|Publication number||USRE37613 E1|
|Application number||US 09/272,936|
|Publication date||Mar 26, 2002|
|Filing date||Mar 18, 1999|
|Priority date||May 25, 1994|
|Also published as||CA2149823A1, EP0684560A1, US5613158|
|Publication number||09272936, 272936, US RE37613 E1, US RE37613E1, US-E1-RE37613, USRE37613 E1, USRE37613E1|
|Inventors||Thomas Warren Savage|
|Original Assignee||Compaq Cupertino Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (17), Non-Patent Citations (1), Referenced by (12), Classifications (10), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention relates to addressing techniques used for bus oriented computer systems. More particularly, this invention relates to an automatic addressing technique for flexibly specifying the individual addresses of a plurality of devices coupled to an information bus.
Bus oriented computer systems are known in which individual devices connected to an information bus are assigned unique addresses specifying the location of each device in a computer system. Typically, individual devices are provided with a mechanical, electrical or electromechanical device capable of being set to provide a unique address for the device. For example, in some known arrangements, each device is provided with jumper terminals which can be connected to an appropriate voltage (e.g., ground) in such a manner that the voltage level on the combined collection of jumper terminals uniquely specifies the physical address of that device, usually using a binary numbering system. In still other known system, manually actuatable switches are employed for the same purpose. In both types of arrangement, the specification of a unique device address requires that a user, usually a technician, manipulate the address specifying device, which creates the possibility for human error in initially configuring the system. In addition, to reconfigure such a system, for example by adding more bus devices, the user must check the setting of each device (or a master list on which the individual addresses are recorded) in order to ensure that the new addresses to be assigned to the additional devices do not duplicate already used addresses.
Still other systems employ active electronic devices capable of responding to interrogation from a host computer by setting a device address generated by the host computer and transmitted thereto over the information bus. Such arrangements require additional active electronic circuits in the bus devices, and also require special programming capability in the host computer to generate the individual device address values and transmit such values to the individual bus devices. Also required is a routine for establishing that an address transmitted to a device was correctly received and stored by that device. This type of arrangement required a relatively sophisticated programming approach and is prone to both software and signaling errors.
The invention comprises an automatic addressing technique for specifying the individual unique addresses of an array of devices which is relatively simple and inexpensive to implement, highly reliable in operation, capable of establishing any address sequence required in a given application, and highly compatible with highly configurable computer system.
From a first aspect, the invention comprises a method of specifying the physical address of a plurality of devices each requiring a unique address in an array, the method including the steps of selecting the individual desired address sequence, creating an anchor pattern representing an initial device address, the anchor pattern having a plurality of multi-bit rank ordered fields each having a superior end position and an inferior end position, creating a bit pattern representing the next address in the sequence by shifting the bit pattern in each field in the direction of the superior end position of that field by an integral multiple (preferably unity) of one rank and relocating the superior end rank position bit of the pattern to the inferior end rank position, and continuing the step of creating a bit pattern representing the next address until the last address in the sequence is attained. For an anchor pattern comprising N bits, the number of the fields can range from N to i, where i is the minimum number of bits required to uniquely specify a total number of J devices. Each address is specified with a bit of predetermined rank from each of the i fields.
From an apparatus standpoint, the invention comprises a multiconductor bus device for specifying a unique physical address for each of a plurality of J devices in an array in accordance with a desired address sequence, the bus device comprising N conductors arranged in parallel and grouped into i fields, where i is the minimum number of bits required to uniquely specify J devices, each field having a plurality of rank ordered bits with a superior end position and an inferior end position, the physical address for each device being determined by i bit values selected from one conductor of each field, and a plurality of transform elements each having an input coupled to the N conductors for converting a physical device address presented at the input thereto to the next physical device address at the output thereof, each transform element including means for shifting the bit pattern in each field of a physical device address presented at the input by an integral multiple (preferably unity) of one rank and relocating the superior end rank position bit of the input address to the inferior rank position of the output address. The number of bit values i used to determine the physical address for each device may be increased up to a value of N, if desired, to create additional potential address sequences.
Each transform element preferably comprises a plurality of N input terminals, a plurality of N output terminals and a plurality of N conductive paths coupled between the input terminals and the output terminals, one path coupling the input terminal corresponding to the superior end rank position of each field to the output terminal located at the inferior rank position of the corresponding field.
From a different method aspect, the invention comprises a method of producing a multi-bit anchor pattern capable of transformation into a desired sequence of physical device addresses, each address comprising a plurality of bits each selected from a different one of a plurality of rank ordered multi-bit fields, each bit in each address corresponding to a preselected rank in the associated multi-bit field. The method proceeds by selecting a desired address sequence, converting each address of the desired sequence to a multi-field binary equivalent having a number of fields equal to the number of bits comprising each address, and arranging the bits in each binary equivalent field in a sequential order related to the manner in which the anchor pattern is transformed into the desired sequence of physical device addresses.
The invention provides a highly reliable device addressing capability which is extremely simple to implement, requiring only passive electrically conductive paths arranged in a repeated predetermined pattern in order to effect the sequential address transformations. Since the desired sequence can be unlimited, the invention is extremely useful and effective is bus oriented computer systems which are intended to be highly configurable. Moreover, the specification of a desired address sequence is completely determined once the anchor pattern has been established for a given desired sequence.
For a fuller understanding of the nature and advantages of the invention, reference should be had to the ensuing detailed description taken in conjunction with the accompanying drawings.
FIG. 1 is a schematic diagram illustrating the preferred embodiment of the invention;
FIG. 2 illustrates the address transform element used in the preferred embodiment;
FIG. 3 illustrates the anchor pattern required to effect a first address sequence; and
FIGS. 4 and 5 illustrate other anchor patterns required to effect two different address sequences.
Turning now to the drawings, FIG. 1 is a schematic diagram illustrating the invention as implemented in conjunction with a Small Computer Systems Interface (SCSI) bus system. The SCSI bus standards are defined in ANSI document X3.131, the disclosure of which is hereby incorporated by reference. As seen in this figure, a SCSI bus generally designated with reference numeral 10 has a conventional terminator 12 coupled to each end thereof, and a plurality of computer devices coupled together in a serial or daisy-chain configuration. The first such device is an initiator 14, typically a computer. The remaining seven devices are target devices, usually storage devices (such as disc drives or tape drives) 15-21.
The SCSI bus 10 is supplemented by a multi-bit address bus generally designated with reference numeral 25 which conveys physical device addresses. Bus 25 is coupled to a host device (not illustrated) which generates a multi-bit character termed an anchor pattern, which is applied to one end of bus 25, for example upper end 27 illustrated in FIG. 1. Each bus device 14-21 is coupled to address bus 25, and address bus 25 includes a plurality of anchor pattern transform elements 28 at locations intermediate the connection points of the SCSI bus devices 14-21 to the address bus 25. In the specific embodiment illustrated in FIG. 1 and described below, each SCSI bus device 14-21 is connected to the same three conductors within address bus 25, these connection points being termed the tap points. Since the addressing scheme is a binary based system, three lines, each corresponding to a bit of address information, are the minimum required to uniquely specify eight different devices. Also, for reasons explained more fully below, fourteen conductors are used in address bus 25 to generate any one of a large number of desired sequences of physical device addresses 0-7 in the FIG. 1 system.
FIG. 2 illustrates the address transform element 28 used in the embodiment of FIG. 1. In FIG. 2, the fourteen conductors within address bus 25 are ordered 0-13, as indicated, and the three tap connections are taken from conductor Nos. 1, 4 and 10. As will become more evident below, the three bit address is composed of three fields: the first field comprised of conductors 0 and 1, a second field composed of conductors 2-5, and a third field composed of conductors 6-13. As will be appreciated by those skilled in the art, these fields correspond to the 20, 21 and 22 fields, respectively, of a three bit binary number. In each transform element 28, the conductors in each field are connected to effect a one order shifting of each conductor in a fixed direction, field by field, with the last conductor in the direction of the shift being coupled back to the first conductor in the field. This may be most readily understood by referring to the 22 field comprising conductors 6-13. The input side of conductor 13 emerges from transform element 28 as conductor 12, 12 emerges as 11, 11 emerges as 10, etc. Thus, conductors 13-7 are effectively right shifted one rank in this field. Conductor 6, which is the edgemost conductor in the shifted direction (right), emerges as conductor 13 on the output side. Thus, the field composed of conductors 6-13 is shifted right by one rank, with the end conductor in the direction of the shift, hereinafter termed the superior end position, relocated to the opposite end of the rank, termed the inferior end position.
Similarly, in the field composed of conductors 2-5, conductors 3, 4 and 5 are right shifted between the input side and the output side, and the superior end conductor (conductor 2) is relocated to the inferior end position (as emerging conductor 5).
The same transform is effected on the conductors in the fields composed of conductors 0 and 1; however, since there are only two conductors, the shifting and relocation of the superior end conductor to the inferior end position merely requires the transposition of the two conductors 0 and 1.
As noted above, each transform element 28 in the address bus 25 has the identical configuration to that just described with reference to FIG. 2.
An important feature of the invention is the addressing flexibility afforded by the transform elements 28 when combined with the anchor pattern selection described below. In essence, the SCSI bus devices 14-21 can be assigned physical addresses in a wide variety of different sequential address orders by selecting the correct anchor pattern. For example, SCSI bus devices 14-21 may simply be addressed in their order of physical proximity to anchor terminal 27, designating initiator 14 with address 0, target device 15 with address 1, target device 16 with address 2, etc.
FIG. 3 illustrates the anchor pattern required to effect this particular address sequence. In this figure, the bus conductors are numbered in the same order as that described above with reference to FIG. 2 and the three field grouping is indicated by the vertical separation lines. Also, the tap locations 0-7 corresponding to the SCSI bus devices 14-21 are indicated in the leftmost column. The anchor pattern is the pattern to the immediate right of tap location 0 and the transformed patterns at tap locations 1-7 are sequentially presented in rows from top to bottom of the figure. The three tap lines are chosen in this example are conductor Nos. 1 (field 0), 4 (field 1) and 10 (field 2), which is the same tap arrangement as that depicted in FIG. 2. Thus, at tap location 0, the designated three bit address is binary 000, which corresponds to 0 decimal. At tap location 1, the three bit address is 001, which is decimal 1, the three bit address at tap location 2 is binary 010 corresponding to decimal 2, etc. Thus, by applying a 14-bit anchor pattern of 00001111001101, the address sequence 0-7 is obtained automatically by virtue of the transform operation effected on the 14-bit pattern by each transform element 28.
The manner in which the anchor pattern is established is as follows. Initially, the desired decimal address sequence is selected. Next, the address sequence is decomposed into a three-digit binary number comprised of the field components 22+21+20. For the 14-bit pattern example under discussion, the pattern of each binary field must be checked for compliance with two specific rules. First, the 21 sequence must repeat every four bits. Second, the 20 sequence must alternate every bit. If the sequence does not satisfy both rules, another sequence must be chosen until a sequence is found which does comply.
Once a sequence which satisfies both rules is found, the bit sequence for the 22 field is laid out across bits 6-13 of the anchor pattern (corresponding to bus 25 conductors 6-13) starting at the tap bit and proceeding in the opposite direction of the shift effected by the transform elements 28. Since the transform element 28 produces a right shift, the bit sequence is laid out from right to left, starting at the tap bit, wrapping from bit position 13 to the first bit position in the 22 field (bit 6) and finishing at bit 9. Next, the bit sequence for the 21 field is laid out across bits 2-5 from the anchor pattern, starting at the tap bit in the same manner as in the previous layout for the 22 field. Only four bits are required, since the pattern repeats after four bits (rule 1). Lastly, the two bits of the 20 field are laid out across bits 0 and 1 of the anchor pattern starting at the tap bit in the same manner as in the previous steps. Only two bits are required since the pattern alternates (rule 2). The anchor pattern is now completely determined.
The following is an example of the construction of an anchor pattern for the desired address sequence 3, 4, 5, 2, 7, 6, 1, 2. This is sequence is first decomposed into a three-digit binary number comprised of the components 22+21+20:
Next, the binary number is checked to see whether the sequence satisfies rules 1 and 2. The 21 sequence does repeat every four bits, and the 20 sequence does alternate every digit. Consequently, the bit sequence layout may commence. For the 22 sequence, the first bit in the sequence (0) is placed at the tap location (bit 10) and the succeeding bits in the sequence are laid out as described above. The resulting bit sequence is as follows:
Next, the four bits of the 21 sequence are laid out across bits 2-5 of the anchor pattern beginning at tap bit 4 and proceeding in the opposite direction of the transform. The result is as follows:
Lastly, the first two bits of the 20 sequence is laid out across bits O and 1 of the anchor pattern starting at the tap bit 1. The completed anchor pattern and the patterns resulting from the successive transformations of the anchor pattern are shown in FIG. 4. By checking the values for each row in tap position 1, 4 and 10, it can be seen that the desired address sequence has been obtained.
As noted above, with the 14-bit anchor pattern implemented in the arrangement of FIG. 1, there are some sequences which cannot be chosen due to the constraints imposed by the choice of fields 20 and 21 of diminished size from field 22. Where additional potential sequences are necessary or desirable, the size of field 20 or field 21 or both fields can be increased. For example, field 0 can be increased to a three-bit field, field 21 can be increased to a seven-bit field or the like. In order to provide all possible sequences, the anchor pattern should be expanded to three fields of eight bits each. However, where not all possible sequences are required, it is convenient to reduce the number of bus conductors or to reduce the number of the required tap connections and the size of the address bus 25.
The following illustrates how a desired address sequence which cannot be converted to a proper anchor pattern using the 14-bit implementation may be converted into a usable anchor pattern in a 24-bit anchor implementation.
The desired address sequence of 1, 3, 5, 7, 0, 2, 4, 6 is decomposed into a three-digit binary number comprised of the components 22+21+20.
Checking the binary numbers against the rules, rule 1 is satisfied since the 21 sequence does repeat every four bits. However, the 20 sequence does not alternate every digit and the sequence cannot be constructed using a 14-bit pattern.
Taking the same sequence, and selecting a 24-bit anchor pattern, the anchor pattern layout proceeds as follows. No rule checks are necessary with a 24-bit pattern, since all patterns can be implemented. Consequently, the bit sequence for the 22 sequence is first laid out in a similar fashion to that described above starting at the tap bit (position 19 in this case) and proceeding in the opposite direction of the transform (i.e., right to left), wrapping from 23 to 16, and finishing at bit 18. The resulting bit sequence is as follows:
Next, the bit sequence for the 21 sequence is laid out across bits 15-9 of the anchor pattern starting at the tap bit (position 9). The result is as follows:
Lastly, the bit sequence of the 20 field is laid out across bits 7-0 of the anchor pattern starting at the tap bit (position 4). The complete anchor pattern, and the remaining transformations are shown in FIG. 5. The sequence can be checked row by row and compared to the desired sequence.
As will now be apparent, the invention affords a highly flexible addressing arrangement for an array of devices which is particularly suited for highly configurable computer systems. Further, the invention can be implemented in a relatively straightforward and simple fashion using essentially only passive devices (the tap connections and the transform elements 28), which can be prefabricated and tested prior to installation. Once connected, the desired address sequence is selected, is converted to a suitable anchor pattern, and the desired addressing sequence is automatically provided to the individual devices connected to the information bus.
Implementation of the transform elements 28 can be accomplished in a number of ways. Firstly, the address bus can be incorporated into suitable cables, such as flat flexible cables, and the conductive patterns required to shift and rotate the individual conductor lines can be formed within the cable itself. Alternatively, the transform elements may be individual bridge connector elements incorporating the shift and rotate wire patterns and provision may be made to insert such discreet devices into the address bus path. Still further, individual segments of information bus 10 may be implemented as fixed length multiple conductor paths on substrates in close proximity to the bus devices 14-21 and, in such applications, the transform elements 28 can be incorporated into portions of such substrates along with appropriate connectors to route the address bus conductors to the input portion of the transform elements 28 and to route the output conductors from the transform elements 28 back to the address bus 25.
While the above provides a full and complete disclosure of the preferred embodiments of the inventions, various modifications, alternate constructions and equivalents may be employed, as desired. For example, anchor patterns with different numbers of bits per field, as well as different numbers of fields, may be selected and employed, depending on the address requirements of a particular application. Therefore, the above description and illustrations should not be construed as limiting the scope of the invention which is defined by the appended claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3993980 *||Dec 12, 1974||Nov 23, 1976||International Computers Limited||System for hard wiring information into integrated circuit elements|
|US4016369 *||Mar 9, 1976||Apr 5, 1977||Bell Telephone Laboratories, Incorporated||Addressing arrangements for communications networks|
|US4050098 *||Nov 17, 1975||Sep 20, 1977||Gulf & Western Industries, Inc.||Self-addressing modules for programmable controller|
|US4161634 *||Jul 31, 1978||Jul 17, 1979||Bell Telephone Laboratories, Incorporated||Count-down addressing system|
|US4253087 *||Mar 8, 1979||Feb 24, 1981||Nestar Systems Incorporated||Self-assigning address system|
|US4373181 *||Jul 30, 1980||Feb 8, 1983||Chisholm Douglas R||Dynamic device address assignment mechanism for a data processing system|
|US4626846 *||May 22, 1984||Dec 2, 1986||Northern Telecom Limited||Bus arrangement for addressing equipment units and a method therefor|
|US4727475 *||May 18, 1984||Feb 23, 1988||Frederick Kiremidjian||Self-configuring modular computer system with automatic address initialization|
|US4873671 *||Jan 28, 1988||Oct 10, 1989||National Semiconductor Corporation||Sequential read access of serial memories with a user defined starting address|
|US4963862 *||Jul 8, 1988||Oct 16, 1990||Mitsubishi Denki Kabushiki Kaisha||Terminal equipment identifier control circuit|
|US4969126 *||Jan 12, 1989||Nov 6, 1990||Mitsubishi Denki Kabushiki Kaisha||Semiconductor memory device having serial addressing and operating method thereof|
|US5065154 *||Apr 20, 1989||Nov 12, 1991||Hewlett-Packard Company||Digitally addressble electronic device with interchanged and inverted address lines|
|US5157658 *||Dec 26, 1990||Oct 20, 1992||Nec Corporation||Method and apparatus for identifying newly installed modem in dial-up network|
|US5386515 *||May 27, 1992||Jan 31, 1995||Intel Corporation||Automatic input/output address conflict resolution|
|US5437019 *||Mar 23, 1994||Jul 25, 1995||Hewlett-Packard Company||Addressing method and apparatus for a computer system|
|US5522077 *||May 19, 1994||May 28, 1996||Ontos, Inc.||Object oriented network system for allocating ranges of globally unique object identifiers from a server process to client processes which release unused identifiers|
|EP0491480A2 *||Nov 28, 1991||Jun 24, 1992||Hewlett-Packard Company||Computer addressing apparatus|
|1||*||Article by D. Meltzer, entitled "Selective Addressing of Devices Using Primitive Polynomial Transformations" published by IBM Technical Disclosure Bulletin vol. 22, No. 3, Aug. 1979, pp. 1124-1127.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US6510481 *||Nov 1, 1999||Jan 21, 2003||Lsi Logic Corporation||Method for out-of-band network communication|
|US6772232 *||Aug 26, 1999||Aug 3, 2004||Hewlett-Packard Development Company, L.P.||Address assignment procedure that enables a device to calculate addresses of neighbor devices|
|US6782437 *||Nov 14, 2002||Aug 24, 2004||Lsi Logic Corporation||Method of communicating on a network|
|US7702879||Dec 8, 2005||Apr 20, 2010||International Business Machines Corporation||Assigning alias addresses to base addresses|
|US7778188 *||Jan 26, 2007||Aug 17, 2010||Abrahams Paul C||Network diagnostic systems and methods for transmitting and receiving network messages|
|US7782845||Nov 19, 2004||Aug 24, 2010||International Business Machines Corporation||Arbitrated loop address management apparatus method and system|
|US9082462 *||Oct 23, 2013||Jul 14, 2015||Infineon Technologies Ag||Methods and systems for addressing memory with variable density|
|US20060109844 *||Nov 19, 2004||May 25, 2006||Bomhoff Matthew D||Arbitrated loop address management apparatus method and system|
|US20070005836 *||Jun 7, 2005||Jan 4, 2007||Sandeep Jain||Memory having swizzled signal lines|
|US20070189171 *||Jan 26, 2007||Aug 16, 2007||Finisar Corporation||Network diagnostic systems and methods for transmitting and receiving network messages|
|US20110289247 *||Nov 24, 2011||National Semiconductor Corporation||Autonomous positional addressing in stacked multi-board systems|
|US20150098292 *||Oct 23, 2013||Apr 9, 2015||Infineon Technologies Ag||Methods and systems for addressing memory with variable density|
|U.S. Classification||710/9, 711/202, 340/3.21, 340/3.51, 711/5, 710/104|
|International Classification||G06F13/14, G06F12/06|
|Dec 28, 2001||AS||Assignment|
|Aug 11, 2004||FPAY||Fee payment|
Year of fee payment: 8
|Sep 22, 2008||REMI||Maintenance fee reminder mailed|
|Mar 13, 2009||LAPS||Lapse for failure to pay maintenance fees|