« PreviousContinue »
PROCESS FOR DETERMINING THE QUALITY PARAMETERS OF TRANSMISSION LINK FOR DIGITAL DATA STREAMS HAVING A CELLULAR
BACKGROUND OF THE INVENTION
The present invention relates to a process for determining quality parameters of a transmission link for 1Q digital data streams having a cellular structure, in which each cell has a useful field. This useful field can be freely occupied with data such as, a test signal to be transmitted on the transmission link and to be evaluated upon receipt. 15
Besides physical transmission lines, transmission links for digital data streams include other components, such as switching equipment, standardized terminals, terminal adapters, as well as terminal equipment (for example, telephones or fax units). In digital communication 20 systems, digital data streams are transmitted at a high speed in an organized form over the transmission links. To evaluate a digital communication system, knowing quality parameters of this type of transmission link is important. Quality criteria, may include, for example, 25 whether portions of the data stream have been lost on the transmission link, what transit time the data has required on the transmission link, whether the sequence of the data was retained, and whether single binarycoded data (bits) were transmitted incorrectly. From 30 communications technology, testing transmission links by injecting a test signal on the input side and by evaluating the test signal after it has been transmitted through the transmission link is known.
To accomplish this, U.S. Pat. No. 4,441,192 discloses 35 a process in which a test signal having a pronounced, peak autocorrelation function is applied to a transmission link on the input side. A compensation circuit is arranged on the output side, downstream from the transmission link to be analyzed. After being appropriately adjusted, this compensation circuit is used to compensate distortions of or disturbances in data signals to be transmitted. Those distortions or disturbances result from transmission characteristics peculiar to the trans- 45 mission links. In the known process, to adjust the compensation circuit; the impulse response of the transmission link to be analyzed is determined from the crosscorrelation of the test signal and the output signal. In this manner, the downstream compensation circuit can 5Q be optimally adjusted, even given transmission links having unknown or time-variant transmission characteristics, so that a substantially undistorted data signal is received on the output side.
In principle, the test signal can preferably be let-in in 55 empty areas of a digital data stream to be transmitted under field conditions. However, one must be able to reliably recognize the measuring signal that has been let into the digital data stream.
The test signal can preferably be placed in empty go areas of a digital data stream to be transmitted under field conditions. However, the ability to reliably recognize the test signal that has been placed into the digital data stream necessary.
The present invention seeks to provide a process for 65 generating and evaluating a test signal for determining the aforementioned quality criteria and which is reliably recognizable as a test signal.
SUMMARY OF THE INVENTION
The present invention meets the aforementioned goal by providing a process including steps of:
using a sequence of test cells as a test signal, the useful fields of the test cells being occupied, in each case, with a test pattern consisting of a number of binarycoded data (bits), wherein the auto-correlation function of the test pattern roughly corresponds to a Dirac pulse and wherein, within the sequence of test cells, the position of the test pattern in the useful fields of the test cells is varied without repetition;
transmitting the sequence of test cells through the
transmission link; after being transmitted through the transmission link, correlating each cell of the data stream with a reference pattern identical to the test pattern or the inverse of the test pattern; and evaluating the maximum of the obtained cross-correlation function with respect to its position in relation to the start of the particular cell and with respect to its height. A Dirac pulse is understood to be a theoretical pulse with an infinitely high amplitude and an mfinitely small duration having a surface area 1 (compare, for example, Otto Mildenberger, Grundlagen der Systemtheorie fur Nachrichtentechniker Fundamentals of System Theory for Communication Engineers, Hansa Publishers, 1981, pp. 48-50). One advantage of the process according to the present invention is that a test cell can be recognized in a digital data stream under field conditions, without needing to transport additional data along with it in the data stream to indicate that the data is a test cell. In this manner, depending on the structure of the cells, in the best case, the entire cell can be tested, because no reserved cell locations are required for such data which characterize the test pattern. Furthermore, bit errors can be established during transmission by evaluating the maximum of the obtained cross-correlation function with respect to its height. In a relatively wide range, there exists, an inversely proportional linear correlation between the height of the maximum and the number of bit errors (i.e., when the height of the maximum is less than an expected threshold value, bit errors exist and the lower the height of the maximum, the more bit errors exist). Therefore, depending on the threshold value for the height of the maximum, test patterns can also be recognized, which have one or more bit errors. However, the danger also increases (provided that the digital data stream also has cells transmitted under service conditions) that a cell transmitted correctly under service conditions contains a data pattern in its useful field, which corresponds to the test pattern or to the test pattern corrupted by bit errors. However, as the length of the test pattern increases, this danger can be considerably minimized. In addition, evaluating the position of the maximum in relation to the start of the evaluated cell permits the number of the test cell to be determined without requiring data for the number of the test cell. Otherwise, such test-cell number data would have to be additionally transmitted, and therefore would not be able to be tested and would reduce the useful field. The location of test pattern within the test cell is shifted in each test cell of the sequence with respect to the start of the test cell which enables systematic errors to be additionally recognized in an advantageous manner. When all variations of the test pattern in the useful fields of the
test cells, taken together, cover the entire useful field of the test cell, each location in the useful field is occupied at least once with a binary-coded data of the test pattern.
If a particularly large number of test cells is to be 5 used, then, according to the present invention, the test signal is supplemented by other test cells, whose useful fields are partially occupied, in each case, with a test pattern that is the inverse of the test pattern of the original test cell. In this manner, the number of test cells that 10 can be recognized without additional data is doubled. If a still further increase in the number of test cells is desired, then an additional sequential number, which is input in binary form in the useful field, is applied. This sequential number should preferably have an exponen- 15 tial form. This means that after reaching a certain number (for example 100) of test cells, whose number can be recognized through the cross-correlation, the exponent is set and the varying of the test cells starts from the beginning again. Thus, in the example, the number of 20 test cells is increased to 200.
To carry out the process according to the present invention with a measuring signal having a very large number of different test cells, a further development of the process according to the present invention includes 25 steps of:
in addition to the test pattern provided in each test cell in one section of their useful fields, providing in a second section of their useful fields, a second test pattern consisting of a number of binary-coded 30 data (bits), wherein the autocorrelation of the second test pattern corresponds roughly to a Dirac pulse;
varying the position of the second test pattern in the second section of the useful fields of the test cells 35 within the sequence of test cells when the nonrepetitive variation of the position of the first test pattern is exhausted in the first section of the useful fields in the sequence of the test locations;
after running through the transmission link, correlat- 40 ing each cell of the data stream with a reference pattern identical to, or the inverse of, the second test pattern; and
evaluating the maximum of the obtained cross-correlation function with respect to its position in rela- 45 tion to the start of the particular cell and in relation to its height.
This specific embodiment of the process according to the present invention is particularly advantageous in that a test signal having a very large number of different 50 test cells may be used without having to write a number into the header or into the useful field of the particular test cell to characterize the particular test cell. As a result, simultaneously recognizing bit errors is possible by evaluating the maximum of the obtained cross-corre- 55 lation function.
To enable the number of test cells used in the test signal to be increased according to a numerical factor by means of the additional test pattern, another advantageous refinement of the process according to the pres- 60 ent invention foresees varying the position of the additional test pattern, without repetition, by a same interval in each case, after exhausting, each time, the non-repetitive variation of the position of the first test pattern in the first section of the useful fields in the sequence of the 65 test cells. Thus, if a certain number of test cells is reached by exhaustively varying, without repetition, the position of the first test pattern in the first section of
the useful fields in the sequence of the test cells, then, by changing the position of the second test pattern, an equally large number of test cells can be achieved by again varying, without repetition, the position of the first test pattern in the first section of the useful fields. A further doubling of the range of available numbers of different test cells can be advantageously achieved by supplementing the test signal with other test cells, each of which are provided with a test pattern that is the inverse of the second test pattern in the second section of their useful fields.
The present invention will be clarified in the following in light of its application in a broad-band communication system (B-ISDN) based on the drawings.
FIG. 1 illustrates the temporal occurrence of certain cells of certain data streams to clarify the differences between the asynchronous transfer mode and the synchronous transfer mode.
FIG. 2 is a schematic of an arrangement for implementing the process according to the present invention.
FIG. 3 illustrates a possible structure of a cell of a digital data stream.
FIG. 4 illustrates a test cell structure.
FIG. 5 illustrates the shifting of a test pattern in the available useful field based on the test cell number.
FIG. 6 schematically illustrates the evaluation of the cross-correlation function.
FIG. 7 is a block diagram of the structure of the ATM analyzer shown in FIG 2.
FIG. 8 illustrates the conditions when other test cells are used, whose useful fields are partially occupied in each case with an inverse test pattern.
FIG. 9 illustrates various test cells of one measuring signal with variably shifted test patterns as well as second, variably shifted, test patterns.
FIG. 10 is a block diagram of an exemplified embodiment of an analyzer provided.
The process according to the present invention can be applied, in particular, to the so-called asynchronous transfer mode (ATM) for the wide-band ISDN (Integrated Service Digital Network) process. The ATM method appears to be a universal solution for transmitting different services covering a broad spectrum of bit rates.
Simultaneously handling different services over one link connection, places ever greater demands on communication networks, such that developing a new transmission technology is necessary. Such a new network should have a flexible transmission rate and should be open to low transmission rates. The wide-band ISDN, with a transmission capacity of 140 MBit/s is well suited for video transmission (for example, for video conferences), or for transmitting large quantities of data, as is required in the magazine and periodical industry. The ATM transmission method can be advantageously applied in B-ISDN technology for all services known today, and its flexible bit rate makes it a viable system for future services.
In the future, the B-ISDN will be introduced to expand on the previous ISDN (Integrated Service Digital Network). Its important distinction from ISDN and SONET (Synchronous, Optical Network) is an asynchronous transmission mode.
FIG. la shows an irregular occurrence of cells in the data streams 1,2,3,4 during the asynchronous transfer mode ATM of the B-ISDN, in contrast to the regular