Method for supervising participation and reducing bit errors for switched communications in a digital system.

p0001Data transmission on lines and cables is disturbing influences, such as electromagnetic fields, exposure that change the information to be transmitted by, for. Example, be one or more bits during transmission 'flipped'. In particular, since international long-haul connections to bit errors are prone to have international standards bodies - such as CCITT - set minimum requirements for bit error rates. Optical transmission lines are subject to external influences considerably less, so that the reliability of transmission is very high when using optical transmission links.

p0002In general, sets each link of a total of two parts:<ul><li>A portion relates to the transmission of information over a physical link between two nodes. These routes are increasingly implemented with optical communication cables.</li><li>Another portion relates to the through-connection of the information in the node where communications systems are installed.</li></ul>

p0003Through the use of optical transmission systems, the potential interference effect has been significantly reduced during transmission system share; while the through-switching operations in the switching networks of communication systems occur frequently in contrast to error.

p0004Of its central importance for the communication systems, the coupling fields usually are doubled or even several times. This makes it possible to switch through the information about another switching network level in case of malfunction of a switching network level. On the one hand this can be done, that there is a switching network level at rest in such a way while the other switching network level performs the active switching of the information; on the other hand can actively switch the relevant information both switching network levels.

p0005To be able to turn about the respective switching network levels, the Durchschalteweg within the matrix must first be determined and set. This means that signals arriving at the switching matrix, enter information to be through an already established, or only to be determined before the through-connect process generally input into the switch and exit via an also be determined starting again. Determining which input / output is used in the through-connection of a multi-information words information stream, thereby by evaluating the signaling information by the central control platform of the communication system that determines the Durchschalteweg through the switch and adjust. Under a central control platform is generally understood the interaction between the central and local control units of a communication system. The central control platform has memorized the current data on the current assignment of inputs / outputs of the entire switching network and the paths followed. This may at any time assign unassigned inputs / outputs as well as those associated with through-connect paths through the switching network further, still to be through information flows, the central control platform. A possible termination or termination of an existing connection is also performed by the central control platform, resulting in the release of occupied inputs / outputs as well as for use through-connect routes. The associated data must therefore constantly maintained by the central control platform, that are updated as needed. Since this is dynamically changing data, the maintenance is carried out in the memory of the central control platform; a saving to an external storage medium, such. as a hard disk is not possible because of the associated long access time.

p0006When switching of information on a switching network level sporadic and permanent errors can occur (bit error rate). They manifest themselves mostly in the fact that one or more bits are corrupted during the through-switching process, an inadmissible bit error occurs, or even a false information stream is output. Temporary or permanent false output can be the result; Sporadic errors originate in electromagnetic couplings, in thermal influences or aging components, while permanent errors from defective or adulterated latch areas, defective speech stores, faulty addressing logic for holding and voice mail, from defective or obsolete components (such., laser) or a faulty power supply results.

p0007Generally, such errors are difficult to analyze and to eliminate when they are the switching unit have no influence on the function of the switching matrix.

p0008Error detection methods for the detection and correction of bit errors in switching networks for example be realized in the form of a Mitlaufüberwachung previously. From the German Offenlegungsschrift 24 27 668, such a method is known. There is added an additional test information before the actual through-connection process in the switching matrix to the information formative information words per channel. The test information may be, for example, a parity bit. This refers to the binary checksum of the individual bits of the corresponding information word. a binary checksum of the through-connected information word again and compared their parity with the transmitted parity - After circuit switching operation is - in the case of the parity bits. Turnout both from each other, then there is a transmission error, which is held in a corresponding table memory. Thus can be over periods of time to evaluate the condition of the corresponding switching network level make. These statistical statements that switching network level can be used for switching of information flows that has exhibited the slightest bit error rate in the past.

p0009Furthermore, a method for Mitlaufüberwachung and Bitfehlerreduktion is proposed in German Patent Application P 41 28 412.7. This check information is determined for the information forming information words in peripheral units before the through-connection process. This is appended to the respective information word. Thereafter, a duplication of the movement document the test information information word and a subsequently conducted through connection via both switching network levels. After the through-switching process that supplemented the testing information information words a renewed Prüfinformationsermittlung undergo and the result is compared with the co-transmitted check information. In parallel, the information words themselves are compared bit by bit with each other. From two reviews then conclusions can be drawn as to whether a transmission error has occurred or not, or can - albeit limited - statements are made about which bit was corrupted, and be carried out in this case corrections. The problem with this procedure is that a clear statement is not always possible. In this case, on a statistical function f (s) accessed, which then forwards the information to the end user words that have appeared over the switching network level, which has come in the past, the better statistical quality data. In this case, however, there is a certain probability is the danger that the corrupted information word is forwarded to the end user.

p0010Furthermore, in this method, an accurate fault location possible only in coarse scale. Although statements to be taken from that switching network level that caused the transmission error, but a more detailed limitation is not possible.

p0011Serious impact also have errors that cause faulty switching through the switching network, ie that a penetration at a given input of the switching network information stream not the switching network relies on the intended output, but because of a sporadic or permanent error occurred during the adjustment on an incorrect output to any end user is forwarded. Such errors can not be detected centrally; often provides only the end user the faulty switching through firmly.

p0012Furthermore, the fact that all the dynamic timing data are stored in the memory of the central control platform, cause problems. So relevant parts of the central control platform are also lost all stored therein setting of the switching network and the data input / output assignment as in the case of a total failure. Outsourcing to an external storage medium such as a hard disk is not possible, because the setting must be updated dynamically and therefore the read / write operations would take on the hard drive to time-consuming, especially in times of high traffic load. This means that all existing connections are broken and the associated data memory must be initialized or need to be updated in a complex manner to the hardware settings.

p0013The invention is based, to minimize the effect of errors in particular in view of an erroneous switching through to end stations in the switching networks of digital communications systems as low as possible the task.

p0014The object is achieved by the method described in claim. 1

p0015Claim 2 relates to a first embodiment of the Prüfbewertungsinformation. It is a material derived from a binary checksum check bit. The binary checksum is formed from the information word and the appended Quellennummernbit. However, other Prüfbewertungsverfahren are conceivable; so for example, the binary checksum be approximately formed only with every other bit. Furthermore, the Prüfbewertung can with the aid of mathematical functions - are formed - such as multiplication or division.

p0016The source number is encrypted in a 1-bit information. To grasp the full source number, so these bits must be stored as long and filled until the source number is complete. The source number is a characteristic number for the place where the information flow has occurred in the communication system. This point of entry is formed by the input side interface elements. The through-switching process itself is characterized in that each point of entry, is associated with a selected output, which is formed by the output side interface elements. Since the source number is switched along with the information stream can be observed on the output side interface elements if the through-connection input / output was correctly through the switch with respect to the assignment.

p0017Claim 4 relates to a first embodiment of the invention. It is contemplated that the source number is filed with the connection in the output side interface elements via a communication channel between the central control platform and interface element. If the express warranties provided by both switching network levels information flows differently and contain different source numbers on the output side interface elements can independently select and output the correct through-connected information current. A second embodiment of the invention is defined in claim fifth It is provided that in the event of permanent inequality of products supplied by the respective switching network levels information flows and a mismatch of the associated source numbers a message from the output side interface elements to the central control platform is transferred. This decides on the basis of its internally stored connection data and the information contained in the message on the destination number and source numbers, which the through-connected to the interface element information streams has been transmitted correctly, and divides the result of the output side interface elements. A third embodiment of the invention is defined in claim sixth A transmission procedure transmits the data stored in the central control platform source numbers cyclically to the respective output-side interface elements. This can therefore decide in the case of a permanent discrepancy between the through-connected information streams themselves, which have been transmitted the information flows properly. Such an approach combines 'dynamic' aspects with a quick response in case of error.

p0018In claim 7, the application of the invention to asynchronously set switching matrices is defined. It must be noted that the application of the inventive method is carried out only after a sufficiently long period of protection, since it is ensured only after this period that both switching network levels are switched through. Furthermore, the inventive method can be applied to individual switching network components of a switching matrix. Such a configuration ensures that the relayed information streams and bit errors are adjusted already within the switching matrix.

p0019An advantage of the invention is in particular that after a total failure of the central control platform, there stored connection data can be regenerated directly from the source numbers available in the output side interface elements that the maintenance of existing connections much easier.

p0020The invention is further illustrated with reference to three exemplary embodiments illustrated figuratively.

p00211 shows a switching network SN of a digital communication system KS. The switching network SN is doubly designed and divided into two switching network levels SN₀, SN₁. As an interface between the communication network and the communication system KS act on the input side and output side interface elements DIU (A), DIU (B).

p0022In the input-side interface elements DIU (A), a penetration in the communication system KS, from a plurality of information words I^(O)... I^(N) existing information flow before the through-connection process through the switching network SN with a for the respective information word I^(X) characteristic additional information ZU provided.

p0023Generally the input of the input interface elements DIU (A) to which the information words are entered, a source number Q is permanently assigned. Furthermore, by analyzing the signaling information a corresponding to the source number Q target number Z, ie the output, leaving from the information words the output side interface elements DIU (B) after the through-connection process again determined. The source number Q, and the target number Z are stored in tabular form in the central control platform CP. From two sizes, the central control platform CP determined with the aid of a suitable algorithm the Durchschalteweg the switching network SN and sets him. This puts the central control platform CP determine which path through the switching network SN of durchzuschaltende information stream I⁽⁰⁾ ... I^(N) must use.

p0024In the exemplary embodiment it is assumed that an information word I^(X) consists of 8 bits. In the input-side interface elements DIU (A) the incoming information words I^(O)... I^(N) first provided with, the respective input of the input interface elements DIU (A) dedicated Quellennnummer Q, and subsequently as the ninth bit of the information word I^(X) bit over several information words appended away. Then a parity formation over the information word I is^(X) and the respective Quellennummernbit Q, the result of which is recorded as the parity bit P in Bit 10th Each information word I^(X) So prior to switching operation with a 2-bit wide, consisting of the parity P and Q Quellennummernbit Additional Information TO applied. Thereafter, a mirror image splitting of the information word I^(X) together with the additional information ZU into two identical structures, each information word I^(X) with its additional information ZU via the respective switching network level SN₀, SN₁ is conducted taking into account the framework set by the central control platform CP through-switching path.

p0025In FIG 2 a the procedure of the invention set forth in detail.

p0026In the output side interface elements DIU (B) after the switching through the information words I^(X) along with her co-transmitted additional information subjected TO re-evaluation. For a place in a first review a new parity check both circuit-switched to the Q Quellennummerbit supplemented information words I^(X) instead of. The result of those parity check is compared with the co-transmitted parity bit P. In a second vote, both through-connected information words I^(X) compared bit by bit with coassigned Quellennummernbit Q together. A combination of both ratings can draw conclusions about it, that the information words I^(X) was transferred falsified. Not match the information words I^(X) both switching network levels SN₀, SN₁ is the information word I^(X) forwarded to the end user, which is consistent in terms of its co-transmitted parity information P including Quellennummernbit Q. If this condition is not satisfied, the information word is I^(X) forwarded to the end user, which has passed through the switching network level having the better statistical quality data. These are recorded in a statistical function f (s).

p0027FIG 2b shows the inventive method in the event of a discrepancy the source numbers Q.

p0028In the present embodiment will now be assumed that by the presence of temporary or permanent error a false switching through in a switching network level is present, so that in the downstream interface element DIU (B) two permanently different, in terms of their parity but internally consistent information streams I⁽⁰⁾... I^(N) and J^(L)... J^(K) arrive. This can be determined by the work accomplished in the second review bitwise comparison. Thus, it is considered in the present case the fact that the processing performed in the initial evaluation parity formation no conclusions about permits, which of the two routed through switching network level and SN₀ SN₁ information streams I^(X) was transferred falsified.

p0029For a more accurate determination another criterion must now be consulted. In a first embodiment of the invention, the output-side interface elements DIU (B) using a communication channel between the central control platform and the output-side interface elements DIU (B) shall be communicated to the respective outputs associated source numbers Q per connection. Thus, the source numbers Q currently in the output side interface elements DIU (B) are stored. In the above-described case of failure can thus independently to determine and correct the through-connected information stream the output side interface element DIU (B). Error data may be stored or peripherally reported to the central control platform CP.

p0030In a further embodiment of the invention send the output side interface elements DIU (B) in the case of difference between the supplied via both switching network levels SN₀, SN₁ information streams I⁽⁰⁾... I^(N), J^(L)... J^(K) a message to the central control platform CP. This includes the destination number and the source numbers Q of the switched-through via both switching network levels SN₀, SN₁ information streams I⁽⁰⁾... I^(N), J^(L)... J^(K). The central control platform CP determined for both source numbers Q using its internally stored call data associated destination numbers. This can be determined in a simple manner, which was transferred to the two through-connected information flows properly. The result of this review will be informed by the central control platform CP the output side interface elements DIU (B) on a message to select the correct information stream.

p0031In a third embodiment of the invention, the data stored in the central control platform CP source numbers Q are cyclically transmitted to the corresponding output side interface units DIU (B) and stored there by a transfer procedure U. This exists in the output side interface units DIU (B) an image of the performed in the central control platform CP source numbers Q. In the described case of a fault so can the output side interface elements DIU (B) likely decide immediately on its own, which the delivered of two switching network levels SN₀, SN₁ information stream is to be forwarded to the end user. This procedure is for digital communication systems both under dynamic aspects and from the point of immediate failure activation very beneficial as the decision-making process can be largely completed locally without interrupting other software processes without providing simultaneously high demands on the processor dynamics of the central control platform CP.

p0032In the event of total failure of relevant parts of the central control platform CP therein stored connection data are lost. The inventive method in the output side interface units DIU (B) existing source numbers Q are to update the data connection to the central control platform CP available. This means in practice that the central control platform CP can be taken very quickly back into operation, without interrupting existing connections.

p0033The inventive method is also applicable to asynchronously set switching network levels. They are characterized in that the through-connection of information flows through the respective switching network levels takes place at different times and thus in the output side interface elements DIU (B) briefly pending information streams from different sources. Therefore, it should be noted that a sufficiently long period of protection must be established, in which causes the difference of traffic routed via both switching network levels SN₀, SN₁ information streams no error response in the output side interface elements DIU (B). Only after that period of protection, the inventive method can be carried out.

p0034In FIG 3 is 1 known from FIG switching network SN with the switching network levels SN₀, SN₁ and the input side and output side interface elements DIU (A), DIU (B) shown in greater detail.

p0035A switching network level SN₀, SN₁ consists of several - in the present embodiment three - switching network components. For the switching network level SN₀ these are the switching network components SN₁₋₀, SN₂₋₀ SN₃₋₀ and the switching network level SN₁ the switching network components SN₁₋₁, SN₂₋₁, SN₃₋₁. For safety reasons, the individual switching network components are mutually crossover. This means that in case of failure identified in the switching network components switch units S can be allocated to the respective other switching network component.

p0036The inventive method can also on the through-connection of information words I⁽⁰⁾... I^(N) be applied in individual switching network components. This means that already in the switching network components SN_xy (X = 2.3; y = 0.1), it is decided which information word I^(X) correctly on the front of it switching matrix component SN_{(X-1) -y} was turned on. Additionally, here, by evaluating the source number Q in the switching network components SN_xy be determined in the manner described where the transmission error has occurred. In this case, the fault localization would be even more precisely carry out than when using the method according to the invention the switching network SN as a whole.