Communications subsystem 776, as discussed above, may be a conventional communications service that provides a network manager 780 and a mail gateway manager 782. Mail filters 784 may be provided to automatically route objects 300 and other VDE information to/tom the outside world. Communications subsystem 776 may support a real time content feed 684 from a cable, satellite or other telecommunications link.
Secure Processing Environment 503
As discussed above in connection with
In the preferred embodiment, an SPE 503 is supported by the hardware resources of an SPU 500. An HPE 655 may be supported by general purpose processor resources and rely on software techniques for security/protection. HPE 655 thus gives ROS 602 the capability of assembling and executing certain component assemblies 690 on a general purpose CPU such as a microcomputer, minicomputer, mainframe computer or supercomputer processor. In the preferred embodiment, the overall software architecture of an SPE 503 may be the same as the software architecture of an HPE 655. An HPE 655 can “emulate” SPE 503 and associated SPU 500, i.e., each may include services and resources needed to support an identical set of service requests from ROS 602 (although ROS 602 may be restricted from sending to an HPE certain highly secure tasks to be executed only within an SPU 500).
Some electronic appliance 600 configurations might, include both an SPE 503 and an HPE 655. For example, the HPE 655 could perform tasks that need lesser (or no) security protections, and the SPE 503 could perform all tasks that require a high degree of security. This ability to provide serial or concurrent processing using multiple SPE and/or HPE arrangements provides additional flexibility, and may overcome limitations imposed by limited resources that can practically or cost-effectively be provided within an SPU 500. The cooperation of an SPE 503 and an EPE 655 may, in a particular application, lead to a more efficient and cost effective but nevertheless secure overall processing environment for supporting and providing the secure processing required by VDE 100. As one example, an HPE 655 could provide overall processing for allowing a user to manipulate released object 300 ‘contents,’ but use SPE 503 to access the secure object and release the information from the object.
As shown in
Channel Services Manager 562
SPE RPC Manager 550′
Time Base Manager 554
Encryption/Decryption Manager 556
Key and Tag Manager 558
Summary Services Manager 560
Authentication Manager/Service Communications
Random Value Generator 565
Secure Database Manager 566
Other Services 592.
Each of the major functional blocks of PPE 650 is discussed in detail below.
I. SPE Kernel/Dispatcher 552
The Kernel/Dispatcher 552 provides an operating system “kernel” that runs on and manages the hardware resources of SPU 500. This operating system “kernel” 552 provides a self-contained operating system for SPU 500; it is also a part of overall ROS 602 (which may include multiple OS kernels, including one for each SPE and HPE ROS is controlling/managing). Kernel/dispatcher 552 provides SPU task and memory management, supports internal, SPU hardware interrupts, provides certain “low level services,” manages “DTD” data structures, and manages the SPU bus interface unit 530. Kernel/dispatcher 552 also includes a load module execution manager 568 that can load programs into secure execution space for execution by SPU 500.
In the preferred embodiment, kernel/dispatcher 552 may include the following software/functional components:
load module execution manager 568
task manager 576
memory manager 578
virtual memory manager 580
“low level” services manager 582
internal interrupt handlers 584
BIU handler 586 (may not be present in HPE 655)
Service interrupt queues 588
DTD Interpreter 590.
At least parts of the kernel/dispatcher 552 are preferably stored in SPU firmware loaded into SPU ROM 532. An example of a memory map of SPU ROM 532 is shown in
One of the functions performed by kernel/dispatcher 552 is to receive RPC calls from ROS RPC manager 732. As explained above, the ROS Kernel RPC manager 732 can route RPC calls to the SPE 503 (via SPE Device Driver 736 and its associated RSI 736 a) for action by the SPE. The SPE kernel/dispatcher 552 receives these calls and either handles them or passes them on to SPE RPC manager 550 for routing internally to SPE 503. SPE 503 based processes can also generate RPC requests. Some of these requests can be processed internally by the SPE 503. If they are not internally serviceable, they may be passed out of the SPE 503 through SPE kernel/dispatcher 552 to ROS RPC manager 732 for routing to services external to SPE 503.
A. Kernel/Dispatcher Task Management
Kernel/dispatcher task manager 576 schedules and oversees tasks executing within SPE 503 (PPE 650). SPE 503 supports many types of tasks. A “channel” (a special type of task that controls execution of component assemblies 690 in the preferred’ embodiment) is treated by task manager 576 as one type of task. Tasks are submitted to the task manager 576 for execution. Task manager 576 in turn ensures that the SPE 503/SPU 500 resources necessary to execute the tasks are made available, and then arranges for the SPU microprocessor 520 to execute the task.
Any call to kernel/dispatcher 552 gives the kernel an opportunity to take control of SPE 503 and to change the task or tasks that are currently executing. Thus, in the preferred embodiment kernel/dispatcher task manager 576 may (in conjunction with virtual memory manager 580 and/or memory manager 578) “swap out” of the execution space any or all of the tasks that are currently active, and “swap in” additional or different tasks.
SPE tasking managed by task manager 576 may be either “single tasking” (meaning that only one task may be active at a time) or “multi-tasking” (meaning that multiple tasks may be active at once). SPE 503 may support single tasking or multi-tasking in the preferred embodiment. For example, “high end” implementations of SPE 503 (e.g., in server devices) should preferably include multi-tasking with “preemptive scheduling.” Desktop applications may be able to use a simpler SPE 503, although they may still require concurrent execution of several tasks. Set top applications may be able to use a relatively simple implementation of SPE 503, supporting execution of only one task at a time. For example, a typical set top implementation of SPU 500 may perform simple metering, budgeting and billing using subsets of VDE methods combined into single “aggregate” load modules to permit the various methods to execute m a single tasking environment. However, an execution environment that supports only single tasking may limit use with more complex control structures. Such single tasking versions of SPE 503 trade flexibility in the number and types of metering and budgeting operations for smaller run time RAM size requirements. Such implementations of SPE 503 may (depending upon memory limitations) also be limited to metering a single object 300 at a time. Of course, variations or combinations are possible to increase capabilities beyond a simple single tasking environment without incurring the additional cost required to support “full multitasking.”
In the preferred embodiment, each task in SPE 503 is represented by a “swap block,” which may be considered a “task” in a traditional multitasking architecture A “swap block” in the preferred embodiment is a bookkeeping mechanism, used by task manager 576 to keep track of tasks and subtasks. It corresponds to a chunk of code and associated references that “fits” within the secure execution environment provided by SPU 500. In the preferred embodiment, it contains a list of references to shared data elements (e.g., load modules 1100 and UDEs 1200), private data elements (e.g., method data and local stack), and swapped process “context” information (e.g., the register set for the process when it is not processing).
Task manager 576 may use Memory Manager 578 to help it perform this swapping operation. Tasks may be swapped out of the secure execution space by reading appropriate information from RAM and other storage internal to SPU 500, for example, and writing a “swap block” to secondary storage 652. Kernel 552 may swap a task back into the secure execution space by reading the swap block from secondary storage 652 and writing the appropriate information back into SPU RAM 532. Because secondary storage 652 is not secure, SPE 503 must encrypt and cryptographically seal (e.g., using a one-way bash function initialized with a secret value known only inside the SPU 500) each swap block before it writes it to secondary storage. The SPE 503 must decrypt and verify the cryptographic seal for each swap block read from secondary storage 652 before the swap block, can be returned to the secure execution space for further execution.
Loading a “swap block” into SPU memory may require one or more “paging operations” to possibly first save, and then flush, any “dirty pages” (i.e., pages changed by SPE 503) associated with the previously loaded swap blocks, and to load all required pages for the new swap block context.
Kernel/dispatcher 522 preferably manages the “swap block” using service interrupt queues 588. These service interrupt queues 588 allow kernel/dispatcher 552 to track tasks (swap blocks) and their status (running, “swapped out,” or “asleep”). The kernel/dispatcher 552 in the preferred embodiment may maintain the following service interrupt queues 588 to help it manage the “swap blocks”:
Those tasks that are completely loaded in the execution space and are waiting for and/or using execution cycles from microprocessor 502 are in the RUN queue. Those tasks that are “swapped” out (e.g., because they are waiting for other swappable components to be loaded) are referenced in the SWAP queue. Those tasks that are “asleep” (e.g., because they are “blocked” on some resource other than processor cycles or are not needed at the moment) are referenced in the SLEEP queue. Kernel/dispatcher task manager 576 may, for example, transition tasks between the RUN and SWAP queues based upon a “round-robin” scheduling algorithm that selects the next task waiting for service, swaps in any pieces that need to be paged in, and executes the task. Kernel/dispatcher 552 task manager 576 may transition tasks between the SLEEP queue and the “awake” (i.e., RUN or SWAP) queues as needed;
When two or more tasks try to write to the same data structure in a multi-tasking environment, a situation exists that may result in “deadly embrace” or “task starvation.” A “multi-threaded” tasking arrangement may be used to prevent “deadly embrace” or “task starvation” from happening. The preferred embodiment kernel/dispatcher 552 may support “single threaded” or “multi-threaded” tasking.
In single threaded applications, the kernel/dispatcher 552 “locks” individual data structures as they are loaded. Once locked, no other SPE 503 task may load them and will “block” waiting for the data structure to become available. Using a single threaded SPE 503 may, as a practical matter, limit the ability of outside vendors to create load modules 1100 since there can be no assurance that they will not cause a “deadly embrace” with other VDE processes about which outside vendors may know little or nothing. Moreover, the context swapping of a partially updated record might destroy the integrity of the system, permit unmetered use, and/or lead to deadlock. In addition, such “locking” imposes a potentially indeterminate delay into a typically time critical process, may limit SPE 503 throughput, and may increase overhead.
This issue notwithstanding, there are other significant processing issues related to building single-threaded versions of SPE 503 that may limit its usefulness or capabilities under some circumstances. For example, multiple concurrently executing tasks may not be able to process using the same often-needed data structure in a single-threaded SPE 503. This may effectively limit the number of concurrent tasks to one. Additionally, single-threadedness may eliminate the capability of producing accurate summary budgets based on a number of concurrent tasks since multiple concurrent tasks may not be able to effectively share the same summary budget data structure. Single-threadedness may also eliminate the capability to support audit processing concurrently with other processing. For example, real-time feed processing might have to be shut down in order to audit budgets and meters associated with the monitoring process.
One way to provide a more workable “single-threaded” capability is for kernel/dispatcher 552 to use virtual page handling algorithms to track “dirty pages” as data areas are written to. The “dirty pages” can be swapped in and out with the task swap block as part of local data associated with the, swap block. When a task exits, the “dirty pages” can be merged with the current data structure (possibly updated by another task for SPU 500) using a three-way merge algorithm (i.e., merging the original data structure, the current data structure, and the “dirty pages” to form a new current data structure). During the update process, the data structure can be locked as the pages are compared and swapped. Even though this virtual paging solution might be workable for allowing single threading in some applications, the vendor limitations mentioned above may limit the use of such-single threaded implementations in some cases to dedicated hardware. Any implementation that supports multiple users (e.g., “smart home” set tops, many desk tops and certain PDA applications, etc.) may hit limitations of a single threaded device in certain circumstances.
It is preferable when these limitations are unacceptable to use a full “multi-threaded” data structure write capabilities. For example, a type of “two-phase commit” processing of the type used by database vendors may be used to allow data structure sharing between processes. To implement this “two-phase commit” process, each swap block may contain page addresses for additional memory blocks that will be used to store changed information. A change page is a local copy of a piece of a data element that has been written by an SPE process. The changed page(s) references associated with a specific data structure are stored locally to the swap block in the preferred embodiment.
For example, SPE 503 may support two (change pages) per data structure. This limit is easily alterable by changing the size of the swap block structure and allowing the update algorithm to process all of the changed pages. The “commit” process can be invoked when a swap block that references changed pages is about to be discarded. The commit process takes the original data element that was originally loaded (e.g., UDE0), the current data element (e.g., UDEn) and the changed pages, and merges them to create a new copy of the data element (e.g., UDEn+1). Differences can be resolved by the DTD interpreter 590 using a DTD for the data element. The original data element is discarded (e.g., as determined by its DTD use count) if no other swap block references it.
B. Kernel/Dispatcher Memory Management
Memory manager 578 and virtual memory manager 580 in the preferred embodiment manage ROM 532 and RAM 534 memory within SPU 500 in the preferred embodiment Virtual memory manager 580 provides a fully “virtual” memory system to increase the amount of “virtual” RAM available in the SPE secure execution space beyond the amount of physical RAM 534 a provided by SPU 500. Memory manager 578 manages the memory in the secure execution space, controlling how it is accessed, allocated and deallocated. SPU MMU 540, if present, supports virtual memory manager 580 and memory manager 578 in the preferred embodiment. In some “minimal” configurations of SPT 500 there may be no virtual memory capability and all memory management functions will be handled by memory manager 578. Memory management can also be used to help enforce the security provided by SPE 503. In some classes of SPUs 500, for example, the kernel memory manager 578 may use hardware memory management unit (MMU) 540 to provide page level protection within the SPU 500. Such a hardware-based memory management system provides an effective mechanism for protecting VDE component assemblies 690 from compromise by “rogue” load modules.
In addition, memory management provided by memory manager 578 operating at least in part based on hardware-based MMU 540 may securely implement and enforce a memory architecture providing multiple protection domains. In such an architecture, memory is divided into a plurality of domains that are largely isolated from each other and share only specific memory areas under the control of the memory manager 578. An executing process cannot access memory outside its domain and can only communicate with other processes through services provided by and mediated by privileged kernel/dispatcher software 552 within the SPU 500. Such an architecture is more secure if it is enforced at least in part by hardware within MMU 540 that cannot be modified by any software-based process executing within SPU 500.
In the preferred embodiment, access to services implemented in the ROM 532 and to physical resources such as NVRAM 534 b and RTC 528 are mediated by the combination of privileged kernel/dispatcher software 552 and hardware within MMU 540. ROM 532 and RTC 528 requests are privileged in order to protect access to critical system component routines (e.g., RTC 528).
Memory manager 578 is responsible for allocating and deallocating memory; supervising sharing of memory resources ‘between processes; and enforcing memory access/use restriction. The SPE kernel/dispatcher memory manager 578 typically initially allocates all memory to kernel 552, and may be configured to permit only process-level access to pages as they are loaded by a specific process. In one example SPE operating system configuration, memory manager 578 allocates memory using a simplified allocation mechanism. A list of each memory page accessible in SPE 503 may be represented using a bit map allocation vector, for example. In a memory block, a group of contiguous memory pages may start at a specific page number. The size of the block is measured by the number of memory pages it spans. Memory allocation may be recorded by setting/clearing the appropriate bits in the allocation vector.
To assist in memory management functions, a “dope vector” may be prepended to a memory block. The “dope vector” may contain information allowing memory manager 578 to manage that memory block. In its simplest form, a memory block may be structured as a “dope vector” followed by the actual memory area of the block. This “dope vector” may include the block number, support for dynamic paging of data elements, and a marker to detect memory overwrites. Memory manager 578 may track memory blocks by their block number and convert the block number to an address before use. All accesses to the memory area can be automatically offset by the size of the “dope vector” during conversion from a block memory to a physical address. “Dope vectors” can also be used by virtual memory manager 580 to help manage virtual memory.
The ROM 532 memory management task performed by memory manager 578 is relatively simple in the preferred embodiment. All ROM 532 pages may be flagged as “read only” and as “non-pagable.” EEPROM 532B memory management may be slightly more complex since the “burn count” for each EEPROM page may need to be retained. SPU EEPROM 532B may need to be protected from all uncontrolled writes to conserve the limited writable lifetime of certain types of this memory. Furthermore, EEPROM pages may in some cases not be the same size as memory management address pages.
SPU NVRAM 534 b is preferably battery backed RAM that has a few access restrictions. Memory manager 578 can ensure control structures that must be located in NVRAM 534 b are not relocated during “garbage collection” processes. As discussed above, memory manager 578 (and MMU 540 if present) may protect NVRAM 534 b and RAM 534 a at a page level to prevent tampering by other processes.
Virtual memory manager 580 provides paging for programs and data between SPU external memory and SPU internal RAM 534 a. It is likely that data structures and executable processes will exceed the limits of any SPU 500 internal memory. For example, PERCs 808 and other fundamental control structures may be fairly large, and “bit map meters” may be, or become, very large. This eventuality may be addressed in two ways:
(1) subdividing load modules 1100; and
(2) supporting virtual paging.
Load modules 1100 can be “subdivided” in that in many instances they can be broken up into separate components only a subset of which must be loaded for execution. Load modules 1100 are the smallest pagable executable element in this example. Such load modules 1100 can be broken up into separate components (e.g., executable code and plural data description blocks), only one of which must be loaded for simple load modules to execute. This structure permits a load module 1100 to initially load only the executable code and to load the data description blocks into the other system pages on a demand basis. Many load modules 1100 that have executable sections that are too large to fit into SPU 500 can be restructured into two or more smaller independent load modules. Large load modules may be manually “split” into multiple load modules that are “chained” together using explicit load module references.
Although “demand paging” can be used to relax some of these restrictions, the preferred embodiment uses virtual paging to manage large data structures and executables. Virtual Memory Manager 580 “swaps” information (e.g., executable code and/or data structures) into and out of SPU RAM 534 a, and provides other related virtual memory management services to allow a full virtual memory management capability. Virtual memory management may be important to allow limited resource SPU 500 configurations to execute large and/or multiple tasks.
C. SPE Load Module Execution Manager 568
The SPE (HPE) load module execution manager (“LMEM”) 568 loads executables into the memory managed by memory manager 578 and executes them. LMEM 568 provides mechanisms for tracking load modules that are currently loaded inside the protected execution environment. LMEM 568 also provides access to basic load modules and code fragments stored within, and thus always available to, SPE 503. LMEM 568 may be called, for example, by load modules 1100 that want to execute other load modules.
In the preferred embodiment, the load module execution manager 568 includes a load module executor (“program loader”) 570, one or more internal load modules 572, and library routines 574. Load module executor 570 loads executables into memory (eg, after receiving a memory allocation from memory manager 578) for execution. Internal load module library 572 may provide a set of commonly used basic load modules 1100 (stored in ROM 532 or NVRAM 534 b, for example). Library routines 574 may provide a set of commonly used code fragments/routines (e.g., bootstrap routines) for execution by SPE 503.
Library routines 574 may provide a standard set of library functions in ROM 532. A standard list of such library functions along with their entry points and parameters may be used. Load modules 1100 may call these routines (e.g., using an interrupt reserved for this purpose). Library calls may reduce the size of load modules by moving commonly used code into a central location and permitting a higher degree of code reuse. All load modules 1100 for use by SPE 503 are preferably referenced by a load module execution manager 568 that maintains and scans a list of available load modules and selects the appropriate load module for execution. If the load module is not present within SPE 503, the task is “slept and LMEM 568 may request that the load module 1100 be loaded from secondary storage 562. This request may be in the form of an RPC call to secure database manager 566 to retrieve the load module and associated data structures, and a call to encrypt/decrypt manager 556 to decrypt the load module before storing it in memory allocated by memory manager 578.
In somewhat more detail, the preferred embodiment executes a load module 1100 by passing the load module execution manager 568 the name (e.g., VDE ID) of the desired load module 1100. LMEM 568 first searches the list of “in memory” and “built-in” load modules 572. If it cannot find the desired load module 1100 in the list, it requests a copy from the secure database 610 by issuing an RPC request that may be handled by ROS secure database manager 744 shown in
In response to this “initialization” event, the control method may construct the channel detail records 594(1), . . . 594(N) used to handle further events other than the “initialization” event. The control method executing “within” the channel may access the various components it needs to construct associated component assemblies 690 based on the “blueprint” accessed at step 1127 (block 1137). Each of these components is bound to the channel 594 (block 1139) by constructing an associated channel detail record specifying the method core(s) 1000′, load module(s) 1100, and associated data structure(s) (eg, UDE(s) 1200 and/or MDE(s) 1202) needed to respond to the event. The number of channel detail records will depend on the number of events that can be serviced by the “right,” as specified by the “blueprints (i.e., URT 464). During this process, the control method will construct “swap blocks” to, in effect, set up all required tasks and obtain necessary memory allocations from kernel 562. The control method will, as necessary, issue calls to secure database manager 566 to retrieve necessary components from secure database 610, issue calls to encrypt/decrypt manager 556 to decrypt retrieved encrypted information, and issue calls to key and tag manager 558 to ensure that all retrieved components are validated. Each of the various component assemblies 690 so constructed are “bound” to the channel through the channel header event code/pointer records 598 and by constructing appropriate swap blocks referenced by channel detail records 594(1), . . . 594(N). When this process is complete, the channel 594 has been completely constructed and is ready to respond to further events. As a last step, the
Once a channel 594 has been constructed in this fashion, it will respond to events as they arrive. Channel services manager 562 is responsible for dispatching events to channel 594. Each time a new event arrives (e.g., via an RPC call), channel services manager 562 examines the event to determine whether a channel already exists that is capable of processing it. If a channel does exist, then the channel services manager 562 passes the event to that channel. To process the event, it may be necessary for task manager 576 to “swap in” certain “swappable blocks” defined by the channel detail records as active tasks. In this way, executable component assemblies 690 formed during the channel open process shown in
To destroy a channel, the various swap blocks defined by the channel detail records are destroyed, the identification information in the channel header 596 is wiped clean, and the channel is made available for re-allocation by the “channel 0” “open channel” task.
D. SPE Interrupt Handlers 584
As shown in
“tick” of RTC 528
interrupt from bus interface 530
power fail interrupt
watchdog timer interrupt interrupt from encrypt/decrypt engine 522
memory interrupt (e.g., from MMU 540).
When an interrupt occurs, an interrupt controller within microprocessor 520 may cause the microprocessor to begin executing an appropriate interrupt handler. An interrupt handler is a piece of software/firmware provided by kernel/dispatcher 552 that allows microprocessor 520 to perform particular functions upon the occurrence of an interrupt. The interrupts may be “vectored” so that different interrupt sources may effectively cause different interrupt handlers to be executed.
A “timer tick” interrupt is generated when the real-time RTC 528 “pulses”. The timer tick interrupt is processed by a timer tick interrupt handler to calculate internal device date/time and to generate timer events for channel processing.
The bus interface unit 530 may generate a series of interrupts. In the preferred embodiment, bus interface 530, modeled after a USART, generates interrupts for various conditions (e.g., “receive buffer full,” “transmitter buffer empty,” and “status word change”). Kernel/dispatcher 552 services the transmitter buffer empty interrupt by sending the next character from the transmit queue to the bus interface 530. Kernel/dispatcher interrupt handler 584 may service the received buffer full interrupt by reading a character, appending it to the current buffer, and processing the buffer based on the state of the service engine for the bus interface 530. Kernel/dispatcher 552 preferably processes a status word change interrupt and addresses the appropriate send/receive buffers accordingly.
SPU 500 generates a power fail interrupt when it detects an imminent power fail condition. This may require immediate action to prevent loss of information. For example, in the preferred embodiment, a power fail interrupt moves all recently written information (i.e., “dirty pages”) into non-volatile NVRAM 534 b, marks all swap blocks as “swapped out,” and sets the appropriate power fail flag to facilitate recovery processing. Kernel/dispatcher 552 may then periodically poll the “power fail bit” in a status word until the data is cleared or the power is removed completely.
SPU 500 in the example includes a conventional watchdog timer that generates watchdog timer interrupts on a regular basis. A watchdog timer interrupt handler performs internal device checks to ensure that tampering is not occurring. The internal clocks of the watchdog timer and RTC 528 are compared to ensure SPU 500 is not being paused or probed, and other internal checks on the operation of SPU 500 are made to detect tampering.
The encryption/decryption engine 522 generates an interrupt when a block of data has been processed. The kernel interrupt handler 584 adjusts the processing status of the block being encrypted or decrypted, and passes the block to the next stage of processing. The next block scheduled for the encryption service then has its key moved into the encrypt/decrypt engine 522, and the next cryptographic process started.
A memory management unit 540 interrupt is generated when a task attempts to access memory outside ‘the areas assigned to it. A memory management interrupt handler traps the request, and takes the necessary action (e.g., by initiating a control transfer to memory manager 578 and/or virtual memory manager 580). Generally, the task will be failed, a page fault exception will be generated, or appropriate virtual memory page(s) will be paged in.
E. Kernel/Dispatcher Low Level Services 582
Low level services 582 in the preferred embodiment provide “low level” functions. These functions in the preferred embodiment may include, for example, power-on initialization, device POST, and failure recovery routines. Low level services 582 may also in the preferred embodiment provide (either by themselves or in combination with authentication manager/service communications manager 564) download response-challenge and authentication communication protocols, and may provide for certain low level management of SPU 500 memory devices such as EEPROM and FLASH memory (either alone or in combination. with memory manager 578 and/or virtual memory manager 580).
F. Kernel/Dispatcher BIU Handler 586
BIU handler 586 in the preferred embodiment manages the bus interface unit 530 (if present). It may, for’ example, maintain read and write buffers for the BIU 530, provide BIU startup initialization, etc.
G Kernel/Dispatcher DTD Interpreter 690
DTD interpreter 590 in the preferred embodiment handles data formatting issues. For example, the DTD interpreter 590 may automatically open data structures such as UDEs 1200 based on formatting instructions contained within DTDs.
The SPE kernel/dispatcher 552 discussed above supports all of the other services provided by SPE 503. Those other services are, discussed below.
II. SPU Channel Services Manager 562
“Channels” are the basic task processing mechanism of SPE 503 (HPE 655) in the preferred embodiment. ROS 602 provides an event-driven interface for “methods.” A “channel” allows component assemblies 690 to service events. A “channel” is a conduit for passing “events” from services supported by SPE 503 (HPE 655) to the various methods and load modules that have been specified to process these events, and also supports the assembly of component assemblies 690 and interaction between component assemblies. In more detail, “channel” 594 is a data structure maintained by channel manager 593 that “binds” together one or more load modules 1100 and data structures (e.g., UDEs 1200 and/or MDEs 1202) into a component assembly 690. Channel services manager 562 causes load module execution manager 569 to load the component assembly 690 for execution, and may also be responsible for passing events into the channel 594 for response by a component assembly 690. In the preferred embodiment, event processing is handled as a message to the channel service manager 562.
The channel 594 is set up by the channel services manager 562 in response to the occurrence of an event. Once the channel is created, the channel services manager 562 may issue function calls to load module execution manager 568 based on the channel 594. The load module execution manager 568 loads the load modules 1100 referenced by a channel 594, and requests execution services by the kernel/dispatcher task manager 576. The kernel/dispatcher 552 treats the event processing request as a task, and executes it ‘by executing the code within the load modules 1100 referenced by the channel.
The channel services manager 562 may be passed an identification of the event (e.g., the “event code”). The channel services manager 562 parses one or more method cores' 1000′ that are part of the component assembly(ies) 690 the channel services manager is to assemble. It performs this parsing to determine which method(s) and data structure(s) are invoked by the type of event. Channel manager 562 then issues calls (e.g., to secure database manager 566) to obtain the methods and ‘data structure(s) needed to build the component assembly 690. These called-for method(s) and data structure(s) (e.g., load modules 1100, ‘UDEs 1200 and/or MDEs ‘1202) are each decrypted using encrypt/decrypt manager 556 (if necessary), and are then each validated using key and tag manager 558. Channel manager 562 constructs any necessary “jump table” references to, in effect, “link” or “bind” the elements into a single cohesive executable so the load module(s) can reference data structures and any other load module(s) in the component assembly. Channel manager 562 may then issue calls to LMEM 568 to load the executable as an active task.
“Channel header” 596 in the preferred embodiment is (or references) the data structure(s) and associated control program(s) that queues events from channel, event sources, processes these events, and releases the appropriate tasks specified in the “channel detail record” for processing. A “channel detail record” in the preferred embodiment links an event to a “swap block” (i.e., task) associated with that event. The “swap block” may reference one or more load modules 1100, UDEs 1200 and private data areas required to properly process the event. One swap block and a corresponding channel detail item is created for each different event the channel can respond to.
In the preferred embodiment, Channel Services Manager 562 may support the following (internal) calls to support the creation and maintenance ‘of channels 562:
As described in connection with
RPC manager 550 within SPE 503 (HPE 655) is not the same as RPC manager 732 shown in
SPE RPC Manager 550 and its integrated service manager uses two tables to dispatch remote procedure calls: an RPC services table, and an optional RPC dispatch table. The RPC services table describes where requests for specific services are to be routed for processing. In the preferred embodiment, this table is constructed in SPU RAM 534 a or NVRAM 534 b, and lists each RPC service “registered” within SPU 500. Each row of the RPC services table contains a service ID, its location and address, and a control byte. In simple implementations, the control byte indicates only that the service is provided internally or externally. In more complex implementations, the control byte can indicate an instance of the service (e.g., each service may have multiple “instances” in a multi-tasking environment). ROS RPC manager 732 and SPE 503 may have symmetric copies of the RPC services table in the preferred embodiment. If an RPC service is not found in the RPC services table, SPE 503 may either reject it or pass it to ROS RPC manager 732 for service.
The SPE RPC manager 550 accepts the request from the RPC service table and processes that request in accordance with the internal priorities associated with the specific service. In SPE 503, the RPC service table is extended by an RPC dispatch table. The preferred embodiment RPC dispatch table is organized as a list of Load Module references for each RPC service supported internally by SPE-503. Each row in the table contains a load module ID that services the call, a control byte that indicates whether the call can be made from an external caller, and whether the load module needed to service the call is permanently resident in SPU 500. The RPC dispatch table may be constructed in SPU ROM 532 (or EEPROM) when SPU firmware 508 is loaded into the SPU 500. If the RPC dispatch table is in EEPROM, it flexibly allows for updates to the services without load module location and version control issues.
In the preferred embodiment, SPE RPC manager 550 first references a service request against the RPC service table to determine the location of the service manager that may service the request. The RPC manager 550 then routes the service request to the appropriate service manager for action. Service requests are handled by the service manager within the SPE 503 using the RPC dispatch table to dispatch the request. Once the RPC manager 550 locates the service reference in the RPC dispatch table, the load module that services the request is called and loaded using the load module execution manager 568. The load module execution manager 568 passes control to the requested load module after performing all required context configuration, or if necessary may first issue a request to load it from the external management files 610.
SPU Time Base Manager 554
The time base manager 554 supports calls that relate td the real time clock (“RTC”) 528. In the preferred embodiment, the time base manager 554 is always loaded and ready to respond to time based requests.
The table below lists examples of basic calls that may be supported by the time base manager 554:
The Encryption/Decryption Manager 556 supports call's to the various encryption/decryption techniques supported by SPE 503/HPE 655. It may be supported ‘by a hardware-based encryption/decryption engine 522 within SPU 500. Those encryption/decryption technologies not supported by SPU encrypt/decrypt engine 522 may be provided by encrypt/decrypt manager 556 in software. The primary bulk encryption/decryption load modules preferably are loaded at all times, and ‘the’ load modules necessary for other algorithms are preferably paged in as needed. Thus, if the primary bulk encryption/decryption algorithm is DES, only the DES load modules need be permanently resident in the RAM 534 a of SPE 503/HPE 655.
The following are examples of RPC calls supported by Encrypt/Decrypt Manager 556 in the preferred embodiment
The call parameters passed may include the key to be used; mode (encryption or decryption); any needed Initialization Vectors; the desired cryptographic operating (eg, type of feedback), the identification of the cryptographic instance to be used, and the start address, destination address, and length of the block to be encrypted or decrypted.
SPU Key and Tag Manager 558
The SPU Key and Tag Manager 558 supports calls for key storage, key and management file tag look up, key convolution, and the generation of random keys, tags, and transaction numbers.
The following table shows an example of a list of SPE/HPE key and tag manager service 558 calls:
Keys and tags may be securely generated within SPE 503 (HPE 655) in the preferred embodiment. The key generation algorithm is typically specific to each type of encryption supported. The generated keys may be checked for cryptographic weakness before they are used. A request for Key and Tag Manager 558 to generate a key, tag and/or transaction number preferably takes a length as its input parameter. It generates a random number (or other appropriate key value) of the requested length as its output.
The key and tag manager 558 may support calls to retrieve specific keys from the key storage areas in SPU 500 and any keys stored external to the SPU. The basic format of the calls is to request keys by key type and key number. Many of the keys are periodically updated through contact with the VDE administrator, and are kept within SPU 500 in NVRAM 534 b or EEPROM because these memories are secure, updatable and non-volatile.
SPE 503/HPE 655 may support both Public Key type keys and Bulk Encryption type keys. The public key (PK) encryption type keys stored by SPU 500 and managed by key and tag manager 558 may include, for example, a device public key, a device private key, a PK certificate, and a public key for the certificate. Generally, public keys and certificates can be stored externally in non-secured memory if desired, but the device private key and the public key for the certificate should only be stored internally in an SPU 500 EEPROM or NVRAM 534 b. Some of the types of bulk encryption keys used by the SPU 500 may include, for example, general-purpose bulk encryption keys, administrative object private header keys, stationary object private header keys, traveling object private header keys, download/initialization keys, backup keys, trail keys, and management file keys.
As discussed above, preferred embodiment Key and Tag Manager 558 supports requests to adjust or convolute keys to make new keys that are produced in a deterministic way dependent on site and/or time, for example. Key convolution is an algorithmic process that acts on a key and some set of input parameter(s) to yield a new key. It can be used, for example, to increase the number of keys available for use without incurring additional key storage space. It may also be used, for example, as a process to “age” keys by incorporating the value of real-time RTC 528 as parameters. It can be used to make keys site specific by incorporating aspects of the site ID as parameters.
Key and Tag Manager 558 also provides services relating to tag generation and management. In the preferred embodiment, transaction and access tags are preferably stored by SPE 503 (HPE 655) in protected memory (e.g., within the NVRAM 534 b of SPU 500). These tags may be generated by key and tag manager 558. They are used to, for example, check access rights to, validate and correlate data elements. For example, they may be used to ensure components of the secured data structures are not tampered with outside of the SPU 500 Key and tag manager 558 may also support a trail transaction tag and a communications transaction tag.
SPU Summary Services Manager 560
SPE 503 maintains an audit trail in reprogrammable non-volatile memory within the SPU 500 and/or in secure database 610. This audit trail may consist of an audit summary of budget activity for financial purposes, and a security summary of SPU use. When a request is made to the SPU, it logs the request as having occurred and then notes whether the request succeeded or failed. All successful requests may be summed and stored by type in the SPU 500. Failure information, including the elements listed below, may be saved along with details of the failure:
This information may be analyzed to detect cracking attempts or to determine patterns of usage outside expected (and budgeted) norms. The audit trail histories in the SPU 500 may be retained until the audit is reported to the appropriate parties. This will allow both legitimate failure analysis and attempts to cryptoanalyze the SPU to be noted.
Summary services manager 560 may store and maintain this internal summary audit information. This audit information can be used to check for security breaches or other aspects of the operation of SPE 503. The event summaries may be maintained, analyzed and used by SPE 503 (HPE 655) or a VDE administrator to determine and potentially limit abuse of electronic appliance 600. In the preferred embodiment, such parameters may be stored in secure memory (e.g., within the NVRAM 534 b of SPU 500).
There are two basic structures for which summary services are used in the preferred embodiment. One (the “event summary data structure”) is VDE administrator specific and keeps track of events. The event summary structure may be maintained and audited during periodic contact with VDE administrators. The other is used by VDE administrators and/or distributors for overall budget. A VDE administrator may register for event-summaries and an overall budget summary at the time an electronic appliance 600 is initialized. The overall budget summary may be reported to and used by a VDE administrator in determining distribution of consumed budget (for example) in the case of corruption of secure management files 610. Participants that receive appropriate permissions can register their processes (e.g., specific budgets) with summary services manager 560, which may then reserve protected memory space (e.g., within NVRAM 534 b) and keep desired use and/or access parameters. Access to and modification of each summary can be controlled by its own access tag.
The following table shows an example of a list of PPE summary service manager 560 service calls:
In the preferred embodiment, the event summary data structure uses a fixed event number to index into a look up table. The look up table contains a value that can be configured as a counter or a counter plus limit. Counter mode may be used by VDE administrators to determine device usage. The limit mode may be used to limit tampering and attempts to misuse the electronic appliance 600. Exceeding a limit will result in SPE 503 (HPE 655) refusing to service user requests until it is reset by a VDE administrator. Calls to the system wide event summary process may preferably be built into all load modules that process the events that are of interest.
The following table shows examples of events that may be separately metered by the preferred embodiment event summary data structure:
Another, “overall currency budget” summary data structure maintained by the preferred embodiment summary services manager 560 allows registration of VDE electronic appliance 600. The first entry is used for an overall currency budget consumed value, and is registered by the VDE administrator that first initializes SPE 503 (HPE 655). Certain currency consuming load modules and audit load modules that complete the auditing process for consumed currency budget may call the summary services manager 560 to update the currency consumed value. Special authorized load modules may have access to the overall currency summary, while additional summaries can be registered for by individual providers.
SPE Authentication Manager/Service Communications Manager 564
The Authentication Manager/Service Communications Manager 564 supports calls for user password validation and “ticket” generation and validation. It may also support secure communications between SPE 503 and an external node or device (e.g., a VDE administrator or distributor). It may support the following examples of authentication-related service requests in the preferred embodiment:
Not included in the table above are calls to the secure communications service. The secure communications service provided by manager 564 may provide (e.g., in conjunction with low-level services manager 582 if desired) secure communications based on a public key (or others) challenge-response protocol. This protocol is discussed in further detail elsewhere in this document. Tickets identify users with respect to the electronic appliance 600 in the case where the appliance may be used by multiple users. Tickets may be requested by and returned to VDE software applications through a ticket-granting protocol (e.g., Kerberos). VDE components may require tickets to be presented in order to authorize particular services.
SPE Secure Database Manager 566
Secure database manager 566 retrieves, maintains and stores secure database records within secure database 610 on memory external to SPE 503. Many of these secure database files 610 are in encrypted form. All secure information retrieved by secure database manager 566 therefore must be decrypted by encrypt/decrypt manager 556 before use. Secure information (e.g., records of use) produced by SPE 503 (HPE 655) which must be stored external to the secure execution environment are also encrypted by encrypt/decrypt manager 556 before they are stored via secure database manager 566 in a secure database file 610.
For each VDE item loaded into SPE 503, Secure Database manager 566 in the preferred embodiment may search a master list for the VDE item ID, and then check the corresponding transaction tag against the one in the item to ensure that the item provided is the current item. Secure Database Manager 566 may maintain list of VDE item ID and transaction tags in a “hash structure” that can be paged into SPE 503 to quickly locate the appropriate VDE item ID. In smaller systems, a look up table approach may be used. In either case, the list should be structured as a pagable structure that allows VDE item ID to be located quickly.
The “hash based” approach may be used to sort the list into “hash buckets” that may then be accessed to provide more rapid and efficient location of items in the list. In the “hash based” approach, the VDE item IDs are “hashed” through a subset of the full item ID and organized as pages of the “hashed” table. Each “hashed” page may contain the rest of the VDE item ID and current transaction tag for each item associated with that page. The “hash” table page number may be derived from the components of the VDE item ID, such as distribution ID, item ID, site ID, user ID, transaction tag, creator ID, type and/or version. The hashing algorithm (both the algorithm itself and the parameters to be hashed) may be configurable by a VDE administrator on a site by site basis to provide optimum hash page use. An example of a hash page structure appears below:
In this example, each hash page may contain all of the VDE item IDs and transaction tags for items that have identical distributor ID, item ID, and user ID fields (site ID will be fixed for a given electronic appliance 600). These four pieces of information may thus be used as hash algorithm parameters.
The “hash” pages may themselves be frequently updated, and should carry transaction tags that are checked each time a “hash” page is loaded. The transaction tag may also be updated each time a “hash” page is written out.
As an alternative to the hash-based approach, if the number of updatable items is kept small (such as in a dedicated consumer electronic appliance 600), then assigning each updatable item a unique sequential site record number as part of its VDE item ID may allow a look up table approach to be used. Only a small number of bytes of transaction tag are needed per item, and a table transaction tag for all frequently updatable items can be kept in protected memory such as SPU NVRAM 534 b.
Random Value Generator Manager 565
Random Value Generator Manager 565 may generate random values. If a hardware-based SPU random value generator 542 is present, the Random Value Generator Manager 565 may use it to assist in generating random values.
Other SPE RPC Services 592
Other authorized RPC services may be included in SPU 500 by having them “register” themselves in the RPC Services Table and adding their entries to the RPC Dispatch Table. For example, one or more component assemblies 690 may be used to provide additional services as an integral part of SPE 503 and its associated operating system. Requests to services not registered in these tables will be passed out of SPE 503 (HPE 655) for external servicing.
SPE 503 Performance Considerations
Performance of SPE 503 (HPE 655) is a function of:
The complexity of component assembly processes along with the number of simultaneous component assembly processes is perhaps the primary factor in determining performance. These factors combine to determine the amount of code and data and must be resident in SPU 500 at any one time (the minimum device size) and thus the number of device size “chunks” the processes must be broken down into. Segmentation inherently increases run time size over simpler models. Of course, feature limited versions of SPU 500 may be implemented using significantly smaller amounts of RAM 534. “Aggregate” load modules as described above may remove flexibility in configuring VDE structures and also further limit the ability of participants to individually update otherwise separated elements, but may result in a smaller minimum device size. A very simple metering version of SPU 500 can be constructed to operate with minimal device resources.
The amount of RAM 534 internal to SPU 500 has more impact on the performance of the SPE 503 than perhaps any other aspect of the SPU. The flexible nature of VDE processes allows use of a large number of load modules, methods and user data elements. It is impractical to store more than a small number of these items in ROM 532 within SPU 500. Most of the code and data structures needed to support a specific VDE process will need to be dynamically loaded into the SPU 500 for the specific VDE process when the process is invoked. The operating system within SPU 500 then may page in the necessary VDE items to perform the process. The amount of RAM 534 within SPU 500 will directly determine how large any single VDE load module plus its required data can be, as well as the number of page swaps that will be necessary to run a VDE process. The SPU I/O speed, encryption/decryption speed, and the amount of internal memory 532, 534 will directly affect the number of page swaps required in the device. Insecure external memory may reduce the wait time for swapped pages to be loaded into SPU 500, but will still incur substantial encryption/decryption penalty for each page.
In order to maintain security, SPE 503 must encrypt and cryptographically seal each block being swapped out to a storage device external to a supporting SPU 500, and must similarly decrypt, verify the cryptographic seal for, and validate each block as it is swapped into SPU 500. Thus, the data movement and encryption/decryption overhead for each swap block has a very large impact on SPE performance.
The performance of an SPU microprocessor 520 may not significantly impact the performance of the SPE 503 it supports if the processor is not responsible for moving data through the encrypt/decrypt engine 522.
VDE Secure Database 610
VDE 100 stores separately deliverable VDE elements in a secure (e.g., encrypted) database 610 distributed to each VDE electronic appliance 610. The database 610 in the preferred embodiment may store and/or manage three basic classes of VDE items:
VDE process elements, and
VDE data structures.
The following table lists examples of some of the VDE items stored in or managed by information stored in secure database 610:
Each electronic appliance 600 may have an instance of a secure database 610 that securely maintains the VDE items.
Secure database 610 may also include the following additional data structures used and maintained for administrative purposes:
Secure database 610 in the preferred embodiment does not include VDE objects 300, but rather references VDE objects stored, for example, on file system 687 and/or in a separate object repository 728. Nevertheless, an appropriate “starting point” for understanding VDE-protected information may be a discussion of VDE objects 300.
VDE Objects 300
VDE 100 provides a media independent container model for encapsulating content.
The generalized “logical object” structure 800 shown in
The “container” concept is a convenient metaphor used to give a name to the collection of elements required to make use of content or to perform an administrative-type activity. Container 302 typically includes identifying information, control structures and content (e.g., a property or administrative data). The term “container” is often (e.g., Bento/OpenDoc and OLE) used to describe a collection of information stored on a computer system's secondary storage system(s) or accessible to a computer system over a communications network on a “server's” secondary storage system. The “container” 302 provided by the preferred embodiment is not so limited or restricted. In VDE 100, there is no requirement that this information is stored together, received at the same time, updated at the same time, used for only a single object, or be owned by the same entity. Rather, in VDE 100 the container concept is extended and generalized to include real-time content and/or online interactive content passed to an electronic appliance over a cable, by broadcast, or communicated by other electronic communication means.
Thus, the “complete” VDE container 302 or logical object structure 800 may not exist at the user's location (or any other location, for that matter) at any one time. The “logical object” may exist over a particular period of time (or periods of time), rather than all at once. This concept includes the notion of a “virtual container” where important container elements may exist either as a plurality of locations and/or over a sequence of time periods (which may or may not overlap). Of course, VDE 100 containers can also be stored with all required control structures and content together. This represents a continuum: from all content and control structures present in a single container, to no locally accessible content or container specific control structures.
Although at least some of the data representing the object is typically encrypted and thus its structure is not discernible, within a PPE 650 the object may be viewed logically as a “container 302, because its structure and components are automatically and transparently decrypted.
A container model merges well with the event-driven processes and ROS 602 provided by the preferred embodiment. Under this model, content is easily subdivided into small, easily manageable pieces, but is stored so that it maintains the structural richness inherent in unencrypted content. An object oriented container model (such as Bento/OpenDoc or OLE) also provides many of the necessary “hooks” for inserting the necessary operating system integration components, and for defining the various content specific methods.
In more detail, the logical object structure 800 provided by the preferred embodiment includes a public (or unencrypted) header 802 that identifies the object and may also identify one or more owners of rights in the object and/or one or more distributors of the object. Private (or encrypted) header 804 may include a part or all of the information in the public header and further, in the preferred embodiment, will include additional data for validating and identifying the object 300 when a user attempts to register as a user of the object with a service clearinghouse, VDE administrator, or an SPU 500. Alternatively, information identifying one or more rights owners and/or distributors of the object may be located in encrypted form within encrypted header 804, along with any of said additional validating and identifying data.
Each logical object structure 800 may also include a “private body” 806 containing or referencing a set of methods 1000 (i.e., programs or procedures) that control use and distribution of the object 300. The ability to optionally incorporate different methods 1000 with each object is important to making VDE 100 highly configurable. Methods 1000 perform the basic function of defining what users (including, where appropriate, distributors, client administrators, etc.), can and cannot do with an object 300. Thus, one object 300 may come with relatively simple methods, such as allowing unlimited viewing within a fixed period of time for a fixed fee (such as the newsstand price of a newspaper for viewing the newspaper for a period of one week after the paper's publication), while other objects may be controlled by much more complicated (e.g., billing and usage limitation) methods.
Logical object structure 800 shown in
The content portion of the object is typically divided into portions called data blocks 812. Data blocks 812 may contain any sort of electronic information, such as, “content,” including computer programs, images, sound, VDE administrative information, etc. The size and number of data blocks 812 may be selected by the creator of the property. Data blocks 812 need not all be the same size (size may be influenced by content usage, database format, operating system, security and/or other considerations). Security will be enhanced by using at least one key block 810 for each data block 812 in the object, although this is not required. Key blocks 810 may also span portions of a plurality of data blocks 812 in a consistent or pseudo-random manner. The spanning may provide additional security by applying one or more keys to fragmented or seemingly random pieces of content contained in an object 300, database, or other information entity.
Many objects 300 that are distributed by physical media and/or by “out of channel” means (e.g., redistributed after receipt by a customer to another customer) might not include key blocks 810 in the same object 300 that is used to transport the content protected by the key blocks. This is because VDE objects may contain data that can be electronically copied outside the confines of a VDE node. If the content is encrypted, the copies will also be encrypted and the copier cannot gain access to the content unless she has the appropriate decryption key(s). For objects in which maintaining security is particularly important, the permission records 808 and key blocks 810 will frequently be distributed electronically, using secure communications techniques (discussed below) that are controlled by the VDE nodes of the sender and receiver. As a result, permission records 808 and key blocks 810 will frequently, in the preferred embodiment, be stored only on electronic appliances 600 of registered users (and may themselves be delivered to the user as part of a registration/initialization process). In this instance, permission records 808 and key blocks 810 for each property can be encrypted with a private DES key that is stored only in the secure memory of an SPU 500, making the key blocks unusable on any other user's VDE node. Alternately, the key blocks 810 can be encrypted with the end user's public key, making those key blocks usable only to the SPU 500 that stores the corresponding private key (or other, acceptably secure, encryption/security techniques can be employed).
In the preferred embodiment, the one or more keys used to encrypt each permission record 808 or other management information record will be changed every time the record is updated (or after a certain one or more events). In this event, the updated record is re-encrypted with new one or more keys. Alternately, one or more of the keys used to encrypt and decrypt management information may be “time aged” keys that automatically become invalid after a period of time. Combinations of time aged and other event triggered keys may also be desirable; for example keys may change after a certain number of accesses, and/or after a certain duration of time or absolute point in time. The techniques may also be used together for any given key or combination of keys. The preferred embodiment procedure for constructing time aged keys is a one-way convolution algorithm with input parameters including user and site information as well as a specified portion of the real time value provided by the SPU RTC 528. Other techniques for time aging may also be used, including for example techniques that use only user or site information, absolute points in time, and/or duration of time related to a subset of activities related to using or decrypting VDE secured content or the use of the VDE system.
VDE 100 supports many different types of “objects” 300 having the logical object structure 800 shown in
Objects may be classified in another sense based on the nature of the information they contain. A container with information content is called a “Content Object” (see
1. Stationary Objects
As shown in
2. Traveling Objects
Traveling object structure 860 may be the same as stationary object structure 850 shown in
“Traveling” objects are a class of VDE objects 300 that can specifically support “out of channel” distribution. Therefore, they include key block(s) 810 and are transportable from one electronic appliance 600 to another. Traveling objects may come with a quite limited usage related budget so that a user may use, in whole or part, content (such as a computer program, game, or database) and evaluate whether to acquire a license or further license or purchase object content. Alternatively, traveling object PERCs 808 may contain or reference budget records with, for example:
As with standard VDE objects 300, a user may be required to contact a clearinghouse service to acquire additional budgets if the user wishes to continue to use the traveling object after the exhaustion of an available budget(s) or if the traveling object (or a copy thereof) is moved to a different electronic appliance and the new appliance does not have a available credit budget(s) that corresponds to the requirements stipulated by permissions record 808.
For example, a traveling object PERC 808 may include a reference to a required budget VDE 1200 or budget options that may be found and/or are expected to be available. For example, the budget VDE may reference a consumer's VISA, MC, AMEX, or other “generic” budget that may be object independent and may be applied towards the use of a certain or classes of traveling object content (for example any movie object from a class of traveling objects that might be Blockbuster Video rentals). The budget VDE itself may stipulate one or more classes of objects it may be used with, while an object may specifically reference a certain one or more generic budgets. Under such circumstances, VDE providers will typically make information available in such a manner as to allow correct referencing and to enable billing handling and, resulting payments.
Traveling objects can be used at a receiving VDE node electronic appliance 600 so long as either the appliance carries the correct budget or budget type (e.g. sufficient credit available from a clearinghouse such as a VISA budget) either in general or for specific one or more users or user classes, or so long as the traveling object itself carries with it sufficient budget allowance or an appropriate authorization (e.g., a stipulation that the traveling object may be used on certain one or more installations or installation classes or users or user classes where classes correspond to a specific subset of installations or users who are represented by a predefined class identifiers stored in a secure database 610). After receiving a traveling object, if the user (and/or installation) doesn't have the appropriate budget(s) and/or authorizations, then the user could be informed by the electronic appliance 600 (using information stored in the traveling object) as to which one or more parties the user could contact. The party or parties might constitute a list of alternative clearinghouse providers for the traveling object from which the user selects his desired contact).
As mentioned above, traveling objects enable objects 300 to be distributed “Out-Of-Channel;” that is, the object may be distributed by an unauthorized or not explicitly authorized individual to another individual. “Out of channel” includes paths of distribution that allow, for example, a user to directly redistribute an object to another individual. For example, an object provider might allow users to redistribute copies of an object to their friends and associates (for example by physical delivery of storage media or by delivery over a computer network) such that if a friend or associate satisfies any certain criteria required for use of said object, he may do so.
For example, if a software program was distributed as a traveling object, a user of the program who wished to supply it or a usable copy of it to a friend would normally be free to do so. Traveling Objects have great potential commercial significance, since useful content could be primarily distributed by users and through bulletin boards, which would require little or no distribution overhead apart from registration with the “original” content provider and/or clearinghouse.
The “out of channel” distribution may also allow the provider to receive payment for usage and/or elsewise maintain at least a degree of control over the redistributed object. Such certain criteria might involve, for example, the registered presence at a user's VDE node of an authorized third party financial relationship, such as a credit card, along with sufficient available credit for said usage.
Thus, if the user had a VDE node, the user might be able to use the traveling object if he had an appropriate, available budget available on his VDE node (and if necessary, allocated to him), and/or if he or his VDE node belonged to a specially authorized group of users or installations and/or if the traveling object carried its own budget(s).
Since the content of the traveling object is encrypted, it can be used only under authorized circumstances unless the traveling object private header key used with the object is broken—a potentially easier task with a traveling object as compared to, for example, permissions and/or budget information since many objects may share the same key, giving a cryptoanalyst both more information in cyphertext to analyze and a greater incentive to perform cryptoanalysis.
In the case of a “traveling object,” content owners may distribute information with some or all of the key blocks 810 included in the object 300 in which the content is encapsulated. Putting keys in distributed objects 300 increases the exposure to attempts to defeat security mechanisms by breaking or cryptoanalyzing the encryption algorithm with which the private header is protected (e.g., by determining the key for the header's encryption). This breaking of security would normally require considerable skill and time, but if broken, the algorithm and key could be published so as to allow large numbers of individuals who possess objects that are protected with the same key(s) and algorithm(s) to illegally use protected information. As a result, placing keys in distributed objects 300 may be limited to content that is either “time sensitive” (has reduced value after the passage of a certain period of time), or which is somewhat limited in value, or where the commercial value of placing keys in objects (for example convenience to end-users, lower cost of eliminating the telecommunication or other means for delivering keys and/or permissions information and/or the ability to supporting objects going “out-of-channel”) exceeds the cost of vulnerability to sophisticated hackers. As mentioned elsewhere, the security of keys may be improved by employing convolution techniques to avoid storing “true” keys in a traveling object, although in most cases using a shared secret provided to most or all VDE nodes by a VDE administrator as an input rather than site ID and/or time in order to allow objects to remain independent of these values.
As shown in
The methods 1000 contained by a traveling object will typically include an installation procedure for “self registering” the object using the permission records 808 in the object (e.g., a REGISTER method). This may be especially useful for objects that have time limited value, objects (or properties) for which the end user is either not charged or is charged only a nominal fee (e.g., objects for which advertisers and/or information publishers are charged based on the number of end users who actually access published information), and objects that require widely available budgets and may particularly benefit from out-of-channel distribution (e.g., credit card derived budgets for objects containing properties such as movies, software programs, games, etc.). Such traveling objects may be supplied with or without contained budget UDEs.
One use of traveling objects is the publishing of software, where the contained permission record(s) may allow potential customers to use the software in a demonstration mode, and possibly to use the full program features for a limited time before having to pay a license fee, or before having to pay more than an initial trial fee. For example, using a time based billing method and budget records with a small pre-installed time budget to allow full use of the program for a short period of time. Various control methods may be used to avoid misuse of object contents. For example, by setting the minimum registration interval for the traveling object to an appropriately large period of time (e.g., a month, or six months or a year), users are prevented from re-using the budget records in the same traveling object.
Another method for controlling the use of traveling objects is to include time-aged keys in the permission records that are incorporated in the traveling object. This is useful generally for traveling objects to ensure that they will not be used beyond a certain date without re-registration, and is particularly useful for traveling objects that are electronically distributed by broadcast, network, or telecommunications (including both one and two way cable), since the date and time of delivery of such traveling objects aging keys can be set to accurately correspond to the time the user came into possession of the object.
Traveling objects can also be used to facilitate “moving” an object from one electronic appliance 600 to another. A user could move a traveling object, with its incorporated one or more permission records 808 from a desktop computer, for example, to his notebook computer. A traveling object might register its user within itself and thereafter only be useable by that one user. A traveling object might maintain separate budget information, one for the basic distribution budget record, and another for the “active” distribution budget record of the registered user. In this way, the object could be copied and passed to another potential user, and then could be a portable object for that user.
Traveling objects can come in a container which contains other objects. For example, a traveling object container can include one or more content objects and one or more administrative objects for registering the content object(s) in an end user's object registry and/or for providing mechanisms for enforcing permissions and/or other security functions. Contained administrative object(s) may be used to install necessary permission records and/or budget information in the end user's electronic appliance.
Content object structure 880 in the particular example shown in
Administrative object structure 870 in this example includes a public header 802, private header 804 (including a “PERC” 808) and a “private body” 806 containing methods 1000. Administrative object structure 870 in this particular example shown in
In the preferred embodiment, an administrative object may be sent, for example, by a distributor, client administrator, or, perhaps, a clearinghouse or other financial service provider, to an end user, or, alternatively, for example, by an object creator to a distributor or service clearinghouse. Administrative objects, for example, may increase or otherwise adjust budgets and/or permissions of the receiving VDE node to which the administrative object is being sent. Similarly, administrative objects containing audit information in the data area 878 of an event record 872 can be sent from end users to distributors, and/or clearinghouses and/or client administrators, who might themselves further transmit to object creators or to other participants in the object's chain of handling.
Methods 1000 in the preferred embodiment support many of the operations that a user encounters in using objects and communicating with a distributor. They may also specify what method fields are displayable to a user (e.g., use events, user request events, user response events, and user display events). Additionally, if distribution capabilities are supported in the method, then the method may support distribution activities, distributor communications with a user about a method, method modification, what method fields are displayable to a distributor, and any distribution database checks and record keeping (e.g., distribution events, distributor request events, and distributor response events).
Given the generality of the existing method structure, and the diverse array of possibilities for assembling methods, a generalized structure may be used for establishing relationships between methods. Since methods 1000 may be independent of an object that requires them during any given session, it is not possible to define the relationships within the methods themselves. “Control methods” are used in the preferred embodiment to define relationships between methods. Control methods may be object specific, and may accommodate an individual object's requirements during each session.
A control method of an object establishes relationships between other methods. These relationships are parameterized with explicit method identifiers when a record set reflecting desired method options for each required method is constructed during a registration process.
An “aggregate method” in the preferred embodiment represents a collection of methods that may be treated as a single unit. A collection of methods that are related to a specific property, for example, may be stored in an aggregate method. This type of aggregation is useful from an implementation point of view because it may reduce bookkeeping overhead and may improve overall database efficiency. In other cases, methods may be aggregated because they are logically coupled. For example, two budgets may be linked together because one of the budgets represents an overall limitation, and a second budget represents the current limitation available for use. This would arise if, for example, a large budget is released in small amounts over time.
For example, an aggregate method that includes meter, billing and budget processes can be used instead of three separate methods. Such an aggregate method may reference a single “load module” 1100 that performs all of the functions of the three separate load modules and use only one user data element that contains meter, billing and budget data. Using an aggregate method instead of three separate methods may minimize overall memory requirements, database searches, decryptions, and the number of user data element writes back to a secure database 610. The disadvantage of using an aggregate method instead of three separate methods can be a loss of some flexibility on the part of a provider and user in that various functions may no longer be independently replaceable.
A “method” 1000 provided by the preferred embodiment is a collection of basic instructions and information related to the basic instructions, that provides context, data, requirements and/or relationships for use in performing, and/or preparing to perform, the basic instructions in relation to the operation of one or more electronic appliances 600. As shown in
method “cores” 1000′;
Method Data Elements (MDEs) 1202;
User Data Elements (UDEs) 1200; and
Data Description Elements (DTDs).
Method “core” 1000′ in the preferred embodiment may contain or reference one or more data elements such as MDEs 1202 and UDEs 1200. In the preferred embodiment, MDEs 1202 and UDEs 1200 may have the same general characteristics, the main difference between these two types of data elements being that a UDE is preferably tied to a particular method as well as a particular user or group of users, whereas an MDE may be tied to a particular method but may be user independent. These MDE and UDE data structures 1200, 1202 are used in the preferred embodiment to provide input data to methods 1000, to receive data outputted by methods, or both. MDEs 1202 and UDEs 1200 may be delivered independently of method cores 1000′ that reference them, or the data structures may be delivered as part of the method cores. For example, the method core 1000′ in the preferred embodiment may contain one or more MDEs 1202 and/or UDEs 1200 (or portions thereof). Method core 1000′ may, alternately or in addition, reference one or more MDE and/or UDE data structures that are delivered independently of method core(s) that reference them.
Method cores 1000′ in the preferred embodiment also reference one or more “load modules” 1100. Load modules 1100 in the preferred embodiment comprise executable code, and may also include or reference one or more data structures called “data descriptor” (“DTD”) information. This “data descriptor” information may, for example, provide data input information to the DTD interpreter 590. DTDs may enable load modules 1100 to access (e.g., read from and/or write to) the MDE and/or UDE data elements 1202, 1200.
Method cores 1000′ may also reference one or more DTD and/or MDE data structures that contain a textual description of their operations suitable for inclusion as part of an electronic contract. The references to the DTD and MDE data structures may occur in the private header of the method core 1000′, or may be specified as part of the event table, described below.
Method cores 1000′ can be specific to a single user, or they may be shared across a number of users (e.g., depending upon the uniqueness of the method core and/or the specific user data element). Specifically, each user/group may have its own UDE 1200 and use a shared method core 1000′. This structure allows for lower database overhead than when associating an entire method core 1000′ with a user/group. To enable a user to use a method, the user may be sent a method core 1000′ specifying a UDE 1200. If that method core 1000′ already exists in the site's secure database 610, only the UDE 1200 may need to be added. Alternately, the method may create any required UDE 1200 at registration time.
An example of a possible field layout for method core 1000′ public header 802 is shown in the following table:
An example of a possible field layout for private header 804 is shown below:
Referring once again to
Thus, in the preferred embodiment, each method event record 1012 may include an event field 1014, a LM/PERC reference field 1016, and any number of data reference fields 1018. Event fields 1014 in the preferred embodiment may contain a “event code” or other information identifying the corresponding event. The LM/PERC reference field 1016 may provide a reference into the secure database 610 (or other “pointer” information) identifying a load module 1100 and/or a PERC 808 providing (or referencing) executable code to be loaded and executed to perform the method in response to the event. Data reference fields 1018 may include information referencing a UDE 1200 or a MDE 1202. These data structures may be contained in the method local data area 1008 of the method core 1000′, or they may be stored within the secure database 610 as independent deliverables.
The following table is an example of a possible more detailed field layout for a method event record 1012:
Load module 1100 contains code and static data (that is functionally the equivalent of code), and is used to perform the basic operations of VDE 100. Load modules 1100 will generally be shared by all the control structures for all objects in the system, though proprietary load modules are also permitted. Load modules 1100 may be passed between VDE participants in administrative object structures 870, and are usually stored in secure database 610. They are always encrypted and authenticated in both of these cases. When a method core 1000′ references a load module 1100, a load module is loaded into the SPE 503, decrypted, and then either passed to the electronic appliance microprocessor for executing in an HPE 655 (if that is where it executes), or kept in the SPE (if that is where it executes). If no SPE 503 is present, the load module may be decrypted by the HPE 655 prior to its execution.
Load module creation by parties is preferably controlled by a certification process or a ring based SPU architecture. Thus, the process of creating new load modules 1100 is itself a controlled process, as is the process of replacing, updating or deleting load modules already stored in a secured database 610.
A load module 1100 is able to perform its function only when executed in the protected environment of an SPE 503 or an HPE 655 because only then can it gain access to the protected elements (e.g., UDEs 1200, other load modules 1100) on which it operates. Initiation of load module execution in this environment is strictly controlled by a combination of access tags, validation tags, encryption keys, digital signatures and/or correlation tags. Thus, a load module 1100 may only be referenced if the caller knows its ID and asserts the shared secret correlation tag specific to that load module. The decrypting SPU may match the identification token and local access tag of a load module after decryption. These techniques make the physical replacement of any load module 1100 detectable at the next physical access of the load module. Furthermore, load modules 1100 may be made “read only” in the preferred embodiment. The read-only nature of load modules 1100 prevents the write-back of load modules that have been tampered with in non-secure space.
Load modules are not necessarily directly governed by PERCs 808 that control them, nor must they contain any time/date information or expiration dates The only control consideration in the preferred embodiment is that one or more methods 1000 reference them using a correlation tag (the value of a protected object created by the load module's owner, distributed (to authorized parties for inclusion in their methods, and to which access and use is controlled by one or more PERCs 808). If a method core 1000′ references a load module 1100 and asserts the proper correlation tag (and the load module satisfies the internal tamper checks for the SPE 503), then that load module can be loaded and executed, or it can be acquired from, shipped to, updated, or deleted by, other systems.
As shown in
The following is an example of a possible field layout for load module public header 802:
Many load modules 1100 contain code that executes in an SPE 503. Some load modules 1100 contain code that executes in an HPE 655. This allows methods 1000 to execute in whichever environment is appropriate. For example, an INFORMATION method 1000 can be built to execute only in SPE 503 secure space for government classes of security, or in an HPE 655 for commercial applications. As described above, the load module public header 802 may contain an “execution space code” field that indicates where the load module 1100 needs to execute. This functionality also allows for different SPE instruction set as well as different user platforms, and allows methods to be constructed without dependencies on the underlying load module instruction set.
Load modules 1100 operate on three major data areas: the stack, load module parameters, and data structures. The stack and execution memory size required to execute the load module 1100 are preferably described in private header 804, as are the data descriptions from the stack image on load module call, return, and any return data areas. The stack and dynamic areas are described using the same DTD mechanism. The following is an example of a possible layout for a load module private header 1104:
Each load module 1100 also may use DTD 1108 information to provide the information necessary to support building methods from a load module. This DTD information contains the definition expressed in a language such as SGML for the names and data types of all, of the method data fields that the load module supports, and the acceptable ranges of values that can be placed in the fields. Other DTDs may describe the function of the load module 1100 in English for inclusion in an electronic contract, for example.
The next section of load module 1100 is an encrypted executable body 1106 that contains one or more blocks of encrypted code. Load modules 1100 are preferably coded in the “native” instruction set of their execution environment for efficiency and compactness. SPU 500 and platform providers may provide versions of the standard load modules 1100 in order to make their products cooperate with the content in distribution mechanisms contemplated by VDE 100. The preferred embodiment creates and uses native mode load modules 1100 in lieu of an interpreted or “p-code” solution to optimize the performance of a limited resource SPU. However, when sufficient SPE (or HPE) resources exist and/or platforms have sufficient resources, these other implementation approaches may improve the cross platform utility of load module code.
The following is an example of a field layout for a load module DTD 1108:
Some examples of how load modules 1100 may use DTDs 1108 include:
Commonly used load modules 1100 may be built into a SPU 500 as space permits. VDE processes that use built-in load modules 1100 will have significantly better performance than processes that have to find, load and decrypt external load modules. The most useful load modules 1100 to build into a SPU might include scaler meters, fixed price billing, budgets and load modules for aggregate methods that perform these three processes.
User Data Elements (UDEs) 1200 and Method Data Elements (MDEs) 1202
User Data Elements (UDEs) 1200 and Method Data Elements (MDEs) 1202 in the preferred embodiment store data. There are many types of UDEs 1200 and MDEs 1202 provided by the preferred embodiment. In the preferred embodiment, each of these different types of data structures shares a common overall format including a common header definition and naming scheme. Other UDEs 1200 that share this common structure include “local name services records” (to be explained shortly) and account information for connecting to other VDE participants. These elements are not necessarily associated with an individual user, and may therefore be considered MDEs 1202. All UDEs 1200 and all MDEs 1202 provided by the preferred embodiment may, if desired, (as shown in
In the preferred embodiment, PERCs 808 and user rights table records are types of UDE 1200. There are many other types of UDEs 1200/MDEs 1202, including for example, meters, meter trails, budgets, budget trails, and audit trails. Different formats for these different types of UDEs/M])Es are defined, as described above, by SGML definitions contained within DTDs 1108. Methods 1000 use these DTDs to appropriately access UDEs/MDEs 1200, 1202.
Secure database 610 stores two types of items: static and dynamic. Static data structures and other items are used for information that is essentially static information. This includes load modules 1100, PERCs 808, and many components of methods. These items are not updated frequently and contain expiration dates that can be used to prevent “old” copies of the information from being substituted for newly received items. These items may be encrypted with a site specific secure database file key when they are stored in the secure database 610, and then decrypted using that key when they are loaded into the SPE.
Dynamic items are used to support secure items that must be updated frequently. The UDEs 1200 of many methods must be updated and written out of the SPE 503 after each use. Meters and budgets are common examples of this. Expiration dates cannot be used effectively to prevent substitution of the previous copy of a budget UDE 1200. To secure these frequently updated items, a transaction tag is generated and included in the encrypted item each time that item is updated. A list of all VDE item IDs and the current transaction tag for each item is maintained as part of the secure database 610.
UDEs 1200 are preferably encrypted using a site specific key once they are loaded into a site. This site-specific key masks a validation tag that may be derived from a cryptographically strong pseudo-random sequence by the SPE 503 and updated each time the record is written back to the secure database 610. This technique provides reasonable assurance that the UDE 1200 has not been tampered with nor substituted when it is requested by the system for the next use.
Meters and budgets are perhaps among the most common data structures in VDE 100. They are used to count and record events, and also to limit events. The data structures for each meter and budget are determined by the content provider or a distributor/redistributor authorized to change the information. Meters and budgets, however, generally have common information stored in a common header format (e.g., user ID, site ID and related identification information).
The content provider or distributor/redistributor may specify data structures for each meter and budget UDE. Although these data structures vary depending upon the particular application, some are more common than others. The following table lists some of the more commonly occurring data structures for METER and BUDGET methods:
The information in the table above is not complete or comprehensive, but rather is intended to show some examples of types of information that may be stored in meter and budget related data structures. The actual structure of particular meters and budgets is determined by one or more DTDs 1108 associated with the load modules 1100 that create and manipulate the data structure. A list of data types permitted by the DTD interpreter 590 in VDE 100 is extensible by properly authorized parties.
The “usage map” concept provided by the preferred embodiment may be tied to the concept of “atomic elements.” In the preferred embodiment, usage of an object 300 may be metered in terms of “atomic elements.” In the preferred embodiment, an “atomic element” in the metering context defines a unit of usage that is “sufficiently significant” to be recorded in a meter. The definition of what constitutes an “atomic element” is determined by the creator of an object 300. For instance, a “byte” of information content contained in an object 300 could be defined as an “atomic element,” or a record of a database could be defined as an “atomic element,” or each chapter of an electronically published book could be defined as an “atomic element.”
An object 300 can have multiple sets of overlapping atomic elements. For example, an access to any database m a plurality of databases may be defined as an “atomic element.” Simultaneously, an access to any record, field of records, sectors of informations, and/or bytes contained in any of the plurality of databases might also be defined as an “atomic element.” In an electronically published newspaper, each hundred words of an article could be defined as an “atomic element,” while articles of more than a certain length could be defined as another set of “atomic elements.” Some portions of a newspaper (e.g., advertisements, the classified section, etc.) might not be mapped into an atomic element.
The preferred embodiment provides an essentially unbounded ability for the object creator to define atomic element types. Such atomic element definitions may be very flexible to accommodate a wide variety of different content usage. Some examples of atomic element types supported by the preferred embodiment include bytes, records, files, sectors, objects, a quantity of bytes, contiguous or relatively contiguous bytes (or other predefined unit types), logically related bytes containing content that has some logical relationship by topic, location or other user specifiable logic of relationship, etc. Content creators preferably may flexibly define other types of atomic elements.
The preferred embodiment of the present invention provides EVENT methods to provide a mapping between usage events and atomic elements. Generally, there may be an EVENT method for each different set of atomic elements defined for an object 300. In many cases, an object 300 will have at least one type of atomic element for metering relating to billing, and at least one other atomic element type for non-billing related metering (e.g., used to, for example; detect fraud, bill advertisers, and/or collect data on end user usage activities).
In the preferred embodiment, each EVENT method in a usage related context performs two functions: (1) it maps an accessed event into a set of zero or more atomic elements, and (2) it provides information to one or more METER methods for metering object usage. The definition used to define this mapping between access events and atomic elements may be in the form of a mathematical definition, a table, a load module, etc. When an EVENT method maps an access request into “zero” atomic elements, a user accessed event is not mapped into any atomic element based on the particular atomic element definition that applies. This can be, for example, the object owner is not interested in metering usage based on such accesses (e.g., because the object owner deems such accesses to be insignificant from a metering standpoint).
A “usage map” may employ a “bit map image” for storage of usage history information in a highly efficient manner. Individual storage elements in a usage map may correspond to atomic elements. Different elements within a usage map may correspond to different atomic elements (e.g., one map element may correspond to number of bytes read, another map element may correspond to whether or not a particular chapter was opened, and yet another map element may correspond to some other usage event.)
One of the characteristics of a usage map provided by the preferred embodiment of the present invention is that the significance of a map element is specified, at least in part, by the position of the element within the usage map. Thus, in a usage map provided by the preferred embodiment, the information indicated or encoded by a map element is a function of its position (either physically or logically) within the map structure. As one simple example, a usage map for a twelve-chapter novel could consist of twelve elements, one for each chapter of the novel. When the user opens the first chapter, one or more bits within the element corresponding to the first chapter could be changed in value (e.g., set to “one”). In this simple example where the owner of the content object containing the novel was interested only in metering which chapters had been opened by the user, the usage map element corresponding to a chapter could be set to “one” the first time the user opened that corresponding chapter, and could remain “one” no matter how many additional times the user opened the chapter. The object owner or other interested VDE participant would be able to rapidly and efficiently tell which chapter(s) had been opened by the user simply by examining the compact usage map to determine which elements were set to “one.”
Suppose that the content object owner wanted to know how many times the user had opened each chapter of the novel. In this case, the usage map might comprise, for a twelve-chapter novel, twelve elements each of which has a one-to-one correspondence with a different one of the twelve chapters of the novel. Each time a user opens a particular chapter, the corresponding METER method might increment the value contained in the corresponding usage map element. In this way, an account could be readily maintained for each of the chapters of the novel.
The position of elements within a usage map may encode a multi-variable function. For example, the elements within a usage map may be arranged in a two-dimensional array as shown in
Usage map meters are thus an efficient means for referencing prior usage. They may be used to enable certain VDE related security functions such as testing for contiguousness (including relative contiguousness), logical relatedness (including relative logical relatedness), usage randomization, and other usage patterns. For example, the degree or character of the “randomness” of content usage by a user might serve as a potential indicator of attempts to circumvent VDE content budget limitations. A user or groups of users might employ multiple sessions to extract content in a manner which does not violate contiguousness, logical relatedness or quantity limitations, but which nevertheless enables reconstruction of a material portion or all of a given, valuable unit of content. Usage maps can be analyzed to determine other patterns of usage for pricing such as, for example, quantity discounting after usage of a certain quantity of any or certain atomic units, or for enabling a user to reaccess an object for which the user previously paid for unlimited accesses (or unlimited accesses over a certain time duration). Other useful analyses might include discounting for a given atomic unit for a plurality of uses.
A further example of a map meter includes storing a record of all applicable atomic elements that the user has paid to use (or alternatively, has been metered as having used, though payment may not yet have been required or made). Such a usage map would support a very efficient and flexible way to allow subsequent user usage of the same atomic elements.
A further usage map could be maintained to detect fraudulent usage of the same object. For example, the object might be stored in such a way that sequential access of long blocks should never occur. A METER method could then record all applicable atomic elements accesses during, for example, my specified increment of time, such as ten minutes, an hour, a day, a month, a year, or other time duration). The usage map could be analyzed at the end of the specified time increment to check for an excessively long contiguous set of accessed blocks, and/or could be analyzed at the initiation of each access to applicable atomic elements. After each time duration based analysis, if no fraudulent use is detected, the usage map could be cleared (or partially cleared) and the mapping process could begin in whole or in part anew. If a fraudulent use pattern is suspected or detected, that information might be recorded and the use of the object could be halted. For example, the user might be required to contact a content provider who might then further analyze the usage information to determine whether or not further access should be permitted.
Audit trail maps may be generated at any frequency determined by control, meter, budget and billing methods and load modules associated with those methods. Audit trails have a similar structure to meters and budgets and they may contain user specific information in addition to information about the usage event that caused them to be created. Like meters and budgets, audit trails have a dynamic format that is defined by the content provider or their authorized designee, and share the basic element types for meters and budgets shown in the table above. In addition to these types, the following table lists some examples of other significant data fields that may be found in audit trails.
Audit trail records may be automatically combined into single records to conserve header space. The combination process may, for example, occur under control of a load module that creates individual audit trail records.
Permissions Record Overview
In the preferred embodiment, no end user may use or access a VDE object unless a permissions record 808 has been delivered to the end user. As discussed above, a PERC 808 may be delivered as part of a traveling object 860 or it may be delivered separately (for example, within an administrative object). An electronic appliance 600 may not access an object unless a corresponding PERC 808 is present, and may only use the object and related information as permitted by the control structures contained within the PERC.
Briefly, the PERC 808 stores information concerning the methods, method options, decryption keys and rights with respect to a corresponding VDE object 300.
PERC 808 includes control structures that define high level categories or classifications of operations. These high level categories are referred to as “rights.” The “right” control structures, in turn, provide internal control structures that reference “methods” 1000. The internal structure of preferred embodiment PERC 808 organizes the “methods” that are required to perform each allowable operation on an object or associated control structure (including operations performed on the PERC itself). For example, PERC 808 contains decryption keys or the object, and usage of the keys is controlled by the methods that are required by the PERC for performing operations associated with the exercise of a “right.”
PERC 808 for an object is typically created when the object is created, and future substantive modifications of a PERC, if allowed, are controlled by methods associated with operations using the distribution right(s) defined by the same (or different) PERC.
There are other elements that may be included in a PERC 808 hierarchy that describe rules and the rule options to support the negotiation of rule sets and control information for smart objects and for the protection of a user's personal information by a privacy filter. These alternate elements may include:
optional rights records
optional control sets
optional method records
permitted rights records
permitted rights control sets
permitted method records
required DTD descriptions
optional DTD descriptions
permitted DTD descriptions
These alternate fields can control other processes that may, in part, base negotiations or decisions regarding their operation on the contents of these fields. Rights negotiation, smart object control information, and related processes can use these fields for more precise control of their operation.
The PERC 808 shown in
Each rights record 906 defines a different “right” corresponding to an object. A “right” record 906 is the highest level of organization present in PERC 808. There can be several different rights in a PERC 808. A “right” represents a major functional partitioning desired by a participant of the basic architecture of VDE 100. For example, the right to use an object and the right to distribute rights to use an object are major functional groupings within VDE 100. Some examples of possible rights include access to content, permission to distribute rights to access content, the ability to read and process audit trails related to content and/or control structures, the right to perform transactions that may or may not be related to content and/or related control structures (such as banking transactions, catalog purchases, the collection of taxes, EDI transactions, and such), and the ability to change some or all of the internal structure of PERCs created for distribution to other users. PERC 808 contains a rights record 906 for each type of right to object access/use the PERC grants.
Normally, for VDE end users, the most frequently granted right is a usage right. Other types of rights include the “extraction right,” the “audit right” for accessing audit trail information, of end users, and a “distribution right” to distribute an object. Each of these different types of rights may be embodied in a different rights record 906 (or alternatively, different PERCs 808 corresponding to an object may be used to grant different rights).
Each rights record 906 includes a rights record header 908, a CSR (“control set for right”) 910, one or more “right ‘keys” 912, and one or more “control sets” 914. Each “rights” record 906 contains one or more control sets 914 that are either required or selectable options to control an object in the exercise of that “right.” Thus, at the next level, inside of a “right” 906, are control sets 914. Control sets 914, in turn, each includes a control set header 916, a control method 918, and one or more required methods records 920. Required methods records 920, in turn, each includes a required method header 922 and one or more required method options 924.
Control sets 914 exist in two types in VDE 100: common required control sets which are given designations “control set 0” or “control set for right,” and a set of control set options. “Control set 0” 902 contains a list of required methods that are common to all control set options, so that the common required methods do not have to be duplicated in each control set option. A “control set for right” (“CSR”) 910 contains a similar list for control sets within a given right. “Control set 0” and any “control sets for rights” are thus, as mentioned above, optimizations; the same functionality for the control sets can be accomplished by listing all the common required methods in each control set option and omitting “control set 0” and any “control sets for rights.”
One of the control set options, “control set 0” and the appropriate “control set for right” together form a complete control set necessary to exercise a right.
Each control set option contains a list of required methods 1000 and represents a different way the right may be exercised. Only one of the possible complete control sets 914 is used at any one time to exercise a right in the preferred embodiment.
Each control set 914 contains as many required methods records 920 as necessary to satisfy all of the requirements of the creators and/or distributors for the exercise of a right. Multiple ways a right may be exercised, or multiple control sets that govern how a given right is exercised, are both supported. As an example, a single control set 914 might require multiple meter and budget methods for reading the object's content, and also require different meter and budget methods for printing an object's content. Both reading and printing an object's content can be controlled in a single control set 914.
Alternatively, two different control set options could support reading an object's content by using one control set option to support metering and budgeting the number of bytes read, and the other control set option to support metering and budgeting the number of paragraphs read. One or the other of these options would be active at a time.
Typically, each control set 914 will reference a set of related methods, and thus different control sets can offer a different set of method options. For example, one control set 914 may represent one distinct kind of metering methodology, and another control set may represent another, entirely different distinct metering methodology.
At the next level inside a control set 914 are the required methods records 920. Methods records 920 contain or reference methods 1000 in the preferred embodiment. Methods 1000 are a collection of “events,” references to load modules associated with these events, static data, and references to a secure database 610 for automatic retrieval of any other separately deliverable data elements that may be required for processing events (e.g., UDEs). A control set 914 contains a list of required methods that must be used to exercise a specific right (i.e., process events associated with a right). A required method record 920 listed in a control set 914 indicates that a method must exist to exercise the right that the control set supports. The required methods may reference “load modules” 1100 to be discussed below. Briefly, load modules 1100 are pieces of executable code that may be used to carry out required methods.
Each control set 914 may have a control method record 918 as one of its required methods. The referenced control method may define the relationships between some or all of the various methods 1000 defined by a control set 906. For example, a control method may indicate which required methods are functionally grouped together to process particular events, and the order for processing the required methods. Thus, a control method may specify that required method referenced by record 920(a)(1)(i) is the first to be called and then its output is to go to required method referenced by record 920(a)(l)(ii) and so on. In this way, a meter method may be tied to one or more billing methods and then the billing methods may be individually tied to different budget methods, etc.
Required method records 920 specify one or more required method options 924. Required method options are the lowest level of control structure in a preferred embodiment PERC 808. By parameterizing the required methods and specifying the required method options 924 independently of the required methods, it becomes possible to reuse required methods in many different circumstances.
For example, a required method record 920 may indicate that an actual budget method ID must be chosen from the list of budget method IDs in the required method option list for that required method. Required method record 920 in this case does not contain any method IDs for information about the type of method required, it only indicates that a method is required. Required method option 924 contains the method ID of the method to be used if this required method option is selected. As a further optimization, an actual method ID may be stored if only one option exists for a specific required method. This allows the size of this data structure to be decreased.
PERC 808 also contains the fundamental decryption keys for an object 300, and any other keys used with “rights” (for encoding and/or decoding audit trails, for example). It may contain the keys for the object content or keys to decrypt portions of the object that contain other keys that then can be used to decrypt the content of the object. Usage of the keys is controlled by the control sets 914 in the same “right” 906 within PERC 808.
In more detail,
Detailed Example of a PERC 808
The PERC 808 shown in
This PERC 808 includes a control set 0 sub-record 914 (0) that may be used commonly by all of rights 906 within the PERC. This control set 0 record 914(0) may include the following fields:
Each required method record 920, in turn may include:
Each method option sub-record 924 may include:
In this example of PERC 808 also includes one or more rights records 906, and an overall check value field 980.
This example of rights record 906 includes:
Referring once again to
In one embodiment, object registry 450 includes the following tables:
an object registration table 460;
a subject table 462;
a User Rights Table (“URT”) 464;
an Administrative Event Log 442;
a shipping table 444; and
a receiving table 446.
Object registry 460 in the example embodiment is a database of information concerning registered VDE objects 300 and the rights of users and user groups with regard to those objects. When electronic appliance 600 receives an object 300 containing a new budget or load module 1100, the electronic appliance usually needs to add the information contained by the object to secure database 610. Moreover, when any new VDE object 300 arrives at an electronic appliance 600, the electronic appliance must “register” the object within object registry 450 so that it can be accessed. The lists and records for a new object 300 are built in the preferred embodiment when the object is “registered” by the electronic appliance 600. The information for the object may be obtained from the object's encrypted private header, object body, and encrypted name services record. This information may be extracted or derived from the object 300 by SPE 503, and, then stored within secure database 610 as encrypted records.
In one embodiment, object registration table 460 includes information identifying objects within object storage (repository) 728. These VDE objects 300 stored within object storage 728 are not, in the example embodiment, necessarily part of secure database 610 since the objects typically incorporate their own security (as necessary and required) and are maintained using different mechanisms than the ones used to maintain the secure database. Even though VDE objects 300 may not strictly be part of secure database 610, object registry 450 (and in particular, object registration table 460) refers to the objects and thus “incorporates them by reference” into the secure database. In the preferred embodiment, an electronic appliance 600 may be disabled from using any VDE object 300 that has not been appropriately registered with a corresponding registration record stored within object registration table 460.
Subject table 462 in the example embodiment establishes correspondence between objects referred to by object registration table 460 and users (or groups of users) of electronic appliance 600.
Subject table 462 provides many of the attributes of an access control list (“ACL”), as will be explained, below.
User rights table 464 in the example embodiment provides permissioning and other information specific to particular users or groups of users and object combinations set forth in subject table 462. In the example embodiment, permissions records 808 (also shown in
Administrative event log 442, shipping table 444, and receiving table 446 provide information about receipts and deliveries of VDE objects 300. These data structures keep track of administrative objects sent or received by electronic appliance 600 including, for example, the purpose and actions of the administrative objects in summary and detailed form. Briefly, shipping table 444 includes a shipping record for each administrative object sent (or scheduled to be sent) by electronic appliance, 600 to another VDE participant. Receiving table 446 in the preferred embodiment includes a receiving record for each administrative object received (or scheduled to be received) by electronic appliance 600. Administrative event log 442 includes an event log record for each shipped and each received administrative object, and may include details concerning each distinct event specified by received administrative objects.
Administrative Object Shipping and Receiving
In the example embodiment of the secure database 610, shipping table header 444A may include a site record number 444A(1), a user (or group) ID 444A(2), a series of reference fields 444A(3)-444A(6), validation tags 444A(7)-444A(8), and a check value field 444A(9). The fields 444A(3)-444A(6) reference certain recent IDs that designate lists of shipping records 445 within shipping table 444. For example, field 444A(3) may reference to a “first” shipping record representing a completed outgoing shipment of an administrative object, and field 444A(4) may reference to a “last” shipping record representing a completed outgoing shipment of an administrative object. In this example, “first” and “last” may, if desired, refer to time or order of shipment as one example. Similarly, fields 444A(5) and 444A(6) may reference to “first” and “last” shipping records for scheduled outgoing shipments. Validation tag 444A(7) may provide validation from a name services record within name services record table 452 associated with the user (group) ID in the header. This permits access from the shipping record back to the name services record that describes the sender of the object described by the shipping records. Validation tag 444A(8) provides validation for a “first” outgoing shipping record referenced by one or more of pointers 444A(3)-444A(6). Other validation tags may be provided for validation of scheduled shipping record(s).
Shipping record 444(1) shown includes a site record number 445(1)(A). It also includes first and last scheduled shipment date/times 445(1)(B), 445(1)(C) providing a window of time used for scheduling administrative object shipments. Field 445(1)(D) may specify an actual date/time of a completed shipment of an administrative object. Field 445(1)(E) provides an ID of an administrative object shipped or to be shipped, and thus identifies which administrative object within object storage 728 pertains to this particular shipping record. A reference field 445(1)(G) references a name services record within name services record table 452 specifying the actual or intended recipient of the administrative object shipped or to be shipped. This information within name services record table 452 may, for example, provide routing information sufficient to permit outgoing administrative objects manager 754 shown in
Each sub-record may include a sub-record length field 442(J)(1)(a), a data area length field 442(J)(1)(b), an event ID field 442(J)(1)(c), a record type field 442(J)(1)(d), a record ID field 442(J)(1)(e), a data area field 442(J)(1)(f), and a check value field 442(J)(1)(g). The data area 442(J)(1)(f) may be used to indicate which information within secure database 610 is affected by the event specified in the event ID field 442(J)(1)(c), or what new secure database item(s) were added, and may also specify the outcome of the event.
The object registration table 460 in the preferred ‘embodiment includes a record corresponding to each VDE object 300 within object storage (repository) 728. When a new object arrives or is detected (e.g., by redirector 684), a preferred embodiment electronic appliance 600 “registers” the object by creating an appropriate object registration record and storing it in the object registration table 460. In the preferred embodiment, the object registration table stores information that is user independent, and depends only on the objects that are registered at a given VDE electronic appliance 600. Registration activities are typically managed by a REGISTER method associated with an object.
In the example, subject table 462 associates users (or groups of users) with registered objects. The example subject table 462 performs the function of an access control list by specifying which users are authorized to access which registered VDE objects 300.
As described above, secure database 610 stores at least one PERC 808 corresponding to each registered VDE object 300. PERCS 808 specify a set of rights that may be exercised to use or access the corresponding VDE object 300. The preferred embodiment allows user to “customize” their access rights by selecting a subset of rights authorized by a corresponding PERC 808 and/or by specifying parameters or choices that correspond to some or all of the rights granted by PERC 808. These user choices are set forth in a user rights table 464 in the preferred embodiment. User rights table (URT) 464 includes URT records, each of which corresponds to a user (or group of users). Each of these URT records specifies user choices for a corresponding VDE object 300. These user choices may, either independently or in combination with a PERC 808, reference one or more methods 1000 for exercising the rights granted to the user by the PERC 808 in a way specified by the choices contained within the URT record.
site record number field 466(1)
object type field 466(2)
creator ID field 466(3)
object ID field 466(4)
a reference field 466(5) that references subject table 462
an attribute field 466(6)
a minimum registration interval field 466(7)
a tag 466(8) to a subject table record, and
a check value field 466(9).
The site record number field 466(1) specifies the site record number for this object registration record 460(N). In one embodiment of secure database 610, each record stored within the secure database is identified by a site record number. This site record number may be used as part of a database lookup process in order to keep track of all of the cords with in the secure database 610.
Object type field 466(2) may specify the type of registered VDE object 300 (e.g., a content object, an administrative object, etc.).
Creator ID field 466(3) in the example may identify the creator of the corresponding VDE object 300.
Object ID field 466(4) in the example uniquely identifies the registered VDE object 300.
Reference field 466(5) in the preferred embodiment identifies a record within the subject table 462. Through use of this reference, electronic appliance 600 may determine all users (or user groups) listed in subject table 462 authorized to access the corresponding VDE object 300. Tag 466(8) is used to validate that the subject table records accessed using field 466(5) is the proper record to be used with the object registration record 460(N).
Attribute field 466(6) may store one or more attributes or attribute flags corresponding to VDE object 300.
Minimum registration interval field 466(7) may specify how often the end user may re-register as a user of the VDE object 300 with a clearinghouse service, VDE administrator, or VDE provider. One reason to prevent frequent re-registration is to foreclose users from reusing budget quantities in traveling objects until a specified amount of time has elapsed. The minimum registration interval field 466(7) maybe left unused when the object owner does not wish to restrict re-registration.
Check value field 466(9) contains validation information used for detecting corruption or modification of record 460(N) to ensure security and integrity of the record. In the preferred embodiment, many or all of the fields within record 460(N) (as with other records within the secure database 610) may be fully or partially encrypted and/or contain fields that are stored redundantly in each record (once in unencrypted form and once in encrypted form). Encrypted and unencrypted versions of the same fields may be cross checked at various times to detect corruption or modification of the records.
As mentioned above, reference field 466(5) references subject table 462, and in particular, references one or more user/object records 460(M) within the subject table.
Subject registration table header 468 in the example includes a site record number field 468(1) that may uniquely identify the header as a record within secure database 610. Header 468 may also include a creator ID field 468(2) that may be a copy of the content of the object registration table creator ID field 466(3). Similarly, subject registration table header 468 may include an object ID field 468(5) that may be a copy of object ID field 466(4) within object registration table 460. These fields 468(2), 468(5) make user/object registration records explicitly correspond to particular VDE objects 300.
Header 468 may also include a tag 468(7) that permits validation. In one example arrangement, the tag 468(7) within the user/object registration header 468 may be the same as the tag 466(8) within the object registration record 460(N) that points to the user/object registration header. Correspondence between these tags 468(7) and 466(8) permits validation that the object registration record and user/object registration header match up.
User/object header 468 also includes an original distributor ID field 468(3) indicating the original distributor of the corresponding VDE object 300, and the last distributor ID field 468(4) that indicates the last distributor within the chain of handling of the object prior to its receipt by electronic appliance 600.
Header 468 also includes a tag 468(8) allowing validation between the header and the “first” subject record 470(1) which field 468(6) references.
Subject record 470(1) includes a site record number 472(1), a user (or user group) ID field 472(2), a user (or user group) attributes field 472(3), a field 472(4) referencing user rights table 464, a field 472(5) that references to the “next” subject record 470(2) (if there is one), a tag 472(6) used to validate with the header tag 468(8), a tag 472(7) used to validate with a corresponding tag in the user rights table record referenced by field 472(4), a tag 472(9) used to validate with a tag in the “next” subject record referenced to by field 472(5) and a check value field 472(9).
User or user group ID 472(2) identifies a user or a user group authorized to use the object identified in field 468(5). Thus, the fields 468(5) and 472(2) together form the heart of the access control list provided by subject table 462. User attributes field 472(3) may specify attributes pertaining to use/access to object 300 by the user or user group specified in fields 472(2). Any number of different users or user groups may be added to the access control list (each with a different set of attributes 472(3)) by providing additional subject records 470 in the “linked list” structure.
Subject record reference field 472(4) references one or more records within user rights table 464.
Rights record header 476 in the preferred embodiment may include site record number field 476(1), a right ID field 476(2), a field 476(3) referencing the “next” rights record 476(2), a field 476(4) referencing a first set of user choice records 478(1), a tag 476(5) to allow validation with URT header tag 474(5), a tag 476(6) to allow validation with a user choice record tag 478(6), and a check value field 476(7). Right ID field 476(2) may, for example, specify the type of right conveyed by the rights record 476 (e.g., right to use, right to distribute, right to read, right to audit, etc.).
The one or more user choice records 478 referenced by rights record header 476 sets forth the user choices corresponding to access and/or use of the corresponding VDE object 300. There will typically be a rights record 476 for each right authorized to the corresponding user or user group. These rights govern use of the VDE object 300 by that user or user group. For instance, the user may have an “access” right, and an “extraction” right, but not a “copy” right. Other rights controlled by rights record 476 (which is derived from PERC 808 using a REGISTER method in the preferred embodiment) include distribution rights, audit rights, and pricing rights. When an object 300 is registered with the electronic appliance 600 and is registered with a particular user or user group, the user may be permitted to select among various usage methods set forth in PERC 808. For instance, a VDE object 300 might have two required meter methodologies: one for billing purposes, and one for accumulating data concerning the promotional materials used by the user. The user might be given the choice of a variety of meter/billing methods, such as: payment by VISA or MasterCard; choosing between billing based upon the quantity of material retrieved from an information database, based on the time of use, and/or both. The user might be offered a discount on time and/or quantity billing if he is willing to allow certain details concerning his retrieval of content to be provided to third parties (e.g., for demographic purposes). At the time of registration of an object and/or user for the object, the user would be asked to select a particular meter methodology as the “active metering method” for the first acquired meter. A VDE distributor might narrow the universe of available choices for the user to a subset of the original selection array stipulated by PERC 808. These user selection and configuration settings are stored within user choice records 480(1), 480(2), 480(N). The user choice records need not be explicitly set forth within user rights table 464; instead, it is possible for user choice records 480 to refer (e.g., by site reference number) to particular VDE methods and/or information parameterizing those methods. Such reference by user choice records 480 to method 1000 should be validated by validation tags contained within the user choice records. Thus, user choice records 480 in the preferred embodiment may select one or more methods 1000 for use with the corresponding VDE object 300 (as is shown in
In one embodiment provided by the present invention, a conventional database engine may be used to store and organize secure database 610, and the encryption layers discussed above may be “on top of” the conventional database structure. However, if such a conventional database engine is unable to organize the records in secure database 610 and support the security considerations outlined above, then electronic appliance 600 may maintain separate indexing structures in encrypted form. These separate indexing structures can be maintained by SPE 503. This embodiment would require SPE 503 to decrypt the index and search decrypted index blocks to find appropriate “site record IDs” or other pointers. SPE 503 might then request the indicated record from the conventional database engine. If the record ID cannot be checked against a record list, SPE 503 might be required to ask for the data file itself so it can retrieve the desired record SPE 503 would then perform appropriate authentication to ensure that the file has not been tampered with and that the proper block is returned. SPE 503 should not simply pass the index to the conventional database engine (unless the database engine is itself secure) since this would allow an incorrect record to be swapped for the requested one.
Updating Secure Database 610
Typically, the end user's electronic appliance 600 may initiate communications with a clearinghouse (Block 1152). This contact may, for example, be established automatically or in response to a user command. It may be initiated across the electronic highway 108, or across other communications networks such as a LAN, WAN, two-way cable or using portable media exchange between electronic appliances. The process of exchanging administrative information need not occur in a single “on line” session, but could instead occur over time based on a number of different one-way and/or two-way communications over the same or different communications means. However, the process 1150 shown in
The end user's electronic appliance 600 generally contacts a particular VDE administrator or clearinghouse. The identity of the particular clearinghouse is based on the VDE object 300 the user wishes to access or has already accessed. For example, suppose the user has already accessed a particular VDE object 300 and has run out of budget for further access. The user could issue a request which will cause her electronic appliance 600 to automatically contact the VDE administrator, distributor and/or financial clearinghouse that has responsibility for that particular object. The identity of the appropriate VDE participants to contact is provided in the example by information within UDEs 1200, MDEs 1202, the Object Registration Table 460 and/or Subject Table 462, for example. Electronic appliance 600 may have to contact multiple VDE participants (e.g., to distribute audit records to one participant, obtain additional budgets or other permissions from another participant, etc.). The contact 1152 may in one example be scheduled in accordance with the
Once contact is established, the end user's electronic appliance and the clearinghouse typically authenticate one another and agree on a session key to use for the real-time information exchange (Block 1154). Once a secure connection is established, the end user's electronic appliance may determine (e.g., based on Shipping Table 444) whether it has any administrative object(s) containing audit information that it is supposed to send to the clearinghouse (decision Block 1156). Audit information pertaining to several VDE objects 300 may be placed within the same administrative object for transmission, or different administrative objects may contain audit information about different objects. Assuming the end user's electronic appliance has at least one such administrative object to send to this particular clearinghouse (“yes” exit to decision Block 1156), the electronic appliance sends that administrative object to the clearinghouse via the now established secure real-time communications (Block 1158). In one specific example, a single administrative object may be sent an administrative object containing audit information pertaining to multiple VDE objects, with the audit information for each different object compromising a separate “event” within the administrative object.
The clearinghouse may receive the administrative object and process its contents to determine whether the contents are “valid” and “legitimate.” For example, the clearinghouse may analyze the contained audit information to determine whether it indicates misuse of the applicable VDE object 300. The clearinghouse may, as a result of this analysis, may generate one or more responsive administrative objects that it then sends to the end user's electronic appliance 600 (Block 1160). The end user's electronic appliance 600 may process events that update its secure database 610 and/or SPU 500 contents based on the administrative object received (Block 1162). For example, if the audit information received by the clearinghouse is legitimate, then the clearinghouse may send an administrative object to the end user's electronic appliance 600 requesting the electronic appliance to delete and/or compress the audit information that has been transferred. Alternatively or in addition, the clearinghouse may request additional information from the end-user electronic appliance 600 at this stage (e.g., retransmission of certain information that was corrupted during the initial transmission, transmission of additional information not earlier transmitted, etc.). If the clearinghouse detects misuse based on the received audit information, it may transmit an administrative object that revokes or otherwise modifies the end user's right to further access the associated VDE objects 300.
The clearinghouse may, in addition or alternatively, send an administrative object to the end user's electronic appliance 600 that instructs the electronic appliance to display one or more messages to the user. These messages may inform the user about certain conditions and/or they may request additional information from the user. For example, the message may instruct the end user to contact the clearinghouse directly by telephone or otherwise to resolve an indicated problem, enter a PIN, or it may instruct the user to contact a new service company to re-register the associated VDE object. Alternatively, the message may tell the end user that she needs to acquire new usage permissions for the object, and may inform the user of cost, status and other associated information.
During the same or different communications exchange, the same or different clearinghouse may handle the end user's request for additional budget and/or permission pertaining to VDE object 300. For example, the end user's electronic appliance 600 may (e.g., in response to a user input request to access a particular VDE object 300) send an administrative object to the clearinghouse requesting budgets and/or other permissions allowing access (Block 1164). As mentioned above, such requests may be transmitted in the form of one or more administrative objects, such as, for example, a single administrative object having multiple “events” associated with multiple requested budgets, and/or other permissions for the same or different VDE objects 300. The clearinghouse may upon receipt of such a request, check the end user's credit, financial records, business agreements and/or audit histories to determine whether the requested budgets and/or permissions should be given. The clearinghouse may, based on this analysis, send one or more responsive administrative objects which cause the end user's electronic appliance 600 to update its secure database in response (Block 1166, 1168). This updating might, for example, comprise replacing an expired PERC 808 with a fresh one, modifying a PERC to provide additional (or lesser) rights, etc. Steps 1164-1168 may be repeated multiple times in the same or different communications session to provide further updates to the end user's secure database 610.
The keys to decrypt secure database 610 records are, in the preferred embodiment, maintained solely within the protected memory of an SPU 500. Each index or record update that leaves the SPU 500 may be time stamped, and then encrypted with a unique key that is determined by the SPE 503. For example, a key identification number may be placed “in plain view” at the front of the records of secure database 610 so the SPE 503 can determine which key to use the next time the record is retrieved SPE 503 can maintain the site ID of the record or index, the key identification number associated with it, and the actual keys in the list internal to the SPE. At some point, this internal list may fill up. At this point, SPE 503 may call a maintenance routine that re-encrypts items within secure database 610 containing changed information. Some or all of the items within the data structure containing changed information may be read in, decrypted, and then re-encrypted with the same key. These items may then be issued the same key identification number. The items may then be written out of SPE 503 back into secure database 610. SPE 503 may then clear the internal list of item IDs and corresponding key identification numbers. It may then begin again the process of assigning a different key and a new key identification number to each new or changed item. By using this process, SPE 503 can protect the data structures (including the indexes) of secure database 610 against substitution of old items and against substitution of indexes for current items. This process also allows SPE 503 to validate retrieved item IDs against the encrypted list of expected IDs.
SPE 503 may generate a new encryption/decryption key for each new item it is going to store within secure database 610 (block 1086). SPE 503 may use this new key to encrypt the record prior to storing it in the secure database (block 1088). SPE 503 make sure that it retains the key so that it can later read and decrypt the record. Such decryption keys are, in the preferred embodiment, maintained within protected non-volatile memory (e.g., NVRAM 534 b) within SPU 500. Since this protected memory has a limited size, there may not be enough room within the protected memory to store a new key. This condition is tested for by decision block 1090 in the preferred embodiment. If there is not enough room in memory for the new key (or some other event such as the number of keys stored in the memory exceeding a predetermined number, a timer has expired, etc), then the preferred embodiment handles the situation by re-encrypting other records with secure database 610 with the same new key in order to reduce the number of (or change) encryption/decryption keys in use. Thus, one or more secure database 610 items may be read from the secure database (block 1092), and decrypted using the old key(s) used to encrypt them the last time they were stored. In the preferred embodiment, one or more “old keys” are selected, and all secure database items encrypted using the old key(s) are read and decrypted. These records may now be re-encrypted using the new key that was generated at block 1086 for the new record (block 1094). The old key(s) used to decrypt the other record(s) may now be removed from the SPU protected memory (block 1096), and the new key stored in its place (block 1097). The old key(s) cannot be removed from secure memory by block 1096 unless SPE 503 is assured that all records within the secure database 610 that were encrypted using the old key(s) have been read by block 1092 and re-encrypted by block 1904 using the new key. All records encrypted (or re-encrypted) using the new key may now be stored in secure database 610 (block 1098). If decision block 1090 determines there is room within the SPU 500 protected memory to store the new key, then the operations of blocks 1092, 1094, 1096 are not needed and SPE 503 may instead simply store the new key within the protected memory (block 1097) and store the new encrypted records into secure database 610 (block 1098).
The security of secure database 610 files may be further improved by segmenting the records into “compartments.” Different encryption/decryption keys may be used to protect different “compartments.” This strategy can be used to limit the amount of information within secure database 610 that is encrypted with a single key. Another technique for increasing security of secure database 610 may be to encrypt different portions of the same records with different keys so that more than one key may be needed to decrypt those records.
Backup of Secure Database 610
Secure database 610 in the preferred embodiment is backed up at periodic or other time intervals to protect the information the secure database contains. This secure database information may be of substantial value to many VDE participants. Back ups of secure database 610 should occur without significant inconvenience to the user, and should not breach any security.
The need to back up secure database 610 may be checked at power on of electronic appliance 600, when SPE 503 is initially invoked, at periodic time intervals, and if “audit roll up” value or other summary services information maintained by SPE 503 exceeds a user set or other threshold, or triggered by criteria established by one or more content publishers and/or distributors and/or clearinghouse service providers and/or users. The user may be prompted to backup if she has failed to do so by or at some certain point in time or after a certain duration of time or quantity of usage, or the backup may proceed automatically without user intervention.
There are at least two scenarios to backing up secure database 610. The first scenario is “site specific,” and uses the security of SPU 500 to support restoration of the backed up information. This first method is used in case of damage to secure database 610 due for example to failure of secondary storage device 652, inadvertent user damage to the files, or other occurrences that may damage or corrupt some or all of secure database 610. This first, site specific scenario of back up assumes that an SPU 500 still functions properly and is available to restore backed up information.
The second back up scenario assumes that the user's SPU 500 is no longer operational and needs to be, or has been, replaced. This second approach permits an authorized VDE administrator or other authorized VDE participant to access the stored back up information in order to prevent loss of critical data and/or assist the user in recovering from the error.
Both of these scenarios are provided by the example of program control steps performed by ROS 602 shown in
The preferred embodiment also reads the summary services audit information stored within the protected memory of SPU 500 by SPE summary services manager 560, encrypts this information with the newly generated back up key(s), and writes this summary services information to back up store 668 (block 1262).
Finally, back up routine 1250 saves the back up key(s) generated block 1252 and used to encrypt in blocks 1256, 1262 onto back up store 668. It does this in two secure ways in order to cover both of the restoration scenarios discussed above. Back up routine 1250 may encrypt the back up key(s) (along with other information such as the time of back up and other appropriate information to identify the back up) with a further key or keys such that only SPU 500 can decrypt (block 1264). This encrypted information is then written to back up store 668 (block 1264). For example, this step may include multiple encryptions using one or more public keys with corresponding private keys known only to SPU 500. Alternatively, a second back up key generated by the SPU 500 and kept only in the SPU may be used for the final encryption in place of a public key. Block 1264 preferably includes multiple encryption in order to make it more difficult to attack the security of the back up by “cracking” the encryption used to protect the back up keys. Although block 1262 includes encrypted summary services information on the back up, it preferably does not include SF15 device private keys, shared keys, SF15 code and other internal security information to prevent this information from ever becoming available to users even in encrypted form.
The information stored by block 1264 is sufficient to allow the same SPU 500 that performed (or at least in part performed) back up routine 1250 to recover the backed up information. However, this information is useless to any device other than that same SPU because only that SPU knows the particular keys used to protect the back up keys. To cover the other possible scenario wherein the SPU 500 fails in a non-recoverable way, back up routine 1250 provides an additional step (block 1266) of saving the back up key(s) under protection of one or more further set of keys that may be read by an authorized VDE administrator. For example, block 1266 may encrypt the back up keys with an “download authorization key” received during initialization of SPU 500 from a VDE administrator. This encrypted version of back up keys is also written to back up store 668 (block 1266). It can be used to support restoration of the back up files in the event of an SPU 500 failure. More specifically, a VDE administrator that knows the download authorization (or other) keys(s) used by block 1266 may be able to recover the back up key(s) in the back up store 668 and proceed to restore the backed up secure database 610 to the same or different electronic appliance 600.
In the preferred embodiment, the information saved by routine 1250 in back up files can be restored only after receiving a back up authorization from an authorized VDE administrator. In most cases, the restoration process will simply be a restoration of secure database 610 with some adjustments to account for any usage since the back up occurred. This may require the user to contact additional providers to transmit audit and billing data and receive new budgets to reflect activity since the last back up. Current summary services information maintained within SPU 500 may be compared to the summary services information stored on the back up to determine or estimate most recent usage activity.
In case of an SPU 500 failure, an authorized VDE administrator must be contacted to both initialize the replacement SPU 500 and to decrypt the back up files. These processes allow for both SPU failures and upgrades to new SPUs. In the case of restoration, the back up files are used to restore the necessary information to the user's system. In the case of upgrades, the back up files may be used to validate the upgrade process.
The back up files may in some instances be used to transfer management information between electronic appliances 600. However, the preferred embodiment may restrict some or all information from being transportable between electronic appliances with appropriate authorizations. Some or all of the back up files may be packaged within an administrative object and transmitted for analysis, transportation, or other uses.
As a more detailed example of a need for restoration from back up files, suppose an electronic appliance 600 suffers a hard disk failure or other accident that wipes out or corrupts part or all of the secure database 610, but assume that the SPU 500 is still functional SPU 500 may include all of the information (e.g., secret keys and the like) it needs to restore the secure database 610. However, ROS 602 may prevent secure database restoration until a restoration authorization is received from a VDE administrator. A restoration authorization may comprise, for example, a “secret value” that must match a value expected by SPE 503. A VDE administrator may, if desired, only provide this restoration authorization after, for example, summary services information stored within SPU 500 is transmitted to the administrator in an administrative object for analysis. In some circumstances, a VDE administrator may require that a copy (partial or complete) of the back up files be transmitted to it within an administrative object to check for indications of fraudulent activities by the user. The restoration process, once authorized, may require adjustment of restored budget records and the like to reflect activity since the last back up, as mentioned above.
The VDE administrator may at this point restore the summary values and other information within SPU 500 based on the information obtained by block 1272 and from the backup (block 1276). For example, the VDE administrator may reset SPU internal summary values and counters so that they are consistent with the last backup. These values may be adjusted by the VDE administrator based on the “work in progress” recovered by block 1272, the amount of time that has passed since the backup, etc. The goal may typically be to attempt to provide internal SPU values that are equal to what they would have been had the failure not occurred.
The VDE administrator may then authorize SPU 500 to recover its secure database 610 from the backup files (block 1278). This restoration process replaces all secure database 610 records with the records from the backup. The VDE administrator may adjust these records as needed by passing commands to SPU 500 during or after the restoration process.
The VDE administrator may then compute bills based on the recovered values (block 1280), and perform other actions to recover from SPU downtime (block 1282). Typically, the goal is to bill the user and adjust other VDE 100 values pertaining to the failed electronic appliance 600 for usage that occurred subsequent to the last backup but prior to the failure. This process may involve the VDE administrator obtaining, from other VDE participants, reports and other information pertaining to usage by the electronic appliance prior to its failure and comparing it to the secure database backup to determine which usage and other events are not yet accounted for.
In one alternate embodiment, SPU 500 may have sufficient internal, non-volatile memory to allow it to store, some or all of secure database 610. In this embodiment, the additional memory may be provided by additional one or more integrated circuits that can be contained within a secure enclosure, such as a tamper resistant metal container or some form of a chip pack containing multiple integrated circuit components, and which impedes and/or evidences tampering attempts, and/or disables a portion or all of SPU 500 or associated critical key and/or other control information in the event of tampering. The same back up routine 1250 shown in
Event-Driven VDE Processes.
As discussed above, processes provided by/under the preferred embodiment rights operating system (ROS) 602 may be “event driven.” This “event driven” capability facilitates integration and extendibility.
An “event” is a happening at a point in time. Some examples of “events” are a user striking a key of a keyboard, arrival of a message or an object 300, expiration of a timer, or a request from another process.
In the preferred embodiment, ROS 602 responds to an “event” by performing a process in response to the event ROS 602 dynamically creates active processes and tasks in response to the occurrence of an event. For example, ROS 602 may create and begin executing one or more component assemblies 690 for performing a process or processes in response to occurrence of an event. The active processes and tasks may terminate once ROS 602 has responded to the event. This ability to dynamically create (and end) tasks in response to events provides great flexibility, and also permits limited execution resources such as those provided by an SPU 500 to perform a virtually unlimited variety of different processes in different contexts.
Since an “event” may be any type of happening, there are an unlimited number of different events. Thus, any attempt to categorize events into different types will necessarily be a generalization. Keeping this in mind, it is possible to categorize events provided/supported by the preferred embodiment into two broad categories:
user-initiated events, and
Generally, “user-initiated” events are happenings attributable to a user (or a user application). A common “user-initiated” event is a user's request (e.g., by pushing a keyboard button, or transparently using redirector 684) to access an object 300 or other VDE-protected information.
“System-initiated” events are generally happenings, not attributable to a user. Examples of system initiated events include the expiration of a timer indicating that information should be backed to non-volatile memory, receipt of a message from another electronic appliance 600, and a service call generated by another process (which may have been started to respond to a system-initiated event and/or a user-initiated event).
ROS 602 provided by the preferred embodiment responds to an event by specifying and beginning processes to process the event. These processes are, in the preferred embodiment, based on methods 1000. Since there are an unlimited number of different types of events, the preferred embodiment supports an unlimited number of different processes to process events. This flexibility is supported by the dynamic creation of component assemblies 690 from independently deliverable modules such as method cores 1000′, load modules 1100, and data structures such as UDEs 1200. Even though any categorization of the unlimited potential types of processes supported/provided by the preferred embodiment will be a generalization, it is possible to generally classify processes as falling within two categories:
processes relating to use of VDE protected information; and
processes relating to VDE administration.
“Use” and “Administrative” Processes
“Use” processes relate in some way to use of VDE-protected information. Methods 1000 provided by the preferred embodiment may provide processes for creating and maintaining a chain of control for use of VDE-protected information. One specific example of a “use” type process is processing to permit a user to open a VDE object 300 and access its contents. A method 1000 may provide detailed use-related processes such as, for example, releasing content to, the user as requested (if permitted), and updating meters, budgets, audit trails, etc. Use-related processes are often user-initiated, but some use processes may be system-initiated. Events that trigger a VDE use-related process may be called “use events.”
An “administrative” process helps to keep VDE 100 working. It provides processing that helps support the transaction management “infrastructure” that keeps VDE 100 running securely and efficiently. Administrative processes may, for example, provide processing relating to some aspect of creating, modifying and/or destroying VDE-protected data structures that establish and maintain VDE's chain of handling and control. For example, “administrative” processes may store, update, modify or destroy information contained within a VDE electronic appliance 600 secure database 610. Administrative processes also may provide communications services that establish, maintain and support secure communications between different VDE electronic appliances 600. Events that trigger administrative processes may be called “administrative events.”
Some VDE processes are paired based on the way they interact together. One VDE process may “request” processing services from another VDE process. The process that requests processing services may be called a “request process.” The “request” constitutes an “event” because it triggers processing by the other VDE process in the pair. The VDE process that responds to the “request event” may be called a “response process.” The “request process” and “response process” may be called “reciprocal processes.”
The “request event” may comprise, for example, a message issued by one VDE node electronic appliance 600 or process for certain information. A corresponding “response process” may respond to the “request event” by, for example, sending the information requested in the message. This response may itself constitute a “request event” if it triggers a further VDE “response process.” For example, receipt of a message in response to an earlier-generated request may trigger a “reply process.” This, “reply process” is a special type of “response process” that is triggered in response to a “reply” from another “response process.” There may be any number of “request” and “response” process pairs within a given VDE transaction.
A “request process” and its paired “response process” may be performed on the same VDE electronic appliance 600, or the two processes may be performed on different VDE electronic appliances. Communication between the two processes in the pair may be by way of a secure (VDE-protected) communication, an “out of channel” communication, or a combination of the two.
Receipt of the request by VDE node 600 b comprises a response event at that node. Upon receipt of the request, the VDE node 600 b may perform a “reciprocal” process 1454 defined by the same or different method 1000 b to respond to the response event. The reciprocal process 1454 may be based on a component assembly 690 (e.g., one or more load modules 1100, data, and optionally other methods present in the VDE node 600B).
A method 1000 may be parameterized with sets of events that specify related or cooperative functions. Events may be logically grouped by function (e.g., use, distribute), or a set of reciprocal events that specify processes that may operate in conjunction with each other.
Control of event processing, reciprocal events, and their associated methods and method components is provided by PERCs 808 in the preferred embodiment. These PERCs (808) might, reference administrative methods that govern the creation, modification, and distribution of the data structures and administrative methods that permit access, modification, and, further distribution of these items. In this way, each link in the chain of handling and control might, for example, be able to customize audit information, alter the budget requirements for using the content, and/or control further distribution of these rights in a manner specified by prior members along the distribution chain.
In the example shown in
After registering to use the content object, the user 112 would be required to utilize an array of “use” processes 1476C to, for example, open, read, write, and/or close the content object as part of the use process.
Once the distributor 106 has used some or all of her budget, she may desire to obtain additional budget. The distributor 106 might then initiate a process using the BUDGET method request process (1480B). Request process 1480B might initiate a communication (1482AB) with the content creator VDE node 102 requesting more budget and perhaps providing details of the use activity to date (e.g., audit trails). The content creator 102 processes the ‘get more budget’ request event 1482AB using the response process (1484A) within the creator's BUDGET method 1510A. Response process 1484A might, for example, make a determination if the use information indicates proper use of the content, and/or if the distributor is credit worthy for more budget. The BUDGET method response process 1484A might also initiate a financial transaction to transfer funds from the distributor to pay for said use, or use the distribute process 1472A to distribute budget to the distributor 106. A response to the distributor 106 granting more budget (or denying more budget) might be sent immediately as a response to the request communication 1482AB or it might be sent at a later time as part of a separate, communication. The response communication, upon being received at the distributor's VDE node 106, might be processed using the reply process 1475B within the distributor's copy of the BUDGET method 1510B The reply process 1475B might then process the additional budget in the same manner as described above.
The chain of handling and control may, in addition to posting budget information, also pass control information that governs the manner in which said budget maybe utilized. For example, the control information specified in the above example may also contain control information describing the process and limits that apply to the distributor's redistribution of the right to use the creator's content object. Thus, when the distributor responds to a budget request from a user (a communication between a user at VDE node 112 to the distributor at VDE node 106 similar in nature to the one described above between VDE nodes 106 and 102) using the distribute process 1472B within the distributor's copy of the BUDGET method 1510B, a distribution and request/response/reply process similar to the one described above might be initiated.
Thus, in this example, a single method can provide multiple dynamic behaviors based on different “triggering” events. For example, single BUDGET method 1510 might support any or all of the events listed below:
Examples of Reciprocal Method Processes
a use (see
administrative request (see
administrative response (see
administrative reply (see
In general, the “use” mode of BUDGET method 2250 is invoked in response to an event relating to the use of an object or its content. The “administrative request” mode of BUDGET method 2250 is invoked by or on behalf of the user in response to some user action that requires contact with a VDE financial provider, and basically its task is to send an administrative request to the VDE financial provider. The “administrative response” mode of BUDGET method 2250 is performed at the VDE financial provider in response to receipt of an administrative request sent from a VDE node to the VDE financial provider by the “administrative request” invocation of BUDGET method 2250 shown in
In the preferred embodiment, the same BUDGET method 2250 performs each of the four different step sequences shown in
Block 2296 may then communicate the administrative object to a VDE financial provider, or alternatively, block 2296 may pass administrative object to a separate communications process or method that arranges for such communications to occur. If desired, method 2250 may then save a communications audit trail (blocks 2300, 2302) before terminating (at termination point 2304).
Upon receiving the administrative object, BUDGET method 250 at the VDE financial provider site may prime a budget communications and response audit trail (blocks 2306, 2308), and may then unpack the administrative object and retrieve the budget request(s), audit trail(s) and record(s) it contains (block 2310). This information retrieved from the administrative object may be written by the VDE financial provider into its secure database (block 2312). The VDE financial provider may then retrieve the budget request(s) and determine the response method it needs to execute to process the request (blocks 2314, 2316). BUDGET method 2250 may send the event(s) contained in the request record(s) to the appropriate response method and may generate response records and response requests based on the RESPONSE method (block 2318). The process performed by block 2318 may satisfy the budget request by writing appropriate new response records into the VDE financial provider's secure database (block 2320). BUDGET method 2250 may then write these Budget administrative response records into an administrative object (blocks 2322, 2324), which it may then communicate back to the user node that initiated the budget request. BUDGET method 2250 may then save communications and response processing audit trail information into appropriate audit trail UDE(s) (blocks 2326, 2328) before terminating (at termination point 2330).
Sometime later, the VDE user node may retrieve the reply record from the secure database and determine what method is required to process it (blocks 2344,2346). The VDE user node may, optionally, prime an audit trail (blocks 2342, 2343) to record the results of the processing of the reply event. The BUDGET method 2250 may then send event(s) contained in the reply record(s) to the REPLY method, and may generate/update the secure database records as necessary to, for example, insert new budget records, delete old budget records and/or apply changes to budget records (blocks 2348, 2350). BUDGET method 2250 may then delete the reply record from the secure data base (blocks 2352, 2353) before writing the audit trail (if required) (blocks 2354 m 2355) terminating (at terminate point 2356).
The steps shown in
The AUDIT method 2520 operating in the “administrative request” mode as shown in
For example, AUDIT method 2520 at this point could call an external process to perform, for example, an electronic funds transfer against the user's bank account or some other bank account. The AUDIT administrative response can, if desired, call an external process that interfaces VDE to one or more existing computer systems. The external process could be passed the user's account number, PIN, dollar amount, or any other information configured in, or associated with, the VDE audit trail being processed. The external process can communicate with non-VDE hosts and use the information passed to it as part of these communications. For example, the external process could generate automated clearinghouse (ACH) records in a file for submittal to a bank. This mechanism would provide the ability to automatically credit or debit a bank account in any financial institution. The same mechanism could be used to communicate with the existing credit card (e.g. VISA) network by submitting VDE based charges against the charge account.
Once the appropriate Audit response record(s) have been generated, AUDIT method 2520 may write an Audit administrative record(s) into an administrative object for communication back to the VDE user node that generated the Audit request (blocks 2566, 2568). The AUDIT method 2520 may then save communications and response processing audit information in appropriate audit trail(s) (blocks 2570, 2572) before terminating (at terminate point 2574).
Examples of Event-Driven Content-Based Methods
VDE methods 1000 are designed to provide a very flexible and highly modular approach to secure processing. A complete VDE process to service a “use event” may typically be constructed as a combination of methods 1000. As one example, the typical process for reading content or other information from an object 300 may involve the following methods:
an EVENT method
a METER method
a BILLING method
a BUDGET method.
EVENT method 402 filters out events that are not significant with regard to the specific control method involved. EVENT method 402 may pass on qualified events to a METER process 1404, which meters or discards the event based on its own particular criteria.
In addition, the preferred embodiment provides an optimization called “precheck.” EVENT method/process 402 may perform this “precheck” based on metering, billing and budget information to determine whether processing based on an event will be allowed. Suppose, for example, that the user has already exceeded her budget with respect to accessing certain information content so that no further access is permitted. Although BUDGET method 408 could make this determination, records and processes performed by BUDGET method 404 and/or BILLING method 406 might have to be “undone” to, for example, prevent the user from being charged for an access that was actually denied. It may be more efficient to perform a “precheck” within EVENT method 402 so that fewer transactions have to be “undone.”
METER method 404 may store an audit record in a meter “trail” UDE 1200, for example, and may also record information related to the event in a meter UDE 1200. For example, METER method 404 may increment or decrement a “meter” value within a meter UDE 1200 each time content is accessed. The two different data structures (meter UDE and meter trail UDE) may be maintained to permit record keeping for reporting purposes to be maintained separately from record keeping for internal operation purposes, for example.
Once the event is metered by METER method 404, the metered event may be processed by a BILLING method 406. BILLING method 406 determines how much budget is consumed by the event, and keeps records that are useful for reconciliation of meters and budgets. Thus, for example, BILLING method 406 may read budget information from a budget UDE, record billing information in a billing UDE, and write one or more audit records in a billing trail UDE. While some billing trail information may duplicate meter and/or budget trail information, the billing trail information is useful, for example, to allow a content creator 102 to expect a payment of a certain size, and serve as a reconciliation check to reconcile meter trail information sent to creator 102 with budget trail information sent to, for example, an independent budget provider.
BILLING method 406 may then pass the event on to a BUDGET method 408. BUDGET method 408 sets limits and records transactional information associated with those limits. For example, BUDGET method 408 may store budget information in a budget UDE, and may store an audit record in a budget trail UDE. BUDGET method 408 may result in a “budget remaining” field in a budget UDE being decremented by an amount specified by BILLING method 406.
The information content may be released, or other action taken, once the various methods 402, 404, 406, 408 have processed the event.
As mentioned above, PERCs 808 in the preferred embodiment may be provided with “control methods” that in effect “oversee” performance of the other required methods in a control process.
Control methods operate at the level of control sets 906 within PERCs 808. They provide structure, logic, and flow of control between disparate acquired methods 1000. This mechanism permits the content provider to create any desired chain of processing, and also allows the specific chain of processing to be modified (within permitted limits) by downstream/redistributors. This control structure concept provides great flexibility.
Many different methods can be in effect simultaneously.
Representative Examples of VDE Methods
Although methods 1000 can have virtually unlimited variety and some may even be user-defined, certain basic “use” type methods are preferably used in the preferred embodiment to control most of the more fundamental object manipulation and other functions provided by VDE 100. For example, the following high level methods would typically be provided for object manipulation:
An OPEN method is used to control opening a container so its contents may be accessed. A READ method is used to control the access to contents in a container. A WRITE method is used to control the insertion of contents into a container. A CLOSE method is used to close a container that has been opened.
Subsidiary methods are provided to perform some of the steps required by the OPEN, READ, WRITE and/or CLOSE methods. Such subsidiary methods may include the following
DESTROY content method
An ACCESS method may be used to physically access content associated with an opened container (the content can be anywhere). A PANIC method may be used to disable at least a portion of the VDE node if a security violation is detected. An ERROR method may be used to handle error conditions A DECRYPT method is used to decrypt encrypted information. An ENCRYPT method is used to encrypt information. A DESTROY content method is used to destroy the ability to access specific content within a container. An INFORMATION method is used to provide public information about the contents of a container. An OBSCURE method is used to devalue content read from an opened container (e.g., to write the word “SAMPLE” over a displayed image). A FINGERPRINT method is used to mark content to show who has released it from the secure container. An event method is used to convert events into different events for response by other methods.
The OPEN method process starts with an “open event.” This open event may be generated by a user application, an operating system intercept or various other mechanisms for capturing or intercepting control. For example, a user application may issue a request for access to a particular content stored within the VDE container. As another example, another method may issue a command.
In the example shown, the open event is processed by a control method 1502. Control method 1502 may call other methods to process the event. For example, control method 1502 may call an EVENT method 1504, a METER method 1506, a BILLING method 1508, and a BUDGET method 1510. Not all OPEN control methods necessarily call of these additional methods, but the OPEN method 1500 shown in
Control method 1502 passes a description of the open event to EVENT method 1504. EVENT method 1504 may determine, for example, whether the open event is permitted and whether the open event is significant in the sense that it needs to be processed by METER method 1506, BILLING method 1508, and/or BUDGET method 1510. EVENT method 1504 may maintain audit trail information within an audit trail UDE, and may determine permissions and significance of the event by using an Event Method Data Element (MDE). EVENT method 1504 may also map the open event into an “atomic element” and count that may be processed by METER method 1506, BILLING method 1508, and/or BUDGET method 1510.
In OPEN method 1500, once EVENT method 1504 has been called and returns successfully, control method 1502 then may call METER method 1506 and pass the METER method, the atomic element and count returned by EVENT method 1504. METER method 1506 may maintain audit trail information in a METER method Audit Trail UDE, and may also maintain meter information in a METER method UDE. In the preferred embodiment, METER method 1506 returns a meter value to control method 1502 assuming successful completion.
In the preferred embodiment, control method 1502 upon receiving an indication that METER method 1506 has completed successfully, then calls BILLING method 1508. Control method 1502 may pass to BILLING method 1508 the meter value provided by METER method 1506. BILLING method 1508 may read and update billing information maintained in a BILLING method map MDE, and may also maintain and update audit trail in a BILLING method Audit Trail UDE. BILLING method 1508 may return a billing amount and a completion code to control method 1502.
Assuming BILLING method 1508 completes successfully, control method 1502 may pass the billing value provided by BILLING method 1508 to BUDGET method 1510. BUDGET method 1510 may read and update budget information within a BUDGET method UDE, and may also maintain audit trail information in a BUDGET method Audit Trail UDE. BUDGET method 1510 may return a budget value to control method 1502, and may also return a completion code indicating whether the open event exceeds the user's budget (for this type of event).
Upon completion of BUDGET method 1510, control method 1502 may create a channel and establish read/use control information in preparation for subsequent calls to the READ method.
Assuming the proper URT for this user and object is present such that the object is registered for this user (“yes” exit to decision block 1522), control method 1502 may determine whether the object is already open for this user (decision block 1528). This test may avoid creating a redundant channel for opening an object that is already open. Assuming the object is not already open (“no” exit to decision block 1528), control method 1502 creates a channel and binds appropriate open control elements to it (block 1530). It reads the appropriate open control elements from the secure database (or the container, such as, for example, in the case of a travelling object), and “binds” or “links” these particular appropriate control elements together in order to control opening of the object for this user. Thus, block 1530 associates an event with one or more appropriate method core(s), appropriate load modules, appropriate User Data Elements, and appropriate Method Data Elements read from the secure database (or the container) (block 1532). At this point, control method 1502 specifies the open event (which started the OPEN method to begin with), the object ID and user ID (determined by block 1520), and the channel ID of the channel created by block 1530 to subsequent EVENT method 1504, METER method 1506, BILLING method 1508 and BUDGET method 1510 to provide a secure database “transaction” (block 1536). Before doing so, control method 1502 may prime an audit process (block 1533) and write audit information into an audit UDE (block 1534) so a record of the transaction exists even if the transaction fails or is interfered with.
The detail steps performed by EVENT method 1504 are set forth on
Control method 1502 tests the completion code returned by EVENT method 1504 to determine whether it failed or was successful (decision block 1552). If the EVENT method failed (“no” exit to decision block 1552), control method 1502 may “roll back” the secure database transaction (block 1554) and return itself with an indication that the OPEN method failed (block 1556). In this context, “rolling back” the secure database transaction means, for example, “undoing” the changes made to audit trail UDE by blocks 1540, 1548. However, this “roll back” performed by block 1554 in the preferred embodiment does not “undo” the changes made to the control method audit UDE by blocks 1532, 1534.
Assuming the EVENT method 1504 completed successfully, control method 1502 then calls the METER method 1506 shown on
Control method 1502 tests whether the METER method succeeded by examining the completion code, for example (decision block 1572). If the METER method failed (“no” exit to decision block 1572), then control method 1502 “rolls back” a secure database transaction (block 1574), and returns with an indication that the OPEN method failed (block 1576). Assuming the METER method succeeded (“yes” exit to decision block 1572), control method 1502 calls the BILLING method 1508 and passes it the meter value provided by METER method 1506.
An example of steps performed by BILLING method 1508 is set forth in
Other BILLING methods may use site, user and/or usage information to establish, for example, pricing information. For example, information concerning the presence or absence of an object may be used in establishing “suite” purchases, competitive discounts, etc. Usage levels may be factored into a BILLING method to establish price breaks for different levels of usage. A currency translation feature of a BILLING method may allow purchases and/or pricing in many different currencies. Many other possibilities exist for determining an amount of budget consumed by an event that may be incorporated into BILLING methods.
An example of detailed control steps performed by BUDGET method 1510 is set forth in
Control method 1502 then, in the preferred embodiment, tests whether the read channel was established successfully (decision block 1626). If the read channel was not successfully established (“no” exit to decision block 1626), control method 1502 “rolls back” the secured database transaction and provides an indication that the OPEN method failed (blocks 1628, 1630). Assuming the read channel was successfully established (“yes” exit to decision block 1626), control method 1502 may “commit” the secure database transaction (block 1632). This step of “committing” the secure database transaction in the preferred embodiment involves, for example, deleting intermediate values associated with the secure transaction that has just been performed and, in one example, writing changed UDEs and MDEs to the secure database. It is generally not possible to “roll back” a secure transaction once it has been committed by block 1632. Then, control method 1502 may “tear down” the channel for open processing (block 1634) before terminating (block 1636). In some arrangements, such as multi-tasking VDE node environments, the open channel may be constantly maintained and available for use by any OPEN method that starts. In other implementations, the channel for open processing may be rebuilt and restarted each time an OPEN method starts.
Block 1950 (“map event to atomic element # and event count using a Map MDE”) is in some sense the “heart” of EVENT method 1940. This step “maps” the event into an “atomic element number” to be responded to by subsequently called methods. An example of process control steps performed by a somewhat representative example of this “mapping” step 1950 is shown in
EVENT method mapping process 1950 may first look up the event code (READ) in the EVENT method MDE (1952) using the EVENT method map DTD (1948) to determine the structure and contents of the MDE. A test might then be performed to determine if the event code was found in the MDE (1956), and if not (“No” branch), the EVENT method mapping process may the terminate (1958) without mapping the event to an atomic element number and count. If the event was found in the MDE (“Yes” branch), the EVENT method mapping process may then compare the event range (e.g., bytes 1001-1500) against the atomic element to event range mapping table stored in the MDE (block 1960). The comparison might yield one or more atomic element numbers or the event range might not be found in the mapping table. The result of the comparison might then be tested (block 1962) to determine if any atomic element numbers were found in the table. If not (“No” branch), the EVENT method mapping process may terminate without selecting any atomic element numbers or counts (1964). If the atomic element numbers were found, the process might then calculate the atomic element count from the event range (1966). In this example, the process might calculate the number of bytes requested by subtracting the upper byte range from the lower byte range (e.g., 1500−1001+1=500). The example EVENT method mapping process might then terminate (block 1968) and return the atomic element number(s) and counts.
EVENT method 1940 may then write an EVENT audit trail if required to an EVENT method Audit Trail UDE (block 1970, 1972). EVENT method 1940 may then prepare to pass the atomic element number and event count to the calling CONTROL method (or other control process) (at exit point 1978). Before that, however, EVENT method 1940 may test whether an atomic element was selected (decision block 1974). If no atomic element was selected, then the EVENT method may be failed (block 1974). This may occur for a number of reasons. For example, the EVENT method may fail to map an event into an atomic element if the user is not authorized to access the specific areas of content that the EVENT method MDE does not describe. This mechanism could be used, for example, to distribute customized versions of a piece of content and control access to the various versions in the content object by altering the EVENT method MDE delivered to the user. A specific use of this technology might be to control the distribution of different language (e.g., English, French, Spanish) versions of a piece of content.
The BILLING method map MDE in this example may describe the pricing algorithm that should be used in this BILLING method (e.g., bill $0.001 per byte of content released). Block 1988 (“Map meter value to billing amount”) functions in the same manner as block 1950 of the EVENT method; it maps the meter value to a billing value. Process step 1988 may also interrogate the secure database (as limited by the privacy filter) to determine if other objects or information (e.g., user information) are present as part of the BILLING method algorithm.
BILLING method 1980 may then write a BILLING audit trail if required to a BILLING method Audit Trail UDE (block 1990, 1992), and may prepare to return the billing amount to the calling CONTROL method (or other control process). Before that, however, BILLING method 1980 may test whether a billing amount was determined (decision block 1994). If no billing amount was determined, then the BILLING method may be failed (block 1996). This may occur if the user is not authorized to access the specific areas of the pricing table that the BILLING method MDE describes (e.g., you may purchase not more than $100.00 of information from this content object).
ACCESS method 2000 may first prime an ACCESS audit trail (if required) by writing to an ACCESS Audit Trail UDE (blocks 2002, 2004). ACCESS method 2000 may then read and load an ACCESS method DTD in order to determine the format of an ACCESS MDE (blocks 2006, 2008). The ACCESS method MDE specifies the source and routing information for the particular object to be accessed in the preferred embodiment. Using the ACCESS method DTD, ACCESS method 2000 may load the correction parameters (e.g., by telephone number, account ID, password and/or a request script in the remote resource dependent language).
ACCESS method 2000 reads the ACCESS method MDE from the secure database, reads it in accordance with the ACCESS method DTD, and loads encrypted content source and routing information based on the MDE (blocks 2010, 2012). This source and routing information specifies the location of the encrypted content. ACCESS method 2000 then determines whether a connection to the content is available (decision block 2014). This “connection” could be, for example, an on-line connection to a remote site, a real-time information feed, or a path to a secure/protected resource, for example. If the connection to the content is not currently available (“No” exit of decision block 2014), then ACCESS method 2000 takes steps to open the connection (block 2016). If the connection fails (e.g., because the user is not authorized to access a protected secure resource), then the ACCESS method 2000 returns with a failure indication (termination point 2018). If the open connection succeeds, on the other hand, then ACCESS method 2000 obtains the encrypted content (block 2020). ACCESS method 2000 then writes an ACCESS audit trail if required to the secure database ACCESS method Audit Trail UDE (blocks 2022, 2024), and then terminates (terminate point 2026).
Decrypt and Encrypt
CONTENT method 2070 begins by determining whether the derived content to be provided must be derived from secure contents, or whether it is already available in the object in the form of static values (decision block 2070). Some objects may, for example, contain prestored abstracts, indexes, tables of contents, etc., provided expressly for the purpose of being extracted by the CONTENT method 2070. If the object contains such static values (“static” exit to decision block 2072), then CONTENT method 2070 may simply read this static value content information from the object (block 2074), optionally decrypt, and release this content description (block 2076). If, on the other hand, CONTENT method 2070 must derive the synopsis/content description from the secure object (“derived” exit to decision block 2072), then the CONTENT method may then securely read information from the container according to a synopsis algorithm to produce the synopsis (block 2078).
Extract and Embed
EXTRACT method 2080 begins by priming an Audit UDE (blocks 2082, 2084). EXTRACT method then calls a BUDGET method to make sure that the user has enough budget for (and is authorized to) extract content from the original object (block 2086). If the user's budget does not permit the extraction (“no” exit to decision block 2088), then EXTRACT method 2080 may write a failure audit record (block 2090), and terminate (termination point 2092). If the user's budget permits the extraction (“yes” exit to decision block 2088), then the EXTRACT method 2080 creates a copy of the extracted object with specified rules and control information (block 2094). In the preferred embodiment, this step involves calling a method that actually controls the copy. This step may or may not involve decryption and encryption, depending on the particular the PERC 808 associated with the original object, for example. EXTRACT method 2080 then checks whether any control changes are permitted by the rights authorizing the extract to begin with (decision block 2096). In some cases, the extract rights require an exact copy of the PERC 808 associated with the original object (or a PERC included for this purpose) to be placed in the new (destination) container (“no” exit to decision block 2096). If no control changes are permitted then extract method 2080 may simply write audit information to the Audit UDE (blocks 2098, 2100) before terminating (terminate point 2102). If, on the other hand, the extract rights permit the user to make control changes (“yes” to decision block 2096), then EXTRACT method 2080 may call a method or load module that solicits new or changed control information (e.g., from the user, the distributor who created/granted extract rights, or from some other source) from the user (blocks 2104, 2106). EXTRACT method 2080 may then call a method or load module to create a new PERC that reflects these user-specified-control information (block 2104). This new PERC is then placed in the new (destination) object, the auditing steps are performed, and the process terminates.
OBSCURE method 2140 first calls an EVENT method to determine if the content is appropriate and in the range to be obscured (block 2142). If the content is not appropriate for obscuring, the OBSCURE method terminates (decision block 2144 “no” exit, terminate point 2146). Assuming that the content is to be obscured (“yes” exit to decision block 2144), then OBSCURE method 2140 determines whether it has previously been called to obscure this content (decision block 2148). Assuming the OBSCURE 2140 has not previously called for this object/content. (“yes” exit to decision block 2148), the OBSCURE method 2140 reads an appropriate OBSCURE method MDE from the secure database and loads an obscure formula and/or pattern from the MDE (blocks 2150, 2152). The OBSCURE method 2140 may then apply the appropriate obscure transform based on the patters and/or formulas loaded by block 2150 (block 2154). The OBSCURE method then may terminate (terminate block 2156).
Such fingerprints 2161 can be “buried”—that is inserted in a manner that hides the fingerprints from typical users, sophisticated “hackers,” and/or from all users, depending on the file format, the sophistication and/or variety of the insertion algorithms, and on the availability of original, non-fingerprinted content (for comparison for reverse engineering of algorithm(s)). Inserted or embedded fingerprints 2161, in a preferred embodiment, may be at least in part encrypted to make them more secure. Such encrypted fingerprints 2161 may be embedded within released content provided in “clear” (plaintext) form.
Fingerprints 2161 can be used for a variety of purposes including, for example, the often related purposes of proving misuse of released materials and proving the source of released content. Software piracy is a particularly good example where fingerprinting can be very useful. Fingerprinting can also help to enforce content providers' rights for most types of electronically delivered information including movies, audio recordings, multimedia, information databases, and traditional “literary” materials. Fingerprinting is a desirable alternative or addition to copy protection.
Most piracy of software applications, for example, occurs not with the making of an illicit copy by an individual for use on another of the individual's own computers, but rather in giving a copy to another party. This often starts a chain (or more accurately a pyramid) of illegal copies, as copies are handed from individual to individual. The fear of identification resulting from the embedding of a fingerprint 2161 will likely dissuade most individuals from participating, as many currently do, in widespread, “casual” piracy. In some cases, content may be checked for the presence of a fingerprint by a fingerprint method to help enforce content providers' rights.
Different fingerprints 2161 can have different levels of security (e.g., one fingerprint 2161 (1) could be readable/identifiable by commercial concerns, while another fingerprint 2161(2) could be readable only by a more trusted agency. The methods for generating the more secure fingerprint 2161 might employ more complex encryption techniques (e.g., digital signatures) and/or obscuring of location methodologies. Two or more fingerprints 2161 can be embedded in different locations and/or using different techniques to help protect fingerprinted information against hackers. The more secure fingerprints might only be employed periodically rather than each time content release occurs, if the technique used to provide a more secure fingerprint involves an undesired amount of additional overhead. This may nevertheless be effective since a principal objective of fingerprinting is deterrence—that is the fear on the part of the creator of an illicit copy that the copying will be found out.
For example, one might embed a copy of a fingerprint 2161 which might be readily identified by an authorized party—for example a distributor, service personal, client administrator, or clearinghouse using a VDE electronic appliance 600. One might embed one or more additional copies or variants of a fingerprint 2161 (e.g., fingerprints carrying information describing some or all relevant identifying information) and this additional one or more fingerprints 2161 might be maintained in a more secure manner.
Fingerprinting can also protect privacy concerns. For example, the algorithm and/or mechanisms needed to identify the fingerprint 2161 might be available only through a particularly trusted agent.
Fingerprinting 2161 can take many forms. For example, in an image, the color of every N pixels (spread across an image, or spread across a subset of the image) might be subtly shifted in a visually unnoticeable manner (at least according to the normal, unaided observer). These shifts could be interpreted by analysis of the image (with or without access to the original image), with each occurrence or lack of occurrence of a shift in color (or greyscale) being one or more binary “on or off” bits for digital information storage. The N pixels might be either consistent, or alternatively, pseudo-random in order (but interpretable, at least in part, by a object creator, object provider, client administrator, and/or VDE administrator).
Other modifications of an image (or moving image, audio etc.) which provide a similar benefit (that is, storing information in a form that is not normally noticeable as a result of a certain modification of the source information) may be appropriate depending on the application. For example, certain subtle modifications in the frequency of stored audio information can be modified so as to be normally unnoticeable to the listener while still being readable with the proper tools. Certain properties of the storage of information might be modified to provide such slight but interpretable variations in polarity of certain information which is optically stored to achieve similar results. Other variations employing other electronic, magnetic, and/or optical characteristic may be employed.
Content stored in files that employ graphical formats, such as Microsoft Windows word processing files, provide significant opportunities for “burying” a fingerprint 2161. Content that includes images and/or audio provides the opportunity to embed fingerprints 2161 that may be difficult for unauthorized individuals to identify since, in the absence of an “unfingerprinted” original for purposes of comparison, minor subtle variations at one or more time instances in audio frequencies, or in one or more video images, or the like, will be in themselves undiscernible given the normally unknown nature of the original and the large amounts of data employed in both image and sound data (and which is not particularly sensitive to minor variations). With formatted text documents, particularly those created with graphical word processors (such as Microsoft Windows or Apple Macintosh word processors and their DOS and Unix equivalents), fingerprints 216 can normally be inserted unobtrusively into portions of the document data representation that are not normally visible to the end user (such as in a header or other non-displayed data field).
Yet another form of fingerprinting, which may be particularly suitable for certain textual documents, would employ and control the formation of characters for a given font. Individual characters may have a slightly different visual formation which connotes certain “fingerprint” information. This alteration of a given character's form would be generally undiscernible, in part because so many slight variations exist in versions of the same font available from different suppliers, and in part because of the smallness of the variation. For example, in a preferred embodiment, a program such as Adobe Type Align could be used which, in its off-the-shelf versions, supports the ability of a user to modify font characters in a variety of ways. The mathematical definition of the font character is modified according to the user's instructions to produce a specific set of modifications to be applied to a character or font. Information content could be used in an analogous manner (as an alternative to user selections) to modify certain or all characters too subtly for user recognition under normal circumstances but which nevertheless provide appropriate encoding for the fingerprint 2161. Various subtly different versions of a given character might be used within a single document so as to increase the ability to carry transaction related font fingerprinted information.
Some other examples of applications for fingerprinting might include:
Fingerprinting method 2160 is typically performed (if at all) at the point at which content is released from a content object 300. However, it could also be performed upon distribution of an object to “mark” the content while still in encrypted form. For example, a network-based object repository could embed fingerprints 2161 into the content of an object before transmitting the object to the requester, the fingerprint information could identify a content requester/end user. This could help detect “spoof” electronic appliances 600 used to release content without authorization.
Assuming the Meter UDE is not yet expired, the meter method 2220 may update it using the atomic element and event count passed to the METER method from, for example, an EVENT method (blocks 2239, 2240). The METER method 2220 may then save the Meter Use Audit Record in the Meter Audit Trail UDE (blocks 2242, 2244), before terminating (at terminate point 2246).
Additional Security Features Provided by the Preferred Embodiment
VDE 100 provided by the preferred embodiment has sufficient security to help ensure that it cannot be compromised short of a successful “brute force attack,” and so that the time and cost to succeed in such a “brute force attack” substantially exceeds any value to be derived. In addition, the security provided by VDE 100 compartmentalizes the internal workings of VDE so that a successful “brute force attack” would compromise only a strictly bounded subset of protected information, not the entire system.
The following are among security aspects and features provided by the preferred embodiment:
Some of these security aspects and considerations are discussed above. The following provides an expanded discussion of preferred embodiment security features not fully addressed elsewhere.
Management of Keys and Shared Secrets
VDE 100 uses keys and shared secrets to provide security. The following key usage features are provided by the preferred embodiment:
different cryptosystem/key types
secure key length
key “convolution” and key “aging.”
Each of these types are discussed below.
A. Public-Key and Symmetric Key Cryptosystems
The process of disguising or transforming information to hide its substance is called encryption. Encryption produces “ciphertext.” Reversing the encryption process to recover the substance from the ciphertext is called “decryption.” A cryptographic algorithm is the mathematical function used for encryption and decryption.
Most modern cryptographic algorithms use a “key.” The “key” specifies one of a family of transformations to be provided. Keys allow a standard, published and tested cryptographic algorithm to be used while ensuring that specific transformations performed using the algorithm are kept secret. The secrecy of the particular transformations thus depends on the secrecy of the key, not on the secrecy of the algorithm.
There are two general forms of key-based algorithms, either or both of which may be used by the preferred embodiment PPE 650:
Symmetric algorithms are algorithms where the encryption key can be calculated from the decryption key and vice versa. In many such systems, the encryption and decryption keys are the same. The algorithms, also called “secret-key”, “single key” or “shared secret” algorithms, require a sender and receiver to agree on a key before ciphertext produced by a sender can be decrypted by a receiver. This key must be kept secret. The security of a symmetric algorithm rests in the key: divulging the key means that anybody could encrypt and decrypt information in such a cryptosystem. See Schneier, Applied Cryptography at Page 3. Some examples of symmetric key algorithms that the preferred embodiment may use include DES, Skipjack/Clipper, IDEA, RC2, and RC4.
In public-key cryptosystems, the key used for encryption is different from the key used for decryption. Furthermore, it is computationally infeasible to derive one key from the other. The algorithms used in these cryptosystems are called “public key” because one of the two keys can be made public without endangering the security of the other key. They are also sometimes called “asymmetric” cryptosystems because they use different keys for encryption and decryption. Examples of public key algorithrris include RSA, El Gamal and LUC.
The preferred embodiment PPE 650 may operate based on only symmetric key cryptosystems, based on public-key cryptosystems, or based on both symmetric key cryptosystems and public key cryptosystems. VDE 100 does not require any specific encryption algorithms; the architecture provided by the preferred embodiment may support numerous algorithms including PK and/or secret key (non PK) algorithms. In some cases, the choice of encryption/decryption algorithm will be dependent on a number of business decisions such as cost, market demands, compatibility with other commercially available systems, export laws, etc.
Although the preferred embodiment is not dependent on any particular type of cryptosystem or encryption/decryption algorithm(s), the preferred example uses PK cryptosystems for secure communications between PPEs 650, and uses secret key cryptosystems for “bulk” encryption/decryption of VDE objects 300. Using secret key cryptosystems (e.g., DES implementations using multiple keys and multiple passes, Skipjack, RC2, or RC4) for “bulk” encryption/decryption provides efficiencies in encrypting and decrypting large quantities of information, and also permits PPEs 650 without PK-capability to deal with VDE objects 300 in a variety of applications. Using PK cryptosystems for communications may provide advantages such as eliminating reliance on secret shared external communication keys to establish communications, allowing for a challenge/response that doesn't rely on shared internal secrets to authenticate PPEs 650, and allowing for a publicly available “certification” process without reliance on shared secret keys.
Some content providers may wish to restrict use of their content to PK implementations This desire can be supported by making the availability of PK capabilities, and the specific nature or type of PK capabilities, in PPEs 650 a factor in the registration of VDE objects 300, for example, by including a requirement in a REGISTER method for such objects in the form of a load module that examines a PPE 650 for specific or general PK capabilities before allowing registration to continue.
Although VDE 100 does not require any specific algorithm, it is highly desirable that all PPEs 650 are capable of using the same algorithm for bulk encryption/decryption. If the bulk encryption/decryption algorithm used for encrypting VDE objects 300 is not standardized, then it is possible that not all VDE electronic appliances 600 will be capable of handling all VDE objects 300. Performance differences will exist between different PPEs 650 and associated electronic appliances 600 if standardized bulk encryption/decryption algorithms are not implemented in whole or in part by hardware-based encrypt/decrypt engine 522, and instead are implemented in software. In order to support algorithms that are not implemented in whole or in part by encrypt/decrypt engine 522, a component assembly that implements such an algorithm must be available to a PPE 650.
B. Key Length
Increased key length may increase security. A “brute-force” attack of a cryptosystem involves trying every possible key. The longer the key, the more possible keys there are to try. At some key length, available computation resources will require an impractically large amount of time for a “brute force” attacker to try every possible key.
VDE 100 provided by the preferred embodiment accommodates and can use many different key lengths. The length of keys used by VDE 100 in the preferred embodiment is determined by the algorithm(s) used for encryption/decryption, the level of security desired, and, throughput requirements. Longer keys generally require additional processing power to ensure fast encryption/decryption response times. Therefore, there is a tradeoff between (a) security, and (b) processing time and/or resources. Since a hardware-based PPE encrypt/decrypt engine 522 may provide faster processing than software-based encryption/decryption, the hardware-based approach may, in general, allow use of longer keys.
The preferred embodiment may use a 1024 bit modulus (key) RSA cryptosystem implementation for PK encryption/decryption, and may use 56-bit DES for “bulk” encryption/decryption. Since the 56-bit key provided by standard DES may not be long enough to provide sufficient security for at least the most sensitive VDE information, multiple DES encryptions using multiple passes and multiple DES keys may be used to provide additional security. DES can be made significantly more secure if operated in a manner that uses multiple passes with different keys. For example, three passes with 2 or 3 separate keys is much more secure because it effectively increases the length of the key. RC2 and RC4 (alternatives to DES) can be exported for up to 40-bit key sizes, but the key size probably needs to be much greater to provide even DES level security. The 80-bit key length provided by NSA's Skipjack may be adequate for most VDE security needs.
The capability of downloading code and other information dynamically into PPE 650 allows key length to be adjusted and changed dynamically even after a significant number of VDE electronic appliances 600 are in use. The ability of a VDE administrator to communicate with each PPE 650 efficiently makes such after the-fact dynamic changes both possible and cost-effective. New or modified cryptosystems can be downloaded into existing PPEs 650 to replace or add to the cryptosystem repertoire available within the PPE, allowing older PPEs to maintain compatibility with newer PPEs and/or newly released VDE objects 300 and other VDE-protected information. For example, software encryption/decryption algorithms may be downloaded into PPE 650 at any time to supplement the hardware-based functionality of encrypt/decrypt engine 522 by providing different key length capabilities. To provide increased flexibility, PPE encrypt/decrypt engine 522 may be configured to anticipate multiple passes and/or variable and/or longer key lengths. In addition, it may be desirable to provide PPEs 650 with the capability to internally generate longer PK keys.
C. Key Generation
Key generation techniques provided by the preferred embodiment permit PPE 650 to generate keys and other information that are “known” only to it.
The security of encrypted information rests in the security of the key used to encrypt it. If a cryptographically weak process is used to generate keys, the entire security is weak. Good keys are random bit strings so that every possible key in the key space is equally likely. Therefore, keys should in general be derived from a reliably random source, for example, by a cryptographically secure pseudo-random number generator seeded from such a source. Examples of such key generators are described in Schneier, Applied Cryptography (John Wiley and Sans, 1994), chapter 15. If keys are generated outside a given PPE 650 (e.g., by another PPE 650), they must be verified to ensure they come from a trusted source before they can be used. “Certification” may be used to verify keys.
The preferred embodiment PPE 650 provides for the automatic generation of keys. For example, the preferred embodiment PPE 650 may generate its own public/private key pair for use in protecting PK-based external communications and for other reasons. A PPE 650 may also generate its own symmetric keys for various purposes during and after initialization. Because a PPE 650 provides a secure environment, most key generation in the preferred embodiment may occur within the PPE (with the possible exception of initial PPE keys used at manufacturing or installation time to allow a PPE to authenticate initial download messages to it).
Good key generation relies on randomness. The preferred embodiment PPE 650—may, as mentioned above in connection with
If no random number generator 542 is available sin the PPE 650, the SPE 503 may employ a cryptographic algorithm (e.g., DES in Output Feedback Mode) to generate a sequence of pseudo-random values derived from a secret value protected within the SPE. Although these numbers are pseudo-random rather than truly random, they are cryptographically derived from a value unknown outside the SPE 503 and therefore may be satisfactory in some applications.
In an embodiment incorporating an HPE 655 without an SPE 503, the random value generator 565 software may derive reliably random numbers from unpredictable external physical events (e.g., high-resolution timing of disk I/O completions or of user keystrokes at an attached keyboard 612).
Conventional techniques for generating PK and non-PK keys based upon such “seeds” may be used. Thus, if performance and manufacturing costs permit, PPE 650 in the preferred embodiment will generate its own public/private key pair based on such random or pseudo-random “seed” values. This key pair may then be used for external communications between the PPE 650 that generated the key pair and other PPEs that wish to communicate with it. For example, the generating PPE 650 may reveal the public key of the key pair to other PPEs. This allows other PPEs 650 using the public key to encrypt messages that may be decrypted only by the generating PPE (the generating PPE is the only PPE that “knows” the corresponding “private key”). Similarly, the generating PPE 650 may encrypt messages using its private key that, when decrypted successfully by other PPEs with the generating PPE's public key, permit the other PPEs to authenticate that the generating PPE sent the message.
Before one PPE 650 uses a public key generated by another PPE, a public key certification process should be used to provide authenticity certificates for the public key. A public-key certificate is someone's public key “signed” by a trustworthy entity such as an authentic PPE 650 or a VDE administrator. Certificates are used to thwart attempts to convince a PPE 650 that it is communicating with an authentic PPE when it is not (e.g., it is actually communicating with a person attempting to break the security of PPE 650). One or more VDE administrators in the preferred embodiment may constitute a certifying authority. By “signing” both the public key generated by a PPE 650 and information about the PPE and/or the corresponding VDE electronic appliance 600 (e.g., site ID, user ID, expiration date, name, address, etc), the VDE administrator certifying authority can certify that information about the PPE and/or the VDE electronic appliance is correct and that the public key belongs to that particular VDE mode.
Certificates play an important role in the trustedness of digital signatures, and also are important in the public-key authentication communications protocol (to be discussed below). In the preferred embodiment, these certificates may include information about the trustedness/level of security of a particular VDE electronic appliance 600 (e.g., whether or not it has a hardware-based SPE 503 or is instead a less trusted software emulation type HPE 655) that can be used to avoid transmitting certain highly/secure information to less trusted/secure VDE installations.
Certificates can also play an important role in decommissioning rogue users and/or sites. By including a site and/or user ID in a certificate, a PPE can evaluate this information as an aspect of authentication. For example, if a VDE administrator or clearinghouse encounters a certificate bearing and ID (or other information) that meets certain criteria (e.g., is present on a list of decommissioned and/or otherwise suspicious users and/or sites), they may choose to take actions based on those criteria such as refusing to communicate, communicating disabling information, notifying the user of the condition, etc. Certificates also typically include an expiration date to ensure that certificates must be replied periodically, for example, to ensure that sites and/or users must stay in contact with a VDE administrator and/or to allow certification keys to be changed periodically. More than one certificate based on different keys may be issued for sites and/or users so-that if a given certification key is compromised, one or more “backup” certificates may be used. If a certification key is compromised, A VDE administrator may refuse to authenticate based on certificates generated with such a key, and send a signal after authenticating with a “backup” certificate that invalidates all use of the compromised key and all certificates associated with it in further interactions with VDE participants. A new one or more “backup” certificates and keys may be created and sent to the authenticated site/user after such a compromise.
If multiple certificates are available, some of the certificates may be reserved as backups. Alternatively or in addition, one certificate from a group of certificates may be selected (e.g., by using RNG 542) in a given authentication, thereby reducing the likelihood that a certificate associated with a compromised certification key will be used. Still alternatively, more than one certificate may be used in a given authentication.
To guard against the possibility of compromise of the certification algorithm (e.g., by an unpredictable advance in the mathematical foundations on which the algorithm is based), distinct algorithms may used for different certificates that are based on different mathematical foundations.
Another technique that may be employed to decrease the probability of compromise is to keep secret (in protected storage in the PPE 650) the “public” values on which the certificates are based, thereby denying an attacker access to values that may aid in the attack. Although these values are nominally “public,” they need be known only to those components which actually validate certificates (i.e., the PPE 650)
In the preferred embodiment, PPE 650 may generate its own certificate, or the certificate may be obtained externally, such as from a certifying authority VDE administrator. Irrespective of where the digital certificate is generated, the certificate is eventually registered by the VDE administrator certifying authority so that other VDE electronic appliances 600 may have access to (and trust) the public key. For example, PPE 650 may communicate its public key and other information to a certifying authority which may then encrypt the public key and other information using the certifying authority's private key. Other installations 600 may trust the “certificate” because it can be authenticated by using the certifying authority's public key to decrypt it. As another example, the certifying authority may encrypt the public key it receives from the generating PPE 650 arid use it to encrypt the certifying authority's private, key. The certifying authority may then send this encrypted information back to the generating PPE 650. The generating PPE 650 may then use the certifying authority's private key to internally create a digital certificate, after which it may destroy its copy of the certifying authority's private key. The generating PPE 650 may then send out its digital certificate to be stored in a certification repository at the VDE administrator (or elsewhere) if desired. The certificate process can also be implemented with an external key pair generator and certificate generator, but might be somewhat less secure depending on the nature of the secure facility. In such a case, a manufacturing key should be used in PPE 650 to limit exposure to the other keys involved.
A PPE 650 may need more than one certificate. For example, a certificate may be needed to assure other users that a PPE is authentic, and to identify the PPE. Further certificates may be needed for individual users of a PPE 650. These certificates may incorporate both user and site information or may only include user information. Generally, a certifying authority will require a valid site certificate to be presented prior to creating a certificate for a given user. Users may each require their own public key/private key pair in order to obtain certificates. VDE administrators, clearinghouses, and other participants may normally require authentication of both the site (PPE 650) and of the user in a communication or other interaction. The processes described above for key generation and certification for PPEs 650 may also be used to form site/user certificates or user certificates.
Certificates as described above may also be used to certify the origin of load modules 1100 and/or the authenticity of administrative operations. The security and assurance techniques described above may be employed to decrease the probability of compromise for any such certificate (including certificates other than the certificate for a VDE electronic appliance 600's identity).
D. Key Aging and Convolution
PPE 650 also has the ability in the preferred embodiment to generate secret keys and other information that is shared between multiple PPEs 650. In the preferred embodiment, such secret keys and other information may be shared between multiple VDE electronic appliances 600 without requiring the shared secret information to ever be communicated explicitly between the electronic appliances. More specifically, PPE 650 uses a technique called “key convolution” to derive keys based on a deterministic process in response to seed information shared between multiple VDE electronic appliances 600. Since the multiple electronic appliances 600 “know” what the “seed” information is and also “know” the deterministic process used to generate keys based on this information, each of the electronic appliances may independently generate the “true key.” This permits multiple VDE electronic appliances 600 to share a common secret keys without potentially compromising its security by communicating it over an insecure channel.
No encryption key should be used for an indefinite period. The longer a key is used, the greater the chance that it may be compromised and the greater the potential loss if the key is compromised but still in use to protect new information. The longer a key is used, the more information it may protect and therefore the greater the potential rewards for someone to spend the effort necessary to break it. Further, if a key is used for a long time, there may be more ciphertext available to an attacker attempting to break the key using a ciphertext based attack. See Schneier at 150-151. Key convolution in the preferred embodiment provides a way to efficiently change keys stored in secure database 610 on a routine periodic or other basis while ˜simplifying key management issues surrounding the change of keys. In addition, key convolution may be used to provide “time aged keys” (discussed below) to provide “expiration dates” for key usage and/or validity.
The current convolution key 2862 represents an encoding of the site ID 2821 and current time. This transformed value 2862 may be used as a key for another encryption process 2872 to transform the stored key 810 in the object's PERC 808 into the true private body key 2863 for the object's contents.
The “convolution function” performed by blocks 2861, 2871 may, for example, be a one-way function that can be performed independently at both the content creator's site and at the content user's site. If the content user does not use precisely the same convolution function and precisely the same input values (e.g., time and/or site and/or other information) as used by the content creator, then the result of the convolution function performed by the content user will be different from the content creator's result. If the result is used as a symmetrical key for encryption by the content creator, the content user will not be able to decrypt unless the content user's result is the same as the result of the content creator.
The time component for input to the key convolution function may be derived from RTC 528 (care being taken to ensure that slight differences in RTC synchronization between VDE electronic appliances will not cause different electronic appliances to use different time components). Different portions of the RTC 528 output may be used to provide keys with different valid durations, or some tolerance can be built into the process to try several different key values. For example, a “time granularity” parameter can be adjusted to provide time tolerance in terms of days, weeks, or any other time period. As one example, if the “time granularity” is set to 2 days, and the tolerance is ±2 days, then three real-time input values can be tried as input to the convolution algorithm. Each of the resulting key values may be tried to determine which of the possible keys is actually used. In this example, the keys will have only a 4 day life span.
Meanwhile, the creator site may use the convolution step 2871(z) based on his RTC 528 value (adjusted to correspond to the intended validity time for the key) to generate a convoluted key 2862(z), which may then be used to generate the content key 2863 in the object's PERC 808. To decrypt the object's content, the user site may use each of its sequence of convolution key 2862(a-e) to attempt to generate the master content key 810. When this is attempted, as long as the RTC 538 of the creator site is within acceptable tolerance of the RTC 528 at the user site, one of keys 2862(a-e) will match key 2862(z) and the decryption will be successful. In this example, matching is determined by validity of decrypted output, not by direct comparison of keys.
Key convolution as described above need not use both site ID and time as a value. Some keys may be generated based on current real time, other keys might be generated on site ID, and still other keys might be generated based on both current real-time and site ID.
Key convolution can be used to provide “time-aged” keys. Such “time-aged” keys provide an automatic mechanism for allowing keys to expire and be replaced by “new” keys. They provide a way to give a user time-limited rights to make time-limited use of an object, or portions of an object, without requiring user re-registration but retaining significant control in the hands of the content provider or administrator. If secure database 610 is sufficiently secure, similar capabilities can be accomplished by checking an expiration date/time associated with a key, but this requires using more storage space for each key or group of keys.
In the preferred embodiment, PERCs, 808 can include an expiration date and/or time after which access to the VDE-protected information they correspond to is no longer authorized. Alternatively or in addition, after a duration of time related to some aspect of the use of the electronic appliance 600 or one or more VDE objects 300, a PERC 808 can force a user to send audit history information to a clearinghouse, distributor, client administrator, or object creator in order to regain or retain the right to use the object(s). The PERC 808 can enforce such time-based restrictions by checking/enforcing parameters that limit key usage and/or availability past time of authorized use. “Time aged” keys may be used to enforce or enhance this type of time-related control of access to VDE protected information.
“Time aged” keys can be used to encrypt and decrypt a set of information for a limited period of time, thus requiring re-registration or the receipt of new permissions or the passing of audit information, without which new keys are not provided for user use. Time aged keys can also be used to improve system security since one or more keys would be automatically replaced based on the time aging criteria—and thus, cracking secure database 610 and locating one or more keys may have no real value. Still another advantage of using time aged keys is that they can be generated dynamically—thereby obviating the need to store decryption keys in secondary and/or secure memory.
A “time aged key” in the preferred embodiment is not a “true key” that can be used for encryption/decryption, but rather is a piece of information that a PPE 650, in conjunction with other information, can use to generate a “true key.” This other information can be time-based, based on the particular “ID” of the PPE 650, or both. Because the “true key” is never exposed but is always generated within a secure PPE 650 environment, and because secure PPEs are required to generate the “true key,” VDE 100 can use “time aged” keys to significantly enhance security and flexibility of the system.
The process of “aging” a key in the preferred embodiment involves generating a time-aged “true key” that is a function of: (a) a “true key,” and (b) some other information (e.g., real time parameters, site ID parameters, etc.) This information is combined/transformed (e.g., using the “key convolution” techniques discussed above) to recover or provide a “true key.” Since the “true key” can be recovered, this avoids having to store the “true key” within PERC 808, and allow different “true keys” to correspond to the same information within PERC 808. Because the “true key” is not stored in the PERC 808, access to the PERC does not provide access to the information protected by the “true key.” Thus, “time aged” keys allows content creators/providers to impose a limitation (e.g., site based and/or time based) on information access that is, in a sense, “external of” or auxiliary to the permissioning provided by one or more PERCs 808. For example, a “time aged” key may enforce an additional time limitation on access to certain protected information, this additional time limitation being independent of any information or permissioning contained within the PERC 808 and being instead based on one or more time and/or site ID values.
As one example, time-aged decryption keys may be used to allow the purchaser of a “trial subscription” of an electronically published newspaper to access each edition of the paper for a period of one week, after which the decryption keys will no longer work. In this example, the user would need to purchase one or more new PERCs 808, or receive an update to an existing one or more permissions records, to access editions other than the ones from that week. Access to those other editions which might be handled with a totally different pricing structure (e.g., a “regular” subscription rate as opposed to a free or minimal “trial” subscription rate).
In the preferred embodiment, time-aged-based “true keys” can be generated using a one-way or invertible “key convolution” function. Input parameters to the convolution function may include the supplied time-aged keys; user and/or site specific values; a specified portion (e.g., a certain number of high order bits) of the time value from an RTC 528 (if present) or a value derived from such time value in a predefined manner; and a block or record identifier that may be used to ensure that each time aged key is unique. The output of the “key convolution” function may be a “true key” that is used for decryption purposes until discarded. Running the function with a time-aged key and inappropriate time values typically yields a useless key that will not decrypt.
Generation of a new time aged key can be triggered based on some value of elapsed, absolute or relative time (e.g., based on a real time value from a clock such as RTC 528). At that time, the convolution would produce the wrong key and decryption could not occur until the time-aged key is updated. The criteria used to determine when a new “time aged key” is to be created may itself be changed based on time or some other input variable to provide yet another level of security. Thus, the convolution function and/or the event invoking it may change, shift or employ a varying quantity as a parameter.
One example of the use of time-aged keys is as follows:
It performs a convolution function (i.e., the inverse of reverse convolution” algorithm in step (2) above) to obtain the “true” key. If the supplied time and/or other information is “wrong,” the convolution function will not yield the “true” key, and therefore content cannot be decrypted.
Any of the key blocks associated with VDE objects 300 or other items can be either a regular key block or a time-aged key block, as specified by the object creator during the object configuration process, or where appropriate, a distribution or client administrator.
“Time aged” keys can also be used as part of protocols to provide secure communications between PPEs 650. For example, instead of providing “true” keys to PPE 650 for communications, VDE 100 may provide only “partial” communication keys to the PPE. These “partial” keys may be provided to PPE 650 during initialization, for example. A predetermined algorithm may produce “true keys” for use to encrypt/decrypt information for secure communications. The predetermined algorithm can “age” these keys the same way in all PPEs 650, or PPEs 650 can be required to contact a VDE administrator at some predetermined time so a new set of partial communications keys can be downloaded to the PPEs. If the PPE 650 does not generate or otherwise obtain “new” partial keys, then it will be disabled from communicating with other PPEs (a further, “fail safe” key may be provided to ensure that the PPE can communicate with a VDE administrator for reinitialization purposes). Two sets of partial keys can be maintained within a PPE 650 to allow a fixed amount of overlap time across all VDE appliances 600. The older of the two sets of partial keys can be updated periodically.
The following additional types of keys (to be discussed below) can also be “aged” in the preferred embodiment:
individual message keys (i.e., keys used for a particular message),
administrative, stationary and travelling object shared keys, secure database keys, and
private body and content keys.
Initial Installation Key Management
The manufacturer possesses (i.e., knows, and protects from disclosure or modification) one or more public key 2811/private key 2812 key pairs used for signing and validating site identification certificates 2821. For each site, the manufacturer generates a site ID 2821 and list of site characteristics 2822. In addition, the manufacturer possesses the public keys 2813, 2814 for validating load modules and initialization code downloads. To enhance security, there may be a plurality of such certification keys, and each PPE 650 may be initialized using only a subset of such keys of each type.
As part of the initialization process, the PPE 650 may generate internally or the manufacturer may generate and supply, one or more pairs of site-specific public keys 2815 and private keys 2816. These are used by the PPE 650 to prove its identity. Similarly, site-specific database key(s) 2817 for the site are generated, and if needed (i.e., if a Random Number Generator 542 is not available), a random initialization seed 2818 is generated.
The initialization may begin by generating site ID 2821 and characteristics 2822 and the site, public key 2815/private key 2816 pair(s). These values are combined and may be used to generate one or more site identity certificates 2823. The site identity certificates 2823 may be generated by the public key generation process 2804, and may be stored both in the PPE's protected key storage 2802 and in the manufacturer's VDE site certificate database 2803.
The certification process 2804 maybe performed either by the manufacturer or internally to the PPE 650. If performed by the PPE 650, the PPE will temporarily receive the identity certification private key(s) 2812, generate the certificate 2823, store the certificate in local key storage 2802 and transmit it to the manufacturer, after which the PPE 650 must erase its copy of the identity certification private key(s) 2812.
Subsequently, initialization may require generation, by the PPE 650 or by the manufacturer, of site-specific database key(s) 2817 and of site-specific seed value(s) 2818, which are stored in the ‘key storage 2802. In addition, the download certification key(s) 2814 and the load module certification key(s) 2813 maybe supplied by the manufacturer and stored in the key storage 2802.
These may be used by the PPE 650 to validate all further communications with external entities.
At this point, the PPE 650 may be further initialized with executable code and data by downloading information certified by the load module key(s) 2813 and download key(s) 2814. In the preferred embodiment, these keys may be used to digitally sign data to be loaded into the PPE 650, guaranteeing its validity, and additional key(s) encrypted using the site-specific public key(s) 2815 may be used to encrypt such data and protect it from disclosure.
Installation and Update Key Management
To perform this installation, the installer retrieves the destination site's identity certificate(s) 2823, and from that extracts the site public key(s) 2815. These key(s) may be used in an encryption process 2841 to protect the keys being installed. The key(s) being installed are then transmitted inside the destination site's PPE 650. Inside the PPE 650, the decryption process 2842 may use the site private key(s) 2816 to decrypt the transmission. The PPE 650 then stores the installed or updated keys in its key storage 2802.
Object-Specific Key Use
The detailed descriptions of key types below further explain secret-key embodiments, this summary is not intended as a complete description. The preferred embodiment PPE 650 can use different types of keys and/or different “shared secrets” for different purposes. Some key types apply to a Public-Key/Secret Key implementation, other keys apply to a Secret Key only implementation, and still other key types apply to both. The following table lists examples of various key and “shared secret” information used in the preferred embodiment, and where this information is used and stored:
A “master” key is a key used to encrypt other keys. An initial or “master” key may be provided within PPE 650 for communicating other keys in a secure way. During initialization of PPE 650, code and shared keys are downloaded to the PPE. Since the code contains secure convolution algorithms and/or coefficients, it is comparable to a “master key” The shared keys may also be considered “master keys.”
If public-key cryptography is used as the basis for external communication with PPE 650, then a master key is required during the PPE Public-key pair certification process. This master key may be, for example, a private key used by the manufacturer or VDE administrator to establish the digital certificate (encrypted public key and other information of the PPE), or it may, as another example, be a private key used by a VDE administrator to encrypt the entries in a certification repository. Once certification has occurred, external communications between PPEs 650 may be established using the certificates of communicating PPEs.
If shared secret keys are used as the basis for external communications, then an initial secret key is required to establish external communications for PPE 650 initialization. This initial secret key is a “master key” in the sense that it is used to encrypt other keys. A set of shared partial external communications keys (see discussion above) may be downloaded during the PPE initialization process, and these keys are used to establish subsequent external PPE communications.
A manufacturing key is used at the time of PPE manufacture to prevent knowledge by the manufacturing staff of PPE-specific key information that is downloaded into a PPE at initialization time. For example, a PPE 650 that operates as part of the manufacturing facility may generate information for download into the PPE being initialized. This information must be encrypted during communication between the PPEs 650 to keep it confidential, or otherwise the manufacturing staff could read the information. A manufacturing key is used to protect the information. The manufacturing key maybe used, to protect various other keys downloaded into the PPE such as, for example, a certification private key, a PPE public/private key pair, and/or other keys such as shared secret keys specific to the PPE. Since the manufacturing key is used to encrypt other keys, it is a “master key.”
A manufacturing key may be public-key based, or it may be based on a shared secret. Once the information is downloaded, the now initialized PPE 650 can discard (or simply not use) the manufacturing key. A manufacturing key may be hardwired into PPE 650 at manufacturing time, or sent to the PPE as its first key and discarded after it is no longer needed. As indicated in the table above and in the preceding discussion, a manufacturing key is not required if PK capabilities are included in the PPE.
Certification Key Pair
A certification key pair may be used as part of a “certification” process for PPEs 650 and VDE electronic appliances 600. This certification process in the preferred embodiment may be used to permit a VDE electronic appliance to present one or more “certificates” authenticating that it (or its key) can be trusted. As described above, this “certification” process may be used by one PPE 650 to “certify” that it is an authentic VDE PPE, it has a certain level of security and capability set (e.g., it is hardware based rather than merely software based), etc. Briefly, the “certification” process may involve using a certificate private key of a certification key pair to encrypt a message including another VDE node's public-key. The private key of a certification key pair is preferably used to generate a PPE certificate. It is used to encrypt a public-key of the PPE. A PPE certificate can either be stored in the PPE, or it may be stored in a certification repository.
Depending on the authentication technique chosen, the, public key and the private key of a certification key pair may need to be protected. In the preferred embodiment, the certification public key(s) is distributed amongst PPEs such that they may make use of them in decrypting certificates as an aspect of authentication. Since, in the preferred embodiment, this public key is used inside a PPE 650, there is no need for this public key to be available in plaintext, and in any event it is important that such key be maintained and transmitted with integrity (e.g., during initialization and/or update by a VDE administrator). If the certification public key is kept confidential (i.e., only available in plaintext inside the PPE 650), it may make cracking security much more difficult. The private key of a certification key pair should be kept confidential and only be stored by a certifying authority (i.e., should not be distributed).
In order to allow, in the preferred embodiment, the ability to differentiate installations with different levels/degrees of trustedness/security, different certification key pairs may be used (e.g., different certification keys may be used to certify SPEs 503 then are used to certify HPEs 655).
PPE Public/Private Key Pair
In the preferred embodiment, each PPE 650 may have its own unique “device” (and/or user) public/private key pair. Preferably, the private key of this key pair is generated within the PPE and is never exposed in any form outside of the PPE. Thus, in one embodiment, the PPE 650 may be provided with an internal capability for generating key pairs internally. If the PPE generates its own public-key cryptosystem key pairs internally, a manufacturing key discussed above may not be needed. If desired, however, for cost reasons a key pair may be exposed only at the time a PPE 650 is manufactured, and may be protected at that time using a manufacturing key. Allowing PPE 650 to generate its public key pair internally allows the key pair to be concealed, but may in some applications be outweighed by the cost of putting a public-key key pair generator into PPE 650.
Initial Secret Key
The initial secret key is used as a master key by a secret key only based PPE 650 to protect information downloaded into the PPE during initialization. It is generated by the PPE 650, and is sent from the PPE to a secure manufacturing database encrypted using a manufacturing key. The secure database sends back a unique PPE manufacturing ID encrypted using the initial secret key in response.
The initial secret key is likely to be a much longer key than keys used for “standard” encryption due to its special role in PPE initialization. Since the resulting decryption overhead occurs only during the initialization process, multiple passes through the decryption hardware with selected portions of this key are tolerable.
PPE Manufacturing ID
The PPE manufacturing ID is not a “key,” but, does fall within the classic definition of a “shared secret.” It preferably uniquely identifies a PPE 650 and may be used by the secure database 610 to determine the PPE's initial secret key during the PPE initialization process.
Site ID, Shared Code, Shared Keys and Shared Secrets
The VDE site ID along with shared code, keys and secrets are preferably either downloaded into PPE 650 during the PPE initialization process, or are generated internally by a PPE as part of that process. In the preferred embodiment, most or all of this information is downloaded.
The PPE site ID uniquely identifies the PPE 650. The site ID is preferably unique so as to uniquely identify the PPE 650 and distinguish that PPE from all other PPEs. The site ID in the preferred embodiment provides a unique address that may be used for various purposes, such as for example to provide “address privacy” functions. In some cases, the site ID may be the public key of the PPE 650. In other cases, the PPE site ID may be assigned during the manufacturing and/or initialization process. In the case of a PPE 650 that is not public-key capable, it would not be desirable to use the device secret key as the unique site ID because this would expose too many bits of the key—and therefore a different information string should be used as the site ID.
Shared code comprises those code fragments that provide at least a portion of the control program for the PPE 650. In the preferred embodiment, a basic code fragment is installed during PPE manufacturing that permits the PPE to bootstrap and begin the initialization process. This fragment can be replaced during the initialization process, or during subsequent download processing, with updated control logic.
Shared keys may be downloaded into PPE 650 during the initialization process. These keys may be used, for example, to decrypt the private headers of many object structures.
When PPE 650 is operating in a secret key only mode, the initialization and download processes may import shared secrets into the PPE 650. These shared secrets may be used during communications processes to permit PPEs 650 to authenticate the identity of other PPEs and/or users.
Download Authorization Key
The download authorization key is received by PPE 650 during the initialization download process. It is used to authorize further PPE 650 code updates, key updates, and may also be used to protect PPE secure database 610 backup to allow recovery by a VDE administrator (for example) if the PPE fails. It may be used along with the site ID, time and convolution algorithm to derive a site ID specific key. The download authorization key may also be used to encrypt the key block used to encrypt secure database 610 backups. It may also be used to form a site specific key that is used to enable future downloads to the PPE 650. This download authorization key is not shared among all PPEs 650 in the preferred embodiment; it is specific to functions performed by authorized VDE administrators.
External Communications Keys and Related Secret and Public Information
There are several cases where keys are required when PPEs 650 communicate. The process of establishing secure communications may also require the use of related public and secret information about the communicating electronic appliances 600. The external communication keys and other information are used to support and authenticate secure communications. These keys comprise a public-key pair in the preferred embodiment although shared secret keys may be used alternatively or in addition.
Administrative Object Keys
In the preferred embodiment, an administrative object shared key may be used to decrypt the private header of an administrative object 870. In the case of administrative objects, a permissions record 808 may be present in the private header. In some cases, the permissions record 808 may be distributed as (or within) an administrative object that performs the function of providing a right to process the content of other administrative objects. The permissions record 808 preferably contains the keys for the private body, and the keys for the content that can be accessed would be budgets referenced in that permissions record 808. The administrative object shared keys may incorporate tune as a component, and may be replaced when expired.
Stationary Object Keys
A stationary object shared key may be used to decrypt a private header of stationary objects 850. As explained above, in some cases a permissions record 808 may be present in the private header of stationary objects. If present, the permissions record 808 may contain the keys for the private body but will not contain the keys for the content. These shared keys may incorporate time a component, and may be replaced when expired.
Traveling Object Shared Keys
A traveling object shared key may be used to decrypt the private header of traveling objects 860. In the preferred embodiment, traveling objects contain permissions record 808 in their private headers. The permissions record 808 preferably contains the keys for the private body and the keys for the content that can be accessed as permitted by the permissions record 808. These shared keys may incorporate time as a component, and may be replaced when expired.
Secure Database Keys
PPE 650 preferably generates these secure database, keys and never exposes them outside of the PPE. They are site-specific in the preferred embodiment, and may be “aged” as described above. As described above, each time an updated record is written to secure database 610, a new key may be used and kept in a key list within the PPE. Periodically (and when the internal list has no more room), the PPE 650 may generate a new key to encrypt new or old records. A group of keys may be used instead of a single key, depending on the size of the secure database 610.
Private Body Keys
Private body keys are unique to an object 300, and are not dependent on key information shared between PPEs 650. They are preferably generated by the PPE 650 at the time the private body is encrypted, and may incorporate real time as a component to “age” them. They are received in permissions records 808, and their usage may be controlled by budgets.
Content Keys are unique to an object 300, and are not dependent on key information shared between PPEs 650. They are preferably generated by the PPE 650 at the time the content is encrypted. They may incorporate time as a component to “age” them. They are received in permissions records 808, and their usage may be controlled by budgets.
Authorization Shared Secrets
Access to and use of information within a PPE 650 or within a secure database 610 may be controlled using authorization “shared secrets” rather than keys. Authorization shared secrets may be stored within the records they authorize (permissions records 808, budget records, etc.). The authorization shared secret may be formulated when the corresponding record is created. Authorization shared secrets can be generated by an authorizing PPE 650, and may be replaced when record updates occur. Authorization shared secrets have some characteristics associated with “capabilities” used in capabilities based operating systems. Access tags (described below) are an important set of authorization shared secrets in the preferred embodiment.
As described above, the secure database 610 backup consists of reading all secure database records and current audit “roll ups” stored in both PPE 650 and externally. Then, the backup process decrypts and re-encrypts this information using a new set of generated keys. These keys, the time of the backup, and other appropriate information to identify the backup, may be encrypted multiple times and stored with the previously encrypted secure database files and roll up data within the backup files. These files may then all be encrypted using a “backup key” that is generated, and stored within PPE 650. This backup key 500 may be used by the PPE to recover a backup if necessary. The backup keys may also be securely encrypted (e.g., using a download authentication key and/or a VDE administrator public key) and stored within the backup itself to permit a VDE administrator to recover the backup in case of PPE 650 failure.
Sealing is used to protect the integrity of information when it may be subjected to modifications outside the control of the PPE 650, either accidentally or as an attack on the VDE security. Two specific applications may be the computation of check values for database records and the protection of data blocks that are swapped out of an SPE 500.
There are two types of sealing: keyless sealing, also known as cryptographic hashing, and keyed sealing. Both employ a cryptographically strong hash function, such as MD5 or SHA. Such a function takes an input of arbitrary size and yields a size hash, or “digest.” The digest has the property that it is infeasible to compute two inputs that yield the same digest, and infeasible to compute one input that yields a specific digest value, where “infeasible” is with reference to a work factor based on the size of the digest value in bits. If, for example, a 256 bit hash function is to be called strong, it must require approximately on average 10ˆ38 (2ˆ128) trials before a duplicated or specified digest value is likely to be produced.
Keyless seals may be employed as check values in database records (e.g., in PERC 808) and similar applications. A keyless seal may be computed based on the content of the body of the record, and the seal stored with the rest of the record. The combination of seal and record may be encrypted to protect it in storage. If someone modifies the encrypted record without knowing the encryption key (either in the part representing the data or the part representing the seal), the decrypted content will be different, and the decrypted check value will not match the digest computed from the record's data. Even though the hash algorithm is known, it is not feasible to modify both the record's data and its seal to correspond because both are encrypted.
Keyed seals may be employed as protection for data stored outside a protected environment without encryption, or as a validity proof between two protected environments. A keyed seal is computed similarly to a keyless seal, except that a secret initial value is logically prefixed to the data being sealed. The digest value thus depends both on the secret and the data, and it is infeasible to compute a new seal to correspond to modified data even though the data itself is visible to an attacker. A keyed seal may protect data in storage with a single secret value, or may protect data in transit between two environments that share a single secret value.
The choice of keys or keyless seals depends on the nature of the data being protected and whether it is additionally protected by encryption.
Tagging is particularly useful for supporting the secure storage of important component assembly and related information on secondary storage memory 652. Integrated use of information “tagging” and encryption strategies allows use of inexpensive mass storage devices to securely store information that, in part enables, limits and/or records the configuration, management and operation of a VDE node and the use of VDE protected content.
When encrypted or otherwise secured information is delivered into a user's secure VDE processing area (e.g., PPE 650), a portion of this information can be used as a “tag” that is first decrypted or otherwise unsecured and then compared to an expected value to confirm that the information represents expected information. The tag thus can be used as a portion of a process confirming the identity and correctness of received, VDE protected, information.
Three classes of tags that may be included in the control structures of the preferred embodiment:
These Tags Have Distinct Purposes.
An access tag may be used as a “shared secret” between VDE protected elements and entities authorized to read and/or modify the tagged element(s). The access tag may be broken into separate fields to control different activities independently. If an access tag is used by an element such as a method core 1000′, administrative events that affect such an element must include the access tag (or portion of the access tag) for the affected element(s) and assert that tag when an event is submitted for processing. If access tags are maintained securely (e.g., created inside a PPE 650 when the elements are created, and only released from PPE 650 in encrypted structures), and only distributed to authorized parties, modification of structures can be controlled more securely. Of course, control structures (e.g., PERCs 808) may further limit or qualify modifications or other actions expressed in administrative events.
Correlation tags are used when one element references another element. For example, a creator might be required by a budget owner to obtain permission and establish a business relationship prior to referencing their budget within the creator's PERCs. After such relationship was formed, the budget owner might transmit one or more correlation tags