Non-blocking full switching matrix cascading method based on variable routing
Through the non-blocking full-fabric cascade method based on variable routing, the problem of a sharp increase in the number of power division and switching units in large-scale matrices is solved, the matrix scale is simplified and cost is reduced, and the non-blocking full-switching function is realized .
Patent Information
- Application Number
- CN202411442919.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-16
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-10-16
AI Technical Summary
In the existing full matrix cascade technology, with the increase of matrix scale, the number of power division, number of switching units and number of cross-connected cables have increased dramatically, resulting in high structural complexity and cost, making it difficult to achieve the effective composition of large-scale matrices.
Using the non-blocking full matrix cascade method based on variable routing, the M×M scale matrix unit is constructed, and routing planning and switching configuration calculation are carried out to realize synchronous switching between input, bridge and output matrix units, reducing the power score. /Number of switch units.
The composition scale of large-scale matrix is greatly reduced, from exponential to multiples, achieving a non-blocking full-switching function, simplifying structural design and reducing costs.
Smart Images

Figure CN118972347A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of cascade design of full-switching large-scale radio frequency matrix switches, and in particular relates to a non-blocking full-switching matrix cascade method based on variable routing. Background Art
[0002] Matrix switches are usually composed of power dividers, switches and other units. A full switching matrix with "pump-out (branching)" function, an M×N matrix, is composed of M 1:N power dividers + N M:1 switch modules.
[0003] The power division module is generally composed of a cascade of 1:2 basic units, and the switch module is generally composed of a cascade of 2:1 basic units (such as an 8:1 switch requires 4+2+1=7 2:1 switch units). The number of cross-connection lines or cables between the power division unit and the switch is M×N. As the scale of the matrix increases, the number of power division basic units, switch basic units, and cross-connection lines or cables required will be very large, which is basically impossible to achieve due to the structural size and connection complexity, and the cost is also quite expensive.
[0004] The composition statistics of matrices of different sizes are shown in the following table: As can be seen from the table above, the module width required for a 256×256 matrix reaches 3 meters, the number of cross-connected cables reaches more than 60,000, and the number of power dividers and switch basic units reaches more than 130,000. The project is basically unfeasible to implement.
[0005] The matrix scale expansion technology of power splitter + matrix + switch is adopted, and multiple small and medium-sized matrices are cascaded into a large-scale matrix. Using the above matrix scale expansion technology, a 256×256 matrix can be realized in engineering. The number of cross-connection cables and power splitters and switch basic units required for matrix systems of different sizes is shown in the following table: As can be seen from the table above, for a 256×256 matrix, the number of power splitters and switch basic units required remains unchanged, and the number of cross-connection cables required increases slightly. If the cascade matrix is designed to be 32×32 in size and the volume is a standard rack-mounted chassis, the number of cross-connection cables and power splitters and switch basic units required for matrix systems of different sizes is shown in the following table: As can be seen from the table above, for a 256×256 matrix, the number of power dividers and switch basic units required remains unchanged, and the number of cross-connection cables required is slightly less than that of a 16×16 cascade. The number of cables connecting devices across 64×64, 128×128, and 256×256 matrix systems still reaches 256, 1024, and 4096, respectively, which is still very difficult and requires a considerable amount of work in engineering construction.
[0006] Therefore, how to provide a new full-switching matrix cascading method to achieve non-blocking full switching of the cascaded matrix, greatly reduce the number of power dividers / switch units of large-scale, non-blocking, full-switching matrix switches, and reduce the composition scale of large-scale matrices from exponential level to multiple level, is a technical problem that urgently needs to be solved. Summary of the invention
[0007] The purpose of the present invention is to provide a new full-switching matrix cascading method to achieve non-blocking full switching of the cascaded matrix, greatly reduce the number of power dividers / switch units of large-scale, non-blocking, full-switching matrix switches, and reduce the composition scale of the large-scale matrix from exponential level to multiple level.
[0008] In order to solve the above technical problems, the technical solution adopted by the present invention is as follows: A non-blocking full-switch matrix cascading method based on variable routing comprises the following steps: S1: Construct a matrix channel with variable routing: Take the M×M matrix scale as the basic matrix unit and construct M 2 ×M 2 The non-blocking full switching matrix includes M input switching matrix units, M bridge switching matrix units and M output switching matrix units. The input switching matrix unit SR, the bridge switching matrix unit QJ and the bridge switching matrix unit QJ and the output switching SC are cross-connected one by one. S2: Perform complete routing planning for the cascade matrix channels: plan the complete routing combination of all channels, use the routing optimization algorithm, and select a set of mutually non-conflicting routes; S3: Calculate the switching configuration array of the matrix unit: calculate the switching configuration relationship of the input switching matrix unit SR, the bridge switching matrix unit QJ and the output switching matrix unit SC through the result of the routing planning; S4: Execute route planning and realize matrix switching: According to the switching configuration relationship of the matrix unit, calculate the switch control I / O in the board, control all switches to switch synchronously, and simultaneously open all channel planning routes to realize matrix switching.
[0009] Preferably, in step S1, the basic matrix unit based on the scale of M×M is constructed as M 2 ×M 2 The non-blocking full switching matrix comprises M input switching matrix units, M bridge switching matrix units and M output switching matrix units, and at least three groups of sub-matrix units are provided. 2 The M output ports and the M bridge switching matrix units QJ 2 The input ports are cross-connected one by one, and the M bridge switching matrix units QJ have M2 M output ports and M output exchange SC 2 One-to-one cross connection between input ports.
[0010] Preferably, in step S3, the array OUT is calculated and outputted by the routing relationship cascade matrix. M 2 and IN M 2 , IN M 2 Split into , then the bridge input is: ; Will Split into , then the bridge output is: ; OUT QJ Split into , then the matrix output is: ; in, , , They are the internal connection relationship arrays of the input switching matrix unit SR, the bridge switching matrix unit QJ and the output switching matrix unit SC, SRQJ M 2 ×M 2 and QJSC M 2 ×M 2 The cross-connection relationship arrays are respectively between the input switching matrix unit and the bridge switching matrix unit, and between the bridge switching matrix unit and the output switching matrix unit; The SRQJ array value is defined as: ; Wherein, i is the output port of the input switching matrix unit SR, which is 1, 2, ..., M*M respectively; j is the input port of the bridge switching matrix unit QJ, which is 1, 2, ..., M*M respectively; The QJSC array value is defined as: ; Among them, j is the output port of the bridge switching matrix unit QJ, which is 1, 2, ..., M*M respectively; k is the input port of the output switch SC, which is 1, 2, ..., M*M respectively.
[0011] Preferably, the specific process of step S2 is as follows: S21: Define the input route R_SR according to the connection between the input switching matrix unit SR and the bridge switching matrix unit QJ: ; Wherein, m is the serial number of the input switching matrix unit SR, n is the serial number of the bridge switching matrix unit QJ, CHAR() is a function for converting numbers to characters, and SR_ is the input string; S22: Define the output route R_SC according to the connection between the bridge switching matrix unit QJ and the output switch SC: ; Where n is the bridge switch matrix unit QJ serial number, k is the output switch SC serial number, CHAR() is the function for converting numbers to characters; The matrix input port number IN and output port number OUT correspond to m and k: ; ; Among them, TRUNC() is the rounding function; S23: According to the routing plan, each channel selects one of the M routes. The selected route does not overlap with the routes selected by other channels or can overlap with them when the input port numbers are the same. The overlapping routes when the input port numbers are the same are shared routes.
[0012] Preferably, in step S2, the complete route from any input to any output in the cascade matrix, including the input route R_SR and the output route R_SC, passes through the bridge switching matrix units QJ numbered 1 to M, and there are M optional routes: ; When M=4, IN=5, OUT=6, then i=2, k=2, the optional routes are: .
[0013] Preferably, step S3 includes the following specific process: Input channel number: IN, output channel number: OUT, input routing: R_SR, output routing: R_SC, where: the value range of IN and OUT is: 1~M, the value range of i, j, k is: 1~M, i, j, k correspond to the serial numbers of input, bridge, and output matrix units respectively; S31: Input switch matrix unit SR unit According to IN and j, the channel selection relationship of the input switching matrix unit SR unit i (input X, output Y) can be calculated: X=MOD((IN-1),M)+1;Y=j S32: Bridge Switching Matrix Unit QJ Unit According to i and k, the channel selection relationship of unit j of the bridge switching matrix unit QJ can be calculated (input X, output Y): X=i,Y=k S33: Output switching matrix unit SC unit The channel selection relationship (input X, output Y) of the output switching matrix unit SCk can be calculated based on j, k and OUT: X=j, Y= MOD((OUT-1),M)+1.
[0014] The beneficial effects of the present invention include: The non-blocking full-switch matrix cascading method based on variable routing provided by the present invention takes the matrix scale of M×M as the basic matrix unit to construct the M matrix of three-stage switching. 2 ×M 2 Switching matrix. 2 A set of mutually non-conflicting routes is selected from the complete routing planning combination of each channel. According to the routing planning results and the switching configuration array of the matrix cascade, the switching configuration relationship of the input switching matrix unit SR, the bridging switching matrix unit QJ and the output switching matrix unit SC is calculated, and the synchronous switching control of the input switching matrix unit SR, the bridging switching matrix unit QJ and the output switching matrix unit SC is realized.
[0015] By inserting a bridge matrix unit between the input and output matrix units, a variable routing channel for three-level switching is constructed, and non-blocking full switching is achieved through conflict-free combination of variable routing. At the same time, compared with the fixed routing matrix, the number of routes is reduced from M to M in the first-level switching. 2 ×M 2 The M is reduced to three-level switching 2 This greatly reduces the number of power dividers / switch units in the matrix switch, and reduces the composition scale of large-scale matrices from exponential level to multiple level. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 The present invention is a flow chart of a non-blocking full-switching matrix cascading method based on variable routing.
[0017] Figure 2 To cascade M×M basic units into M 2 ×M 2 Matrix composition diagram Figure 3 The composition diagram of 4×4 basic units cascaded into a 16×16 matrix Figure 4 A schematic diagram of the routing of a 16×16 matrix cascaded from 4×4 basic units DETAILED DESCRIPTION
[0018] The following is combined with Figure 1-Figure 4 The present invention is further described in detail: See also Figure 1 , a non-blocking full-switch matrix cascading method based on variable routing, comprising the following steps: S1: Construct a matrix channel with variable routing: Take the M×M matrix scale as the basic matrix unit and construct M 2 ×M 2 The non-blocking full switching matrix includes M input switching matrix units, M bridge switching matrix units and M output switching matrix units. The input switching matrix unit SR, the bridge switching matrix unit QJ and the bridge switching matrix unit QJ and the output switching SC are cross-connected one by one. S2: Perform complete routing planning for the cascade matrix channels: plan the complete routing combination of all channels, use the routing optimization algorithm, and select a set of mutually non-conflicting routes; S3: Calculate the switching configuration array of the matrix unit: calculate the switching configuration relationship of the input switching matrix unit SR, the bridge switching matrix unit QJ and the output switching matrix unit SC through the result of the routing planning; S4: Execute route planning and realize matrix switching: According to the switching configuration relationship of the matrix unit, calculate the switch control I / O in the board, control all switches to switch synchronously, and simultaneously open all channel planning routes to realize matrix switching.
[0019] The non-blocking full switching matrix cascading method based on variable routing of the present invention takes the matrix scale of M×M as the basic matrix unit to construct the M of three-stage switching 2 ×M 2 Switching matrix. 2 A set of mutually non-conflicting routes is selected from the complete routing planning combination of each channel. According to the routing planning results and the switching configuration array of the matrix cascade, the switching configuration relationship of the input switching matrix unit SR, the bridging switching matrix unit QJ and the output switching matrix unit SC is calculated, and the synchronous switching control of the input switching matrix unit SR, the bridging switching matrix unit QJ and the output switching matrix unit SC is realized.
[0020] By inserting a bridge matrix unit between the input and output matrix units, a variable routing channel for three-level switching is constructed, and non-blocking full switching is achieved through conflict-free combination of variable routing. At the same time, compared with the fixed routing matrix, the number of routes is reduced from M to M in the first-level switching. 2 ×M 2 The M is reduced to three-level switching 2 This greatly reduces the number of power dividers / switch units in the matrix switch, and reduces the composition scale of large-scale matrices from exponential level to multiple level.
[0021] In step S1, the basic matrix unit based on the scale of M×M, the specific cascade structure is shown in Figure 2 , construct M 2 ×M 2 The non-blocking full switching matrix comprises M input switching matrix units, M bridge switching matrix units and M output switching matrix units, and at least three groups of sub-matrix units are provided. 2 The M output ports and the M bridge switching matrix units QJ 2 The input ports are cross-connected one by one, and the M bridge switching matrix units QJ have M 2 M output ports and M output exchange SC 2 One-to-one cross connection between input ports.
[0022] In step S3, the routing relationship cascade matrix is used to calculate and output the array OUT M 2 and IN M 2 , IN M 2 Split into , then the bridge input is: ; Will Split into , then the bridge output is: ; OUT QJ Split into , then the matrix output is: ; in, , , They are the internal connection relationship arrays of the input switching matrix unit SR, the bridge switching matrix unit QJ and the output switching matrix unit SC, SRQJ M 2 ×M 2 and QJSC M 2 ×M 2 The cross-connection relationship arrays are respectively between the input switching matrix unit and the bridge switching matrix unit, and between the bridge switching matrix unit and the output switching matrix unit; The SRQJ array value is defined as: ; Wherein, i is the output port of the input switching matrix unit SR, which is 1, 2, ..., M*M respectively; j is the input port of the bridge switching matrix unit QJ, which is 1, 2, ..., M*M respectively; The QJSC array value is defined as: ; Among them, j is the output port of the bridge switching matrix unit QJ, which is 1, 2, ..., M*M respectively; k is the input port of the output switch SC, which is 1, 2, ..., M*M respectively.
[0023] This embodiment takes M=4 as an example. For the principle of 4×4 cascade matrix, see Figure 3 , SRQJ M 2 ×M 2 and QJSC M 2 ×M 2 The relationship array is equal to the following: ; The value in the array is 1, indicating a connection, and 0, indicating no connection.
[0024] when ; but: ; .
[0025] The calculation results show that the corresponding relationship between the input and output of the 16×16 matrix is "sequential" one-to-one selection.
[0026] when , , but, ; ; .
[0027] The calculation results show that the corresponding relationship between the input and output of the 16×16 matrix is "reverse order" and one-to-one selection.
[0028] The specific process of step S2 is as follows: S21: Define the input route R_SR according to the connection between the input switching matrix unit SR and the bridge switching matrix unit QJ: ; Wherein, m is the serial number of the input switching matrix unit SR, n is the serial number of the bridge switching matrix unit QJ, CHAR() is a function for converting numbers to characters, and SR_ is the input string; S22: Define the output route R_SC according to the connection between the bridge switching matrix unit QJ and the output switch SC: ; Where n is the bridge switch matrix unit QJ serial number, k is the output switch SC serial number, CHAR() is the function for converting numbers to characters; The matrix input port number IN and output port number OUT correspond to m and k: ; Among them, TRUNC() is the rounding function; S23: According to the routing plan, each channel selects one of the M routes. The selected route does not overlap with the routes selected by other channels or can overlap with them when the input port numbers are the same. The overlapping routes when the input port numbers are the same are shared routes.
[0029] Preferably, in step S2, the complete route from any input to any output in the cascade matrix, including the input route R_SR and the output route R_SC, passes through the bridge switching matrix units QJ numbered 1 to M, and there are M optional routes: ; When M=4, IN=5, OUT=6, then i=2, k=2, the optional routes are: .
[0030] According to the routing plan, each channel selects one of the M routes. This route does not overlap with the routes selected by other channels. It can overlap with the routes selected by other channels when the input sequence numbers selected by the channels are the same. This is called a shared route.
[0031] In the above scheme, step S3 includes the following specific processes: Input channel number: IN, output channel number: OUT, input routing: R_SR, output routing: R_SC, where: the value range of IN and OUT is: 1~M, the value range of i, j, k is: 1~M, i, j, k correspond to the serial numbers of input, bridge, and output matrix units respectively.
[0032] S31: Input switch matrix unit SR unit According to IN and j, the channel selection relationship of the input switching matrix unit SR unit i (input X, output Y) can be calculated: X=MOD((IN-1),M)+1;Y=j S32: Bridge Switching Matrix Unit QJ Unit According to i and k, the channel selection relationship of unit j of the bridge switching matrix unit QJ can be calculated (input X, output Y): X=i,Y=k S33: Output switching matrix unit SC The channel selection relationship (input X, output Y) of the output switching matrix unit SCk can be calculated based on j, k and OUT: X=j, Y= MOD((OUT-1),M)+1.
[0033] In this embodiment, a routing example of a 16×16 matrix of cascaded 4×4 matrix units is used, with a total of 16 routes, of which output 13 is idle: Calculate the control of each input, bridge, and output matrix unit (input X, output Y) respectively. The control relationship of the input switching matrix unit SR is shown in the following table: The control relationship of the bridge switching matrix unit QJ is as follows: The control relationship of input switching SC is as follows: According to the control relationship of all switching matrix units in the routing calculation, channel control is performed on all switching matrix units of the cascaded 16×16 matrix to realize channel selection of the cascaded matrix. The routing example is as follows Figure 4 .
[0034] In summary, the non-blocking full switching matrix cascading method based on variable routing provided by the present invention uses an M×M matrix scale as a basic matrix unit to construct a three-stage switching M 2 ×M 2 Switching matrix. 2 A set of mutually non-conflicting routes is selected from the complete routing planning combination of each channel. According to the routing planning results and the switching configuration array of the matrix cascade, the switching configuration relationship of the input switching matrix unit SR, the bridging switching matrix unit QJ and the output switching matrix unit SC is calculated, and the synchronous switching control of the input switching matrix unit SR, the bridging switching matrix unit QJ and the output switching matrix unit SC is realized.
[0035] By inserting a bridge matrix unit between the input and output matrix units, a variable routing channel for three-level switching is constructed, and non-blocking full switching is achieved through conflict-free combination of variable routing. At the same time, compared with the fixed routing matrix, the number of routes is reduced from M to M in the first-level switching. 2 ×M 2 The M is reduced to three-level switching 2 This greatly reduces the number of power dividers / switch units in the matrix switch, and reduces the composition scale of large-scale matrices from exponential level to multiple level.
Claims
1. A non-blocking full-switch matrix cascading method based on variable routing, It is characterized in that The following steps are involved: S1: Construct a matrix channel with variable routing: Take the M×M matrix scale as the basic matrix unit and construct M 2 ×M 2 The non-blocking full switching matrix comprises M input switching matrix units, M bridge switching matrix units and M output switching matrix units, and the input switching matrix unit SR, the bridge switching matrix unit QJ and the bridge switching matrix unit QJ, the output switching matrix unit SC are cross-connected one by one; S2: Perform complete routing planning for the cascade matrix channels: plan the complete routing combination of all channels, use the routing optimization algorithm, and select a set of mutually non-conflicting routes; S3: Calculate the switching configuration array of the matrix unit: calculate the switching configuration relationship of the input switching matrix unit SR, the bridge switching matrix unit QJ and the output switching matrix unit SC through the result of the routing planning; S4: Execute route planning and realize matrix switching: According to the switching configuration relationship of the matrix unit, calculate the switch control I / O in the board, control all switches to switch synchronously, and simultaneously open all channel planning routes to realize matrix switching.
2. According to the non-blocking full switching matrix cascading method based on variable routing as described in claim 1, It is characterized in that In step S1, the basic matrix unit based on the scale of M×M is constructed as M 2 ×M 2 The non-blocking full switching matrix comprises M input switching matrix units, M bridge switching matrix units and M output switching matrix units, and at least three groups of sub-matrix units are provided. 2 The M output ports and the M bridge switching matrix units QJ 2 The input ports are cross-connected one by one, and the M bridge switching matrix units QJ have M 2 M output ports and M output exchange SC 2 One-to-one cross connection between input ports.
3. According to the non-blocking full switching matrix cascading method based on variable routing as described in claim 1, It is characterized in that In step S3, the routing relationship cascade matrix is used to obtain the input array IN. M 2 Calculate the output array OUT M 2 The specific process is as follows: IN M 2 Split into , then the bridge input is: ; Will Split into , then the bridge output is: ; OUT QJ Split into , then the matrix output is: ; in, , , The internal connection relationship arrays of the input switching matrix unit SR, the bridge switching matrix unit QJ and the output switching matrix unit SC, and the cross-connection relationship arrays of the input switching matrix unit and the bridge switching matrix unit, and the bridge switching matrix unit and the output switching matrix unit; The SRQJ array value is defined as: ; Wherein, i is the output port of the input switching matrix unit SR, which is 1, 2, ..., M*M respectively; j is the input port of the bridge switching matrix unit QJ, which is 1, 2, ..., M*M respectively; The QJSC array value is defined as: ; Among them, j is the output port of the bridge switching matrix unit QJ, which is 1, 2, ..., M*M respectively; k is the input port of the output switch SC, which is 1, 2, ..., M*M respectively.
4. According to claim 2, a non-blocking full switching matrix cascading method based on variable routing, It is characterized in that The specific process of step S2 is as follows: S21: Define the input route R_SR according to the connection between the input switching matrix unit SR and the bridge switching matrix unit QJ: ; Wherein, m is the serial number of the input switching matrix unit SR, n is the serial number of the bridge switching matrix unit QJ, CHAR() is a function for converting numbers to characters, and SR_ is the input string; S22: Define the output route R_SC according to the connection between the bridge switching matrix unit QJ and the output switch SC: ; Where n is the bridge switch matrix unit QJ serial number, k is the output switch SC serial number, CHAR() is the function for converting numbers to characters; The matrix input port number IN and output port number OUT correspond to m and k: ; ; Among them, TRUNC() is the rounding function; S23: According to the routing plan, each channel selects one of the M routes. The selected route does not overlap with the routes selected by other channels or can overlap with them when the input port numbers are the same. The overlapping routes when the input port numbers are the same are shared routes.
5. According to the non-blocking full switching matrix cascading method based on variable routing as described in claim 2, It is characterized in that In step S2, the complete route from any input to any output in the cascade matrix, including the input route R_SR and the output route R_SC, passes through the bridge switching matrix units QJ numbered 1 to M, and there are M optional routes in total: ; When M=4, IN=5, OUT=6, then i=2, k=2, the optional routes are: 。 6. According to claim 3, a non-blocking full switching matrix cascading method based on variable routing, It is characterized in that Step S3 includes the following specific processes: Input channel number: IN, output channel number: OUT, input routing: R_SR, output routing: R_SC, where: the value range of IN and OUT is: 1~M, the value range of i, j, k is: 1~M, i, j, k correspond to the serial numbers of input, bridge, and output matrix units respectively; S31: Input switch matrix unit SR unit According to IN and j, the channel selection relationship of the input switching matrix unit SR unit i (input X, output Y) can be calculated: X=MOD((IN-1),M)+1;Y=j S32: Bridge Switching Matrix Unit QJ Unit According to i and k, the channel selection relationship of unit j of the bridge switching matrix unit QJ can be calculated (input X, output Y): X=i,Y=k S33: Output switching matrix unit SC unit The channel selection relationship (input X, output Y) of the output switching matrix unit SC unit k can be calculated based on j, k and OUT: X=j, Y= MOD((OUT-1),M)+1.
Citation Information
Patent Citations
Service routing algorithm applied to Clos switching structure
CN107948068A
A large-scale strictly non-blocking optical cross-connect matrix structure and a control method thereof
CN109151623A
Improved three-level CLOS routing algorithm based on FPGA crossover
CN109327410A
System and method for realizing non-blocking channel switching of cascade switching matrix
CN112769647A
Method for mapping a protection scheme over a switching system
EP2112790A1
Cited By
Non-blocking full switching matrix cascading method based on dual variable routing
CN119052159A
A non-blocking full-switch matrix cascading method based on dual variable routing
CN119052159B
Parallel hierarchical exhaustion-based on-line rapid self-checking method for variable routing matrix
CN120416138A
Method for performing multi-configuration simultaneous control on large-scale full switching matrix switch by using All-to-All broadcast algorithm
CN120949629A
Multi-input multi-output full-switching radio frequency switch matrix and control method thereof
CN121966528A
