FI97842C - Configuring a digital cross connection - Google Patents

Description

97842

Configuring the digital crossover - Configuring the digital crossover

The invention relates to a method according to the preamble of claim 1 for configuring a cross-switch for digital transmission lines in the case of change-over situations.

The synchronous digital hierarchy (SDH) comprises a large set for transmitting time-division signals in a telecommunications network, the backbone transmission network of which is evolving into a remote-controlled crossover network. The first level of SDH signals is a Synchronous Transport Module (STM-1) with a transmission rate of 155.520 Mbit / s. The STM-1 basic frame consists of bytes (8 bits) including 2430 control blocks in the frame; in this case, 63 TU 12 (Tributary Unit) 2 Mbit / s signals are transmitted in the STM-1 frame, which may include a 2 Mbit / s signal of a standard 30-channel PCM system. Each byte of the frame forms a 64 kbit / s channel. SDH signals, i.e. transport modules, are generated from the signals of the subsystems by interleaving bytes.

An SDH cross switch (DXC, Digital Cross Connect) can transmit traffic between different SDH levels as well as switch traffic between different signals. In addition, it must be able to handle the flexible reconfiguration of the network, i.e. the re-disconnection of connections, and guarantee the rapid introduction of va-ray connections in the event of a network failure.

20, Digital cross-linking has been extensively studied to find an architecture that meets optimal conditions. A structure that well meets the conditions of capacity, non-blocking and feasibility is the TST (Time Space Time) structure, i.e.: * as described in, for example, our patent PCT / FI / 00174 (or equivalent FI-921834).

• • I

. * This patent sets out in some detail the general principles of a TST cross-switch. Although the method disclosed in PCT / FI / 00174 works quite well: * · there is a need for even more efficient switching, especially in larger cross-switches. for a more control method.

It is an object of the invention to provide a method for cross-coupling between two given points of a digital signal: "": a method by which non-blocking coupling can be implemented and the known drawbacks and disadvantages avoided.

•. ‘: This task is solved in the periodically implemented method according to claim 1 by a configuration calculation method. Other preferred embodiments of the invention are set out in the other dependent claims.

2 97842

The signals to be switched are preferably multiplexed sub-signals of high-speed signals, i.e. in the SDH system this means that the sub-signals are mainly 2 Mbit / s VC-12 virtual containers, the main signals being 155 Mbit / s STM-1 signals. The Game algorithm according to the invention has better properties than other known point-to-point switching algorithms. The advantages are lower data memory requirements and faster performance. Reasonable memory requirements and fast performance are very important as the crossover grows large, especially when it exceeds the current size of 16 * 16. The algorithm also comprises a simple and straightforward reordering routine, which forms the main part of the algorithm. The reorganization routine is thus easy to implement and well optimized so that it is capable of fast computational performances.

In fact, the power of the algorithm can be seen in the fact that the algorithm rearranges only those connections for which rearrangement is necessary. This is accomplished in part by setting the coupling problem in a new way. In this case, the coupling problem is presented in Brochure-15 as an imaginary matrix. The algorithm keeps track of all rearrangements made to the connections to meet the need for cross-connection. Once all the rearrangements have been made, the accounts provide all the necessary information needed to make the cross-connection based on the control information of the switches.

In the method according to the invention, difficult coupling situations are solved in the recursion-20 step, in which the location of the coupling in the imaginary matrix is randomly forced to a new location, after which the operation is continued according to the basic method.

The algorithm of the invention can also be used to route signals between two points through a three-stage Benes switching field.

The invention will now be explained by way of examples with reference to the accompanying drawings.

. Figure 1 shows a simplified example of how to view a cross-connect situation with::: only 5 time slots instead of the usual 64 time slots.

• ', · Figure 2 shows the imaginary matrix and the connections between the input ports and the output ports. In this reduced example, the horizontal axis has input ports with input time slots-30, and the vertical axis has output ports with output time slots.

. ; Figure 3 is a simplified flow chart of a game algorithm according to the invention.

Figure 4 illustrates a recursion step according to the invention.

3,97842

Fig. 5 shows a situation when a first line feed has been implemented in the method of the invention, and Fig. 6 shows a situation of a switching matrix when the first submatrix is rearranged according to the invention by means of a game algorithm.

The efficient implementation of the configuration calculation of the cross-switch according to the invention is based on the feature of the TST cross-switch that the new configuration can be solved one time slot at a time. In other words, when the input side time switch k, k * k through the space switch can be found in a non-blocking state for an arbitrary time between, it can be assured that the other time slots n + 1, n + 2, ..., m may be solved by the same principle. Thus, the configuration of the cross connect can be calculated one time slot 10 so that the non-blocking connection configuration is always the search for the remaining time slots n, n + 1, n + 2, ..., m, the input side time switch and the space switch.

The mode switching of the specified time interval is unobstructed when all outputs of the mode switch are active. In other words, the input side time switch is connected to all the time slots or channels so that the state of the switch inputs aikakvtketyt all the time slots of guide 15 to the various outputs. Unless there is a time slot coming that should route to a particular output, then of course this output does not need to be used. Thus, the space in front of a switch the input side of the time switch is to distribute the channels evenly by time slots so that the space switch can switch them to the correct, desired outputs. More specifically, the mode switch must avoid your blocking, where there would be more than 20 one channel routed to the same slot in the same time slot. With the method according to the invention, the coupling matrix is formed in such a way that the cross-coupling operates unimpeded.

The switching method calculates the configuration of the cross-switch one time slot at a time, i First the input channels, i.e. the bytes in the frame, are divided into as many groups as there are cross-switched signals, i.e. time periods, in the transport module of the interface. For example, support • 28 run I time switch connected to the SDH.. The STM-1 line has a 2-Mbit / s signal-periods of time to 63 transmission rate of 155 Mbit / s, a subsystem containers (e.g. TU 1, • Tributary Unit, which can not. contains a standard 30-channel PCM system V '2 Mbit / s signal). The position of each time period selected in accordance with the invention in the frame is obtained directly by means of or from a standard pointer (pointed).

; 3T) · chem. Correspondingly, the line according to the plesiochronous transmission hierarchy (PDH) has 2 Mbit / s signal time periods 64 at a speed of 140 Mbit / s. Within each group defined in this way, the cross-connect paths are resolved so that it is realized. . the switching configuration also means that the slots that have not yet been calculated can also be switched. The solution of the whole cross-connection field is thus obtained by the time-periodic method according to Fig. 3.

4,97842

The invention can be applied in cross-switches using one-way or two-way traffic switching. In a bidirectional cross switch, the switching configuration of the second direction can be formed in an analogous manner, e.g. as a mirror image.

The coupling problem is presented as an imaginary matrix shown in Figure 2.

5 It should be noted here that Figure 2 is simplified to correspond to the matrix (4 * 5), while the matrix corresponding to the practical switching field would be Koolaan (16 * 63). The small memory requirement of the algorithm is due to the fact that the only data fields actually needed are data corresponding to the points on the matrix axis.

The algorithm keeps track of all reconfigurations made to the connections to meet the need for cross-connection. Once all the rearrangements have been made, the accounts provide all the necessary information needed to make the cross-connection based on the control information of the switches.

The signals to be connected are mainly 2 Mbit / s VC-12 virtual containers, as mentioned above. The connection of the main signal (STM-1) as a whole between the input and output ports 15 need not be considered here, since such connections can be made immediately without any calculation. Thus, the problem is to connect VC-12 signals from incoming main signals to other outgoing main signals.

The general problem field and starting point of the algorithm is a table containing all connections to be connected via cross-connection. The size of the table is 16 * 63, which 20 ,: represents the number of ports and time slots, respectively. Location of table elements • '. · In a row and column, and the value of the element indicates the connection. Figure 1 shows; : simplified table for clarity, only 4 * 5 in size, ie • 4 input ports and 5 time slots. The values of the table elements in the input ports mean • · · ·. '. desired output ports and time slots for the available signals. The connection to be made over the cross-switch, for example from the time slot 2 of the input port 1, must be made to the time slot 3 of the output port 4. In the table elements, the input ports form a row and time slots in a column, and the output port and output time slot are marked in the element. in the form: (port, time slot). In the example of Figure 1, the switching situation is thus considered i t * from the input side. Correspondingly, in the table from the time slot 3 of the input port 2, the connection to the time slot 2 of the output port 1 is resolved, which can also be denoted by 2.3 1,2.

... · A prerequisite for successful configuration in a TST cross-connect architecture is that '·. ‘That the connections must always be arranged in such a way that, within the respective time slot, the input signals are routed to the individual individual outputs of the state switch, i.e. no two or more signals may enter the same output port in a given time slot. This requirement is inherently simple, but the performance of algorithms used in larger cross-connects differs greatly in the way the algorithm resolves the configuration and how long this calculation requires.

Let us now consider the mate-5 underlying rationale of the algorithm of the invention.

Combinatorial mathematics can be applied to the conditions of the TST architecture. The theory includes a definition of System of Distinct Representation (SDR). The system SDR describes N sets (sl, s2, ..., sN) whose distribution of elements is such that a solution can be found for the elements in which each element is different. The sets sX can have the following contents: sl: (1,3,5) s2: (2,4) s3: (1,2) s4: (3,4) 15 s5: (5) Such an SDR made with sX solution may include the following elements: 1, 4, 2, 3, 5 (i.e., element 1 of s1, element 4 of s2, etc.). The condition for forming the SDR of a system is presented as Hall's theorem: “In sets sX there exists an individual solution if for each value k we obtain k different solutions in any combination of sets 20 k”.

From the definitions of the above sets, it can be seen that the theorem holds. . for these sets sX (x = 1 ... 5). When the union of the sets sl and s2 is transformed, 5 • · · representatives are obtained (elements 1, 2, 3, 4 and 5), which is more than the number of <· sets forming the union. If the set s4 contained only element 5, it would make SDR 25 impossible, which is also seen from Hall's theorem. The alliance of the forces s4 and s5 would then contain only one representative, which is less than the number of the forces in the union. *.

«/ *« · * ..; This mathematical situation can be applied to the TST architecture because the sets sX «· can represent input lines. The elements of the sets can then mean signals, which must be routed to the desired output lines. Here it should be noted that the mathematical. · The order of the time slots is irrelevant, because the essential thing in the TST structure is to arrange the connections so that they use different ports in the state switch. The timers take care of the correct order outside of this definition.

6 97842

It can be shown that Hall's theorem holds for the TST architecture because the input and output sides are the same size. The problem is not the non-blocking of the conditionally unobstructed TST architecture, but how to describe an algorithm for finding the correct switching configuration, i.e. how best to find a solution where for each 63 time slots 5 the signals use different ports over the state switch.

One possible algorithm is found by testing the validity of the Hall theorem. In such an algorithm, one element is simply extracted from one of the sets sX, as long as this element is not already part of the solution. The algorithm tests the elements in the set until it finds a new solution. This is continued for all sets, 10 until all sets are solved with individual elements.

If no solution is found among all the elements of a given set, then a recursive operation must be performed. Recursion is initiated by selecting one element from the unresolved set. An overlapping element in another set must be removed from an already made solution, and a new solution must be found for that overlapping set. If there is an element in the overlapping set that increases the number of solutions, then it is a new solution. If, on the other hand, no new solution is found, then the re-course is continued, but now using a new unsolved set as a starting point. Recursion is as described above except that the selection of the next element cannot be applied to the previous overlapping element. The recursion calculation is continued until the number of solutions increases with the new solution.

The following is an example to illustrate the test algorithm. The above sets can be used, i.e.: sl: (1,3,5); s2: (2.4); s3: (1,2); s4: (3,4); s5: (5).

• · · • · ·

From the set sl, element 1 is selected for the solution. Next, element 2 is selected from the set s2, and then an element from the set s3 must be selected. The set s3 contains the elements •. 1 and 2, which already form part of a step-by-step solution indicating that recursion computing is required. Let the element 1 of the set s3 be chosen as the new solution, in which case the element 1 of the set s1 must be removed from the solution. In this case, the set sl must have • · ·:. 'new solution, and element 3 is suitable for this, because it increases the number of solutions, * 3Ör, ie the previous solution (s2 (2), s3 (l)) increases to the form (sl (3), s 2> 2), s3 (l) ). In this case, the recursion is stopped. In the next step, the element 4 of the set s4 is selected and then finally the element 5 of the set s5. Thus, the final solution becomes: • (s 1 (3), s2 (2), s3 (1), s4 (4), s5 (5)).

7 97842

When there is an overlap between the solutions, i.e. a collision or an overlap, the algorithm aims to start from the set that does not yet have a solution. This premise cannot always be considered advisable, as the algorithm may remain in a loop between the same sets until it finds the element-5 tin of the previous sets by which the solution grows. A better starting point is then to select an element of the previous set sX with which a new solution can be obtained. Other solutions are found by going through the algorithm in the opposite direction.

Next, the implementation of a suitable recursion operation is considered. In the event of a collision, the set sX is selected, the element of which produces an additional solution. The previous selection for this set of sX 10 must be deselected. The free solution is now selected from the second set, and the element previously selected for it is removed from the solutions. This selection and deletion operation in one set is continued for all sets in the total solution until the deleted element is found in the set in which the recursion was started, i.e., the set that previously contained only those elements that were already present in the other sets.

Let’s look at the recursion situation above as an example and use a softer method. Then there are two elements in the set s3, but neither can be selected. Both s1 and s2 can increase the solution. In the set sl there are still elements 3 and 5, and in the set s2 there is still an element 4. Let us now take the element 3 of the set sl 3. In this case, the previously selected element 1 of the set sl is deleted, whereby the ele-. ; item 1 can be selected in another set. Examining the set s3, it is seen: that this set also includes element 1, which is thus chosen as the solution of the set s3. Thus, the total solution is: (s 1 (3), s2 (2), s3 (l)).

i.i: The disadvantage of this method or algorithm is its weak power. The time required to increase the solutions, i.e. to form a new connection across the mode switch, takes practically too much time. Increasing the number of solutions from k-1 to read-! . *. the number k can in the worst case mean k new choices. Theoretically, an Ek phase can then be required for the solution of the entire term. In the worst case, the theoretical efficiency of the selection process is (100% * 16) / Ek = 11.76% when kmax = • 30 16, ie when the cross-connect size is 16 * 16.

• ·

In the game algorithm according to the invention, a switching request table is processed, as in Figure 1. It should be noted here that the table formed according to the output ports must have a switching situation in which the elements of the table show the inputs of the cross-switch. Thus, in the table, the row corresponds to the output port, the column to the output port time slot, 35, and the content of the element corresponds to the input port and time slot. From the connection request table, the 8 97842 fields are recorded in two auxiliary tables that describe the connections from the inputs to the outputs.

The main task of the algorithm according to the invention is to solve all connections at once without dividing the connection request table into smaller parts. Said two auxiliary pause-5 locks, i.e. the input table and the output table, are placed in two sets, which are arranged in time slots. These two sets form the two axes of the imaginary matrix. Figure 2 shows an example of these sets arranged in an imaginary matrix based on the data in Figure 1. The x-axis of the matrix corresponds to the request set of input ports and the y-axis to the request set of output ports.

10 In Fig. 2, the inputs are on the x-axis so that the first or top line indicates the input port and the second line indicates the time interval of the input connection. The third line of the X-axis indicates the output port and time slot (port, time) of the signal destination in the input port / slot. The first, or most left-hand column, indicates send; Sport and the second column of the output slot interval, which is presented as the output interface request. The third column 15 indicates the desired content of the output interface, i.e. the connection in the form of a source: input port, input time slot. The structure of this output series is thus the same as in the result series. In addition, the content of the output set shows the same connections as the result set, only the aspects differ from each other. In the rest of the table, the imaginary matrix, the notation ‘1’ indicates the coupling. Thus, each input is associated with one of the outputs at a given intersection of ribs and columns marked “1”.

'' When the size of the matrix is N * N, where N = 4 in Figure 2, the matrix can be implemented as sub-matrices of the total matrix. In this case, one submatrix is formed for each time slot of both the input and output switching series. In the total matrix, the submatrices • · · · * form a diagonal that runs from the upper left corner to the lower right corner. The two: the first 25 such matrices are shown in stronger circumferential lines in Figure 2. These matrices show the connections across the state switch in each time slot. These mat-:. rice also corresponds to the solution of the whole SDR system.

• · * * * Figure 3 shows the main loop of the algorithm. In the main loop of the algorithm, all switching fields are assembled into submatrices. The complete solution is achieved when all • · · * 30 'connections have been transferred to the sub-matrices. Transfers are performed by swapping rows or columns in pairs. Transfers are made in the input and output series, but also in two additional series, not shown in Figure 2. These additional series are the same size as the input and output series, but contain only one element. The function of such a series of states is to indicate the transfers required in the first and second time switches of the TST crossover. Thus, these two sets of states are said to be the first states. «U i Kilu III, i H i 9 97 842 in series and the second series of state corresponding to the input side time switch and an output side time switch.

The transfers are controlled so that the subject of the current transfer is always the submatrix. Connections outside the Alimat rice, i.e. the 5 elements marked “1” by the imaginary m <iris, always have two alternative target sub-matrices, one in the direction of Tom and one in the direction of the column. Transfer to a submatrix is allowed if the connection is the only one that uses the relevant input port and output port, i.e. when a separate or individual connection is obtained in that time slot. The transfer in the row direction precedes the transfer in the direction of the column, due to the starting order of the main loop. Because the transfer is in fact an exchange, it transfers the selected connection to the submatrix, but in addition, the original or valid connection outside the submatrix that was in the row or column to be transferred is replaced by the original location of the selected connection. Thus, the transfer in the row direction takes place by swapping two columns with each other. Correspondingly, the shift in the direction of the column can be illustrated by the interchange of the respective rows. Transfers are made for each connection starting from the first connection and ending with the last connection in the output series.

The algorithm in Figure 3 starts from step 10. In decision block 11, it is examined whether the connection is already in some submatrix. If so, the operation proceeds to block 18, otherwise in block 13 it is examined in block 13 whether the submatrix in question contains:; : lines. If not, decision block 14 examines whether there is a free column in the submatrix in question. If no free column is found, the operation proceeds to the recursion ... Subroutine, which is explained in connection with Figure 4. If the free column is found in block 14,; . ·. in block 15, the interleaving of the rows moves back to block 25- '11. If a free row is found in block 13, the inter-column' swapping of the operations in block 17 moves to block 18. In decision block 18, it is examined whether the last connection has already been processed. If not, the next connection is selected in block 19 and the operation continues from block 11, otherwise the algorithm ends in step 20.

The execution of the algorithm according to the invention can be illustrated in the case of the imaginary matrix shown in Figure 2. The algorithm starts with the first f · connection of the output series, i.e. the result 3.5 of the result of Figure 2 to the output 1.1. This coupling is not in the submatrix because of the input and output time slots or „ri time slots that are examined. ‘. is shown in block 11 of Figure 3. In block 13, it is determined that a connection from the same input port has already been used in the submatrix in the same row. Therefore, 35 column swaps are not made. However, in block 14 of Figure 3, it is indicated that there is free in the sub-matrix in the direction of the column, in which case a row change is made in block 15. The matrix of Figure 10 97842 2 is updated, resulting in the table of Figure 5. Changed values are shown in bold and italics.

Next, we look at position 1.1 in the starting series. The element in this row now combines input 4.4 and output 1.1, so the element is not in the submatrix. There is a free space in the sub-matrix in the direction of the row, and thus a column change is made. The first connection of the output series is then complete. The main loop of Figure 3 then moves to the next row of the imaginary matrix with output port 2, time slot 1. This connection is already in the submatrix, so no switching is required. The next output, 3.1, is then examined. The connection from output 1.3 to output 3.1 is not in the submatrix, and there is no free in the direction of 10 rows. In this case, the rows are changed because there is a free element in column 3.3 (output port 3, time slot 3) of the submatrix. At the same time, the target line, i.e. coupling 2.1 to »3.3, is moved to the position of the processed line, whereby the coupling element enters directly into the submatrix. In this case, the submatrix at the top left is complete. The situation of the total imaginary matrix is now shown in Figure 6 within the reinforced-15 box.

The method described above continues to fill the other sub-matrices.

When row and column swaps are made as described above, the corresponding status sets are also updated. The column feed is marked in the first status set (input side), and the row break is marked in the second status set (output side).

20 Next, the recursion situation will be considered with reference to Fig. 4. This is used when there is no free space for transfer in either possible submatrix. This means that both submatrices are occupied, either in the case of •. ·: At the current input port or output port. When this happens, \ we will break the line break and column break until a free space is found. The block diagram in Figure 3 should be completed with these forced row and column breaks. Recursion starts from the decision block: '. 14, when no free column is found, in which case the recursion calculation is started according to these 16.

: ‘'The recursion calculation resembles the recursion operation of the test algorithm described above. In the game algorithm according to the invention, the previously made selection is released from the submatrix-30, and at the same time it is removed from the submatrix. This corresponds to the operation of a test algorithm in which a new solution is chosen instead of the old one. Now, however, the game algorithm fills the submatrix so that a new working solution is obtained. The problem, however, is that it must be ensured that the game algorithm is not left in a loop when retrieving a free submatrix for a connection that has been switched out of the submatrix. The test algorithm 11 97842 rhythm is successful because it works on the principle that in each shift, i.e., deletion-selection, only a new solution is selected for processing. The game algorithm according to the invention uses the same selection process when new connections are taken into processing.

5 The following is a review of the recursion situation. The basic structure of a large coupling matrix is based on the fact that there is only one coupling in the same row or in the same column at a time. When recursion is started, there are two candidate sub-matrices available for the signal under consideration, with the corresponding port reserved: one sub-matrix on the input side, i.e. a sub-matrix in the row direction, and another sub-matrix on the output side, i.e. a sub-matrix in the column direction. Throughout the recursion process, only these two sub-matrices are processed, as recursion is terminated immediately when a free space is found in either sub-matrix. This means that in the recursion, the connections are shifted between these submatrices until a free space is found. If the recursion then moves to the loop, the exchanges must take place in the middle of these two submatrices. This also means that the recursion process does not involve connections outside the two sub-matrices. Figure 4 considers an issue with two adjacent submatrices.

In Fig. 4, the recursion process starts from changing the position of the valid connection, so that this connection is moved out of the state s 1, which is in the upper-20th right in the figure. The transfer first concerns the connection b1, which is transferred from the state sl to the submatrix; ; s3. Thereafter, the exchange means that the connections b1 and b2 included in the exchange are the only connections that use a common port, which makes the exchange necessary. In the figure, this port in time slot t + 1 (time slot t + 1), with respect to which the signals b1 and I,, b2 overlap, is the output port OI. When recursion is continued, it is required * 2?> * That the next forced switch applies to signals having a common port, in which case the common port must not be the same port that was common to the signal of the previous switch b 1. The switch in this case is in the submatrix s2 on the left above, and '-] * applies to signals b2 and b3. In this case, the common port of these signals is the input port II, in which case the signal b3 must be moved away. Then b ^ falls outside the submatrix s2 • 30 · into the state sl and is transferred from there again to the submatrix s3. S-λ to be traversed • ·? ·· *. signal b4 must again have a common port (02) with signal b3. Thus, b4 is moved to the state sl. Then, connection b5 is processed, which is moved out of the submatrix s2 because it has the same input port as connection b4.

The “Domino” effect of recursion can be continued for all available port 35 paths. In order for a recursion to circulate in a loop, it must have a connection that is changed to a set of forced changes. The only available option for 12 97842 switching is at the beginning of the chain, because the only port port in the chain that has not yet been used is for connection bl, whose input port Ix has not yet been used. In Fig. 4, Ix = 14. Since Ix is also the port that was found to be busy in the second submatrix for connection b1, this means that the connection bz and the overlapping gate Ix are already in the second submatrix. The only way to get the switches bz exchanged with the connection bl is to find another connection to port Ix. Since bl and bz use port Ix, a third circuit that would use the same port Ix in the area covered by the submatrices is out of the question. Therefore, there can also be no recursion that would cause a loop at the beginning of the exchange chain. The chain reaction thus requires a maximum of 2 * (k-1) exchanges to solve the coupling and place it already in the submatrix of the honk by recursion. Then k corresponds to the size of the cross-switch, i.e. kmax = 16. The chain switching process resembles a test algorithm for rearranging solutions within a submatrix, but the recursion presented here also solves two submatrices simultaneously.

15 The strength of the game algorithm is that no actual computational processes are required. The connections are only exchanged for free positions in different submatrices. Another advantage of the game algorithm is that only connections outside the sub-matrices are swapped, which leaves the already established successful connections unprocessed. The switching process also has a very useful feature, when the connection transferred to the submatrix can force the second connection immediately to the second submatrix by only one execution step. The recursion process associated with the algorithm also provides very advantageous properties compared to the test algorithm. In the recursion process, the connections shifted out of the submatrix are used to complement the second submatrix, even so that two matrices can be processed simultaneously. The main advantage of a game algorithm is that it starts from a formation of separate, individual connections, which are then rearranged, so that the game algorithm does not have to start all over «from scratch, as in test algorithms and some other algorithms.

• ». · \ Since the game algorithm does not require any special calculation before selecting the transfer of a connection, the connection can be transferred in two possible directions. The decision to switch the connections ·· ·: r is made on a random basis, which depends only on the distribution of the connections and * * in the output matrix. In particular, forced exchanges can be used to address the difficult situations of women in the story.

Claims (10)

97842 A method for cross-linking digital signals in a digital cross-connect between the respective input and output points so that two or more signals do not enter the same output port in a given time interval, characterized in that it comprises the steps of a) generating a k * k switching request table in which the row has k slots of input port n and the column k has n slots of output port, and the content of the element corresponds to input port and input time slot, and the connection request table records connections in two auxiliary tables describing connections from inputs to outputs, the first corresponding to 10 inputs and the second corresponding to outputs are the two axes of the imaginary matrix formed by the connection request table, b) the n * n submatrices on the diagonal of the matrix are formed, c) the first connection to be examined is selected, d) it is examined (11) whether the connection is in the submatrix, and if this is valid, go to step-g), otherwise e) examine (13) whether there is free space in the row of the connection to be processed in the relevant sub-matrix, and if so, exchange the column of the connection to be processed and the free column in the sub-matrix (17), otherwise, f ) examines (14) whether there is free space in the column of the connection in question in the relevant submatrix 20, and if so, exchanges the row of the connection under examination and the free row in the submatrix (15) and proceeds to step d), otherwise a forced row in different recursion step (16) - and / or column switching and returning to step d), g) examining (18) whether all the connections in the matrix have been processed and, if so, the process; 25 si is terminated (20), otherwise the following to be processed is selected (19) from the coupling matrix:.:: Coupling and return to step d). • · · • · · • » Method according to Claim 1, characterized in that the submatrix. corresponds to the time interval. • «· • · · • · ·« Method according to Claim 1 or 2, characterized in that in the step:! 30 d) only a connection that has not been previously processed is selected for processing. ♦ · ♦ ♦ ♦ · · • · ·:; Method according to one of the preceding claims, characterized in that k = 16 and n = 63. Method according to one of the preceding claims, characterized in that in said recursion step the connections are exchanged between the two sub-matrices in question until a free space is formed in at least one of the sub-matrices. 97842 Method according to Claim 5, characterized in that the switching of the connections is carried out on a random basis. Method according to one of the preceding claims, characterized in that the steps of the method are carried out at my time intervals. Method according to one of the preceding claims, characterized in that the signals to be switched are VC-12 virtual containers of STM-1 signals. Method according to one of the preceding claims, characterized in that, in the case of bidirectional connections, a solution is sought only for one direction and the connection of the other direction is implemented symmetrically. 10 Patentkrav