A non-blocking full-switch matrix cascading method based on dual variable routing

By adopting cross-connection of double variable routing in the full matrix cascade method, the problem of inconsistent number of ports of cascade matrix units is solved, which significantly reduces the blocking probability and search time of routing planning.

CN119052159BActive Publication Date: 2025-02-14CHENGDU CHUANMEI NEW TECH CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202411516340.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-10-29
Publication Date
2025-02-14
Estimated Expiration
2044-10-29

AI Technical Summary

Technical Problem

In the prior art, the non-blocking full matrix cascade method based on variable routing is difficult to effectively solve the problem that the number of ports of the previous level matrix unit is inconsistent with the number of next level matrix units, resulting in high blocking probability and routing search time during routing planning. long.

Method used

Using the non-blocking full-fabric cascade method based on dual variable routing, the routing relationship between input array IN to output array OUT is calculated by establishing a cross connection between input switch SR, bridge switch QJ and output switch SC. Cascading the matrix and planning the complete routes of all channels to select conflict-free routes.

Benefits of technology

It greatly reduces the blocking probability during routing planning and shortens the routing search time, effectively solving the problem that the number of ports of the previous level matrix unit is inconsistent with the number of next level matrix units.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119052159B_ABST
    Figure CN119052159B_ABST
Patent Text Reader

Abstract

The present invention belongs to the technical field of full-switch matrix cascading methods, and specifically relates to a non-blocking full-switch matrix cascading method based on dual variable routing, which adopts M×M matrix units to cascade M*M / 2×M*M / 2 matrices, and connects three groups of matrix units according to a cascade relationship, including input switching SR, bridge switching QJ, and output switching SC; establishes a cross connection between two groups of matrix units, calculates the routing relationship cascade matrix from the input array IN to the output array OUT; plans the complete routing (SR‑ij, SC‑jk) of all channels, and selects a group of mutually non-conflicting routes; calculates the switching configuration array of the matrix units of the input switching SR, bridge switching QJ, and output switching SC, and performs synchronous switching control. The problem of the inconsistency between the number of ports of the upper-level matrix units and the number of the lower-level matrix units is solved, and a double multi-route cross connection is adopted to reduce the blocking probability during routing planning and shorten the routing search time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of full-switch matrix cascading methods, and in particular relates to a non-blocking full-switch matrix cascading method based on dual variable routing. Background Art

[0002] In the prior art, there is a non-blocking full switching matrix cascading method based on variable routing, such as Figure 1 As shown, M×M units are cascaded into M 2 ×M 2 Matrix, where N=8, 12, 16, can only realize 64×64, 144×144, 256×256 scale matrices. If you want to realize M*M / 2×M*M / 2 matrices, such as 32×32, 72×72, 128×128, etc. The cascade matrix includes 3 groups of M / 2 M×M matrix units, which are divided into 3 groups: input switching SR, bridge switching QJ, and output switching SC. The input / output ports of the matrix unit are inconsistent with the number of matrix units in each group.

[0003] Therefore, how to solve the inconsistency between the number of ports of the upper-level matrix unit and the number of the lower-level matrix unit, and how to construct routes that can significantly reduce the blocking probability during route planning and shorten the route search time are technical problems that need to be solved urgently. Summary of the invention

[0004] The purpose of the present invention is to provide a non-blocking full-switch matrix cascading method based on dual variable routing, which is used to solve the inconsistency problem between the number of ports of the upper-level matrix unit and the number of the lower-level matrix unit. A dual multi-route cross connection is used between the input switch SR, the bridge switch QJ, and the bridge switch QJ and the output switch SC. For example, the cross connection between two groups of matrix units is two, such as between SR1 and QJ2, there are only two routing connections, R12A and R12B. And the number of variable routes is doubled, which greatly reduces the blocking probability during route planning and shortens the route search time.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0006] A non-blocking full-switch matrix cascading method based on dual variable routing includes the following steps:

[0007] S1: Use M×M matrix units to cascade M*M / 2×M*M / 2 matrices, and connect at least three groups of matrix units in a cascade relationship, including input switching SR, bridge switching QJ, and output switching SC;

[0008] S2: Establish a cross connection between two groups of matrix units, that is, double connection, and calculate the routing relationship cascade matrix from the input array IN to the output array OUT;

[0009] S3: Plan the complete routes of all channels (SR-ij, SC-jk) and select a set of mutually non-conflicting routes;

[0010] S4: Calculate the switch configuration array of the matrix units of input switch SR, bridge switch QJ, and output switch SC, and synchronize the switching control.

[0011] Preferably, the specific process of calculating the routing relationship cascade matrix from the input array IN to the output array OUT in step S2 is as follows:

[0012] S21: Let m = M / 2, and split the input array IN into:

[0013]

[0014] Then bridge input array IN QJ for:

[0015]

[0016] IN QJ Split into:

[0017]

[0018] Then the bridge output array OUT QJ for:

[0019] ;

[0020] OUT QJ Split into:

[0021]

[0022] The matrix output is:

[0023] ;

[0024] Among them, SR M×M , QJ M×M , SC M×M They are the internal connection relationship arrays of input switching, bridge switching and output switching matrices, SRQJ and SCQJ are the cross connection relationship arrays between input units and bridge units, and bridge units and output units, respectively;

[0025] S22: Through SR M×M , QJ M×M , SC M×M The assignment of arrays, that is, the control of each M×M matrix unit, realizes arbitrary exchange of M*m×M*m matrices.

[0026] Preferably, step S3 includes the following specific process:

[0027] S31: Define the input route R_SR according to the connection between the input switch SR and the bridge switch QJ:

[0028] ;

[0029] Where, i is the input switch SR matrix unit number, j is the bridge switch QJ matrix unit number, A represents the first route, and B represents the second route;

[0030] S32: Define the output route R_SC according to the connection between the bridge switch QJ and the output switch SC:

[0031] ;

[0032] Where: j is the bridge switch QJ matrix unit number, k is the output switch SC matrix unit number, A represents the first route, and B represents the second route;

[0033] S33: i and k corresponding to the matrix input port number IN to the matrix output port number OUT:

[0034]

[0035]

[0036] S34: Calculate the complete route from any input to any output in the cascade matrix;

[0037] S35: Select one of the complete routes from any input to any output in the calculation cascade matrix, and the selected route does not overlap with the routes selected by other channels.

[0038] Preferably, the routing planning result in step S3 includes: input channel number: IN, output channel number: OUT, input route: R_SR, output route: R_SC; input switch SR matrix unit number i; bridge switch QJ matrix unit number j; output switch SC matrix unit number k;

[0039] Among them: the value range of the input channel number IN and the output channel number OUT is 1 to M / 2; the value range of i, j, k is: 1 to M / 2.

[0040] Preferably, in step S35, a route that overlaps with a route selected by another channel when the input sequence number selected by the channel is the same is defined as a common route.

[0041] Preferably, the calculation of the switch configuration array of the matrix unit of the input switch SR, the bridge switch QJ, and the output switch SC in step S4 includes the following specific steps:

[0042] S41: Calculate the channel selection relationship of the input switch SR matrix unit number i according to the input channel number IN and the bridge switch QJ matrix unit number j;

[0043] S42: Calculate the channel selection relationship of the bridge switch QJ matrix unit number j according to the input switch SR matrix unit number i, the output switch SC matrix unit number k and A or B in the route;

[0044] S43: Calculate the channel selection relationship of the output switch SR matrix unit number k according to the bridge switch QJ matrix unit number j and the matrix output port number OUT;

[0045] S44: respectively calculating the control relationship of each input switching SR matrix unit, the control relationship of the bridge switching QJ matrix unit, and the control relationship of the output switching SC matrix unit, and performing channel control on all the switching units of the cascade matrix according to the calculated control relationship to realize channel selection of the cascade matrix.

[0046] The beneficial effects of the present invention include:

[0047] The present invention provides a non-blocking full-switch matrix cascading method based on dual variable routing, which adopts M×M matrix units to cascade M*M / 2×M*M / 2 matrices, and connects at least three groups of matrix units in a cascade relationship, including input switching SR, bridging switching QJ, and output switching SC; establishes a cross-connection between two groups of matrix units to calculate the routing relationship cascade matrix from the input array IN to the output array OUT; plans the complete routing (SR-ij, SC-jk) of all channels, and selects a group of mutually non-conflicting routes; calculates the switching configuration array of the matrix units of the input switching SR, bridging switching QJ, and output switching SC, and performs synchronous switching control. The problem of inconsistency between the number of ports of the upper-level matrix units and the number of the lower-level matrix units is solved, a double multi-routing cross-connection is adopted, and the number of variable routes is doubled, which greatly reduces the blocking probability during routing planning and shortens the routing search time. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The M×M units in the prior art of the present invention are cascaded into M 2 ×M 2 Schematic diagram of the matrix.

[0049] Figure 2 It is a schematic diagram of the M×M units of the present invention being cascaded into an M*M / 2×M*M / 2 matrix.

[0050] Figure 3 It is a schematic diagram of cascading 4×4 units into an 8×8 matrix according to the present invention.

[0051] Figure 4 An example of channel selection routing for cascading 4×4 units into an 8×8 matrix of the present invention. DETAILED DESCRIPTION

[0052] The following is combined with Figure 1~Figure 4 The present invention is further described in detail:

[0053] Example 1

[0054] See attached Figure 2 As shown, a non-blocking full-switching matrix cascading method based on dual variable routing includes the following steps:

[0055] S1: Use M×M matrix units to cascade M*M / 2×M*M / 2 matrices, and connect at least three groups of matrix units in a cascade relationship, including input switching SR, bridge switching QJ, and output switching SC;

[0056] S2: Establish a cross connection between two groups of matrix units, that is, double connection, and calculate the routing relationship cascade matrix from the input array IN to the output array OUT;

[0057] S3: Plan the complete routes of all channels (SR-ij, SC-jk) and select a set of mutually non-conflicting routes;

[0058] S4: Calculate the switch configuration array of the matrix units of input switch SR, bridge switch QJ, and output switch SC, and synchronize the switching control.

[0059] See also Figure 1 The M×M units in the prior art are cascaded into M 2 ×M 2 Matrix, the input / output ports of the matrix unit are consistent with the number of matrix units in each group. The cross connection between two groups of matrix units is a single connection, such as between SR1 and QJ2, there is only one routing connection, R12. Figure 2 , using M×M matrix units cascaded M*M / 2×M*M / 2 matrix, input switching SR matrix units, bridge switching QJ matrix units, output switching SC matrix units 3 groups, each group has only M / 2, the number of input / output ports of the matrix unit is twice the number of matrix units in each group, the cross connection between the two groups of matrix units is a double connection, see Figure 2 .

[0060] There are two routes between any input switch SR matrix unit and bridge switch QJ matrix unit. Similarly, there are two routes between the bridge switch QJ matrix unit and the output switch SC matrix unit. In the bridge switch, it can also be exchanged. Therefore, there are 4 routes from any input switch SR to the bridge switch QJ and then to the output switch SC, which is M×M units cascaded into M 2 ×M 2 The matrix is ​​four times larger, with more routes and greater flexibility, which is more conducive to the rapid success of route planning.

[0061] For example, when cascading a 144×144 matrix with 12×12 units and a 128×128 matrix with 16×16 units, the number of routes from any matrix input port to any matrix output port is shown in Table 1 below:

[0062] Table 1: Routing numbers for a 144×144 matrix cascaded with 12×12 units and a 128×128 matrix cascaded with 16×16 units

[0063]

[0064] The number of cross-connections and total power dividers / switch units required for cascading a 144×144 matrix with 12×12 units and a 128×128 matrix with 16×16 units is shown in Table 2 below:

[0065] Table 2: Number of cells in a 144×144 matrix cascaded with 12×12 cells and a 128×128 matrix cascaded with 16×16 cells

[0066]

[0067] It can be seen that the cascade method provided by the present invention, although the total number of power dividers / switch units is slightly higher by about 30%, the matrix unit is of the standard size of M×M, and the number of matrix units is reduced by 50%. The number of routes from any matrix input port to any matrix output port is increased by 160%. The inconsistency between the number of matrix unit ports at the previous level and the number of matrix units at the next level is effectively solved, and a double multi-route cross connection is adopted, and the number of variable routes is doubled, which greatly reduces the blocking probability and route search time during route planning.

[0068] Example 2

[0069] Based on Example 1, the specific process of calculating the routing relationship cascade matrix from the input array IN to the output array OUT in step S2 is as follows:

[0070] S21: Let m = M / 2, and split the input array IN into:

[0071]

[0072] Then bridge input array IN QJ for:

[0073]

[0074] IN QJ Split into:

[0075]

[0076] Then the bridge output array OUT QJ for:

[0077]

[0078] OUT QJ Split into:

[0079]

[0080] Then the matrix is:

[0081]

[0082] Among them, SR M×M , QJ M×M , SC M×M They are the internal connection relationship arrays of the input switching, bridge switching and output switching matrices respectively, and SRQJ and SCQJ are the cross-connection relationship arrays between the input unit and the bridge unit, and the bridge unit and the output unit respectively.

[0083] The values ​​in the SRQJ array are defined as:

[0084] ;

[0085] Among them, i is the output port number of the input switching SR matrix unit, which is 1, 2, ..., M*M / 2 respectively; j is the input port number of the bridge switching QJ matrix unit, which is 1, 2, ..., M*M / 2 respectively.

[0086] The values ​​in the SCQJ array are defined as:

[0087] ;

[0088] Among them, j is the output port serial number of the bridge switch QJ matrix unit, which is 1, 2, ..., M*M / 2 respectively; k is the output port serial number of the output switch SC matrix unit, which is 1, 2, ..., M*M / 2 respectively.

[0089] S22: Through SR M×M , QJ M×M , SC M×MThe assignment of arrays, that is, the control of each M×M matrix unit, realizes arbitrary exchange of M*m×M*m matrices.

[0090] For example, when M=4, the principle of cascading 4×4 into 8×8 matrix is ​​as follows Figure 3 As shown:

[0091] QUR 8×8 and QJSC 8×8 The relationship array is equal to the following:

[0092] ;

[0093] The value in the array is 1, indicating a connection, and 0, indicating no connection.

[0094] like:

[0095]

[0096] but:

[0097] ;

[0098] ;

[0099] ;

[0100] From the above calculation, it can be seen that the corresponding relationship between the input and output of the 8×8 matrix is ​​"sequential" one-to-one selection.

[0101] when:

[0102]

[0103]

[0104] but:

[0105] ;

[0106] ;

[0107] .

[0108] The calculation results show that the corresponding relationship between the input and output of the 8×8 matrix is ​​"reverse order" and one-to-one selection.

[0109] Example 3

[0110] On the basis of Embodiment 1 and Embodiment 2, step S3 includes the following specific processes:

[0111] S31: Define the input route R_SR according to the connection between the input switch SR and the bridge switch QJ:

[0112] Where i is the input switch SR matrix unit number, j is the bridge switch QJ matrix unit number, A represents the first route, and B represents the second route;

[0113] S32: Define the output route R_SC according to the connection between the bridge switch QJ and the output switch SC:

[0114]

[0115] Where: j is the bridge switch QJ matrix unit number, k is the output switch SC matrix unit number, A represents the first route, and B represents the second route;

[0116] S33: i and k corresponding to the matrix input port number IN to the matrix output port number OUT:

[0117]

[0118]

[0119] S34: Calculate the complete route from any input to any output in the cascade matrix;

[0120] The complete route from any input to any output in the cascade matrix includes the input route R_SR and the output route R_SC, each of which has 2 bridge switches QJ with sequence number J. There are 4 combinations:

[0121] ;

[0122] There are M / 2 bridge switches QJ, which is a complete route from any input to any output, with a total of 4*M / 2 optional routes.

[0123] When M=4, A=5, B=6, then i=2, k=2, the optional routes are:

[0124] ;

[0125] S35: Select one of the complete routes from any input to any output in the calculation cascade matrix, the selected route does not overlap with the routes selected by other channels, and define the routes that overlap with the routes selected by other channels when the input sequence numbers selected by the channels are the same as the common routes.

[0126] The routing planning result in step S3 includes: input channel number: IN, output channel number: OUT, input route: R_SR, output route: R_SC; input switch SR matrix unit number i; bridge switch QJ matrix unit number j; output switch SC matrix unit number k;

[0127] Among them: the value range of the input channel number IN and the output channel number OUT is 1 to M / 2; the value range of i, j, k is: 1 to M / 2.

[0128] Example 4

[0129] On the basis of Embodiment 1 or Embodiment 2 or Embodiment 3, the calculation of the switch configuration array of the matrix unit of the input switch SR, the bridge switch QJ, and the output switch SC in step S4 includes the following specific steps:

[0130] S41: Calculate the channel selection relationship of the input switch SR matrix unit number i according to the input channel number IN and the bridge switch QJ matrix unit number j;

[0131] S42: Calculate the channel selection relationship of the bridge switch QJ matrix unit number j according to the input switch SR matrix unit number i, the output switch SC matrix unit number k and A or B in the route;

[0132] S43: Calculate the channel selection relationship of the output switch SR matrix unit number k according to the bridge switch QJ matrix unit number j and the matrix output port number OUT;

[0133] S44: respectively calculating the control relationship of each input switching SR matrix unit, the control relationship of the bridge switching QJ matrix unit, and the control relationship of the output switching SC matrix unit, and performing channel control on all the switching units of the cascade matrix according to the calculated control relationship to realize channel selection of the cascade matrix.

[0134] Input exchange SR matrix unit, according to IN, j, the channel selection relationship of input exchange SR unit i (input X, output Y) can be calculated:

[0135] X=MOD((IN-1),M)+1

[0136]

[0137] The bridge switching QJ matrix unit can calculate the channel selection relationship (input X, output Y) of the bridge switching QJ unit j according to i, k and A or B in the route:

[0138]

[0139]

[0140] Output exchange SR matrix unit, according to j and OUT, the channel selection relationship of output exchange SR unit k (input X, output Y) can be calculated:

[0141] ;

[0142] Y= MOD((OUT-1),M)+1;

[0143] See Table 3 below, the routing of the 8×8 matrix of cascaded 4×4 matrix units, a total of 8 routes:

[0144] Table 3: Routing of an 8×8 matrix with 4×4 matrix units cascaded

[0145]

[0146] The control relationship of each input switching SR matrix unit, the control relationship of the bridge switching QJ matrix unit, and the control relationship of the output switching SC matrix unit (input X-output Y) are calculated respectively. The control relationship of the input switching SR matrix unit is as shown in Table 4:

[0147] Table 4: Control relationships for input switching SR matrix units

[0148]

[0149] The control relationship of the bridge switching QJ matrix unit is as follows in Table 5:

[0150] Table 5: Control relationship of bridge switching QJ matrix unit

[0151]

[0152] The control relationship of the output switching SC matrix unit is:

[0153] Table 6: Control relationships of output switching SC matrix units

[0154]

[0155] According to the control relationship of all switching units in the routing calculation, channel control is performed on all switching units of the cascaded 8×8 matrix to realize channel selection of the cascaded matrix. The routing is shown in Figure 4 .

[0156] In summary, the present invention provides a non-blocking full-switch matrix cascading method based on dual variable routing, which uses M×M matrix units to cascade M*M / 2×M*M / 2 matrices, and connects at least three groups of matrix units in a cascade relationship, including input switching SR, bridging switching QJ, and output switching SC; establishes a cross-connection between two groups of matrix units to calculate the routing relationship cascade matrix from the input array IN to the output array OUT; plans the complete routing (SR-ij, SC-jk) of all channels, and selects a group of mutually non-conflicting routes; calculates the switching configuration array of the matrix units of the input switching SR, bridging switching QJ, and output switching SC, and synchronizes the switching control. The problem of the inconsistency between the number of ports of the upper-level matrix unit and the number of the lower-level matrix unit is solved, a double multi-route cross-connection is adopted, and the number of variable routes is doubled, which greatly reduces the blocking probability during routing planning and shortens the routing search time.

Claims

1. A non-blocking full-switch matrix cascading method based on dual variable routing, characterized in that: The following steps are involved: S1: Use M×M matrix units to cascade M*M / 2×M*M / 2 matrices, and connect at least three groups of matrix units in a cascade relationship, including input switching SR, bridge switching QJ, and output switching SC; S2: Establish a cross connection between two groups of matrix units, that is, double connection, and calculate the routing relationship cascade matrix from the input array IN to the output array OUT; S3: Plan the complete routes of all channels (SR-ij, SC-jk) and select a set of mutually non-conflicting routes; S4: Calculate the switch configuration array of the matrix units of input switch SR, bridge switch QJ, and output switch SC, and synchronize the switching control; The specific process of calculating the routing relationship cascade matrix from the input array IN to the output array OUT in step S2 is as follows: S21: Let m = M / 2, and split the input array IN into: Then bridge input array IN QJ for: IN QJ Split into: Then the bridge output array OUT QJ for: OUT QJ Split into: The matrix output is: ; Among them, SR M×M , QJ M×M , SC M×M They are the internal connection relationship arrays of input switching, bridge switching and output switching matrices, SRQJ and SCQJ are the cross connection relationship arrays between input units and bridge units, and bridge units and output units, respectively; S22: Through SR M×M , QJ M×M , SC M×M The assignment of arrays, that is, the control of each M×M matrix unit, realizes arbitrary exchange of M*m×M*m matrices.

2. A non-blocking full switching matrix cascading method based on dual variable routing according to claim 1, characterized in that: Step S3 includes the following specific processes: S31: Define the input route R_SR according to the connection between the input switch SR and the bridge switch QJ: ; Describes the route as a string, where: i is the input exchange SR matrix unit number; j is the unit number of the bridge exchange QJ matrix; A represents the first route, and B represents the second route; S32: Define the output route R_SC according to the connection between the bridge switch QJ and the output switch SC: ; Describes the route as a string, where: j is the unit number of the bridge exchange QJ matrix; k is the output exchange SC matrix unit number; A represents the first route, and B represents the second route; S33: i and k corresponding to the matrix input port number IN to the matrix output port number OUT: ; ; S34: Calculate the complete route from any input to any output in the cascade matrix; S35: Select one of the complete routes from any input to any output in the calculation cascade matrix, and the selected route does not overlap with the routes selected by other channels.

3. A non-blocking full switching matrix cascading method based on dual variable routing according to claim 2, characterized in that: The routing planning result in step S3 includes: input channel number: IN, output channel number: OUT, input route: R_SR, output route: R_SC; input switch SR matrix unit number i; bridge switch QJ matrix unit number j; output switch SC matrix unit number k; Among them: the value range of the input channel number IN and the output channel number OUT is 1 to M / 2; the value range of i, j, k is: 1 to M / 2.

4. A non-blocking full switching matrix cascading method based on dual variable routing according to claim 2, characterized in that: In step S35, a route that overlaps with a route selected by another channel when the input sequence number selected by the channel is the same is defined as a common route.

5. The method for cascading a non-blocking full switching matrix based on dual variable routing according to claim 1, characterized in that: The calculation of the switch configuration array of the matrix unit of the input switch SR, the bridge switch QJ, and the output switch SC in step S4 includes the following specific steps: S41: Calculate the channel selection relationship of the input switch SR matrix unit number i according to the input channel number IN and the bridge switch QJ matrix unit number j; S42: Calculate the channel selection relationship of the bridge switch QJ matrix unit number j according to the input switch SR matrix unit number i, the output switch SC matrix unit number k and A or B in the route; S43: Calculate the channel selection relationship of the output switch SR matrix unit number k according to the bridge switch QJ matrix unit number j and the matrix output port number OUT; S44: respectively calculating the control relationship of each input switching SR matrix unit, the control relationship of the bridge switching QJ matrix unit, and the control relationship of the output switching SC matrix unit, and performing channel control on all the switching units of the cascade matrix according to the calculated control relationship to realize channel selection of the cascade matrix.

Citation Information

Patent Citations

  • Non-blocking full switching matrix cascading method based on variable routing

    CN118972347A