CN114363251B - Low-complexity obstacle avoidance routing method and device for Benes network - Google Patents
Low-complexity obstacle avoidance routing method and device for Benes network Download PDFInfo
- Publication number
- CN114363251B CN114363251B CN202210275748.7A CN202210275748A CN114363251B CN 114363251 B CN114363251 B CN 114363251B CN 202210275748 A CN202210275748 A CN 202210275748A CN 114363251 B CN114363251 B CN 114363251B
- Authority
- CN
- China
- Prior art keywords
- input
- output
- layer
- switch unit
- switch
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/28—Routing or path finding of packets in data switching networks using route fault recovery
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a low-complexity obstacle avoidance routing method and device for a Benes network, for the Benes network, the connection priority of a subnet is determined by a fault unit; constructing an incidence relation among an input layer input port, an output layer output port and a subnet connection priority according to the connection condition of the input-output ports; determining the switch states of the input layer and the output layer according to the incidence relation; updating the input-output port connection condition in the subnet according to the switch state; when the input-output port connection condition reaches the middle level, determining the state of a middle level switch unit and ending the process; otherwise, taking the secondary edge level as the latest edge level, and analyzing the fault unit again. The low-complexity obstacle avoidance routing method can realize low-blocking-rate routing under the conditions of any scale of Benes network structures and any port configuration and realize optimization of resource utilization of the switch unit in view of the conditions that the design of the switch unit is not ideal and the matched control circuit and the electric package are not communicated.
Description
Technical Field
The invention relates to the field of network switching, in particular to a low-complexity obstacle avoidance routing method and device for a Benes network.
Background
A network switch is a network device for electrical/optical signal forwarding that can provide a path for any two accessed network switching nodes. The network switching architecture is divided into a limited blocking type, a rearrangement non-blocking type and a strict non-blocking type, and the number of the switch units used by the three architectures including the stage number is increased. The Benes network belongs to a rearrangement non-blocking type framework, namely when a new input-output port connection is established, all paths need to be re-planned, and the states of the switch units can meet all connection requirements after re-planning. Compared with a strict non-blocking network architecture, the Benes network has smaller scale and passes through fewer stages during routing. The device has low loss and small size due to the Benes network architecture, and is widely applied to the optical switching network structure, but the path switching algorithm under the network architecture is complex, so that important attention needs to be paid.
In view of this, d.c. Opferman et al from bell labs proposed a loop routing algorithm (locating algorithm), the basic idea of which is to determine the switching states of the edge stages first, and then determine the switching states of the internal secondary stages in sequence, until the intermediate stage. However, the algorithm is based on the premise of assuming the state of a certain switch unit, and if the assumption is wrong, retrospective correction is needed, so that the algorithm complexity is extremely high. On the basis, a, Karimi et al proposes an Input and Output Routing Algorithm (Input and Output Routing Algorithm), which solves the problem of collision backtracking, but can only work normally under the condition of full configuration, that is, all Input and Output ports are used, and the Algorithm lacks flexibility and usability under the condition of multiple scenes. A. The split Routing Algorithm (Division Routing Algorithm) proposed by Chakrabarty et al can implement path selection under any port configuration condition by adopting a scheme of filling in a switch unit table, but still faces the conflict backtracking problem under the special path selection condition.
In addition, the routing algorithm scheme does not consider the condition that the network unit has faults, and in practical application, because the switch unit is not ideal in design and the matched control circuit and the electric package are not connected, the blocking rate of the routing scheme is increased, and the network switching performance is greatly influenced.
Disclosure of Invention
In order to solve the defects of the prior art, the low-complexity obstacle avoidance routing method and device for the Benes network can be deployed in the Benes network with any scale, conflict backtracking is not required to be executed for different port configuration conditions, algorithm complexity is greatly reduced, the blocking rate of a routing scheme can be effectively reduced for the conditions of individual switch unit damage or non-ideal performance brought by various problems, and the performance and stability of network switching are improved, and the method adopts the following technical scheme:
a low-complexity obstacle avoidance routing method for a Benes network is provided, the Benes network is an array formed by a group of switch units, a plurality of arrays are nested in sequence, each array comprises an input layer and an output layer, and two subnets between the input layer and the output layer, each array input layer is connected with the input ends of the two subnets of the array, the output layer of each array is connected with the output ends of the two subnets of the array, each input layer, each output layer and each subnet respectively comprise a plurality of switch units arranged in an array, each switch unit is a 2 x 2 input-output switch unit, each subnet of the current layer array is respectively nested with a next layer array, for each array, the input layer of the array acting as an edge level input layer of the array and the output layer of the array acting as an edge level output layer of the array, said method comprising the steps of:
and sequentially executing the following operations on each array from the outer layer of the nest to the center until the innermost intermediate-level array of the nest is reached:
step S1, analyzing the array for fault units, and determining the connection priority of two sub-networks according to the analysis result;
step S2, according to the connection relationship between the input port of the input layer and the output port of the output layer and the connection priority of the two subnets, constructing the input port of the input layer, the output end of the output layer and the association relationship between the connection priority of the two subnets;
step S3, determining the switch state of each switch unit in the input layer as the edge level and the switch state of each switch unit in the output layer as the edge level according to the association relationship, wherein the switch states are a cross state or a parallel state;
step S4, according to the switch state of the input layer as the edge level and the switch state of the output layer as the edge level, updating the paths of the two subnets to update the connection relationship of the input-output ports in the two subnets to determine the state of the subsequent internal switch unit;
step S5, judging whether the input-output port connection condition reaches the middle level, if yes, determining the state of the middle level switch unit and ending the process; if not, the next edge level is set as the latest edge level, and the input-output port connection status at this time is analyzed, returning to step S1. From the Benes network, the initial edge level is an input layer and an output layer, the secondary edge level is the edge level of the sub-network, namely two columns of switch units on the left and right of the sub-network, the edge level is gradually increased from two sides to the middle layer, and each layer avoids the sub-network with more faults through the input end and the output end of the layer and the corresponding sub-network priority, so that the performance and the stability of network switching are replaced.
Further, the step S1 specifically includes:
determining the number of subnet fault units and the number of edge level fault units in the array, wherein the subnet fault units comprise failed switch units in a corresponding subnet, and the edge level fault units comprise failed switch units included in an input layer of the array and failed switch units included in an output layer of the array; according to the number of the subnet fault units and the number of the edge level fault units in the array, judging the connection priority of the two subnets as the priority filled in a subnet connection condition table, wherein the priority is higher for subnets with few subnet fault units; when the number of the subnet fault units is the same, the priority of the subnet with fewer edge fault units is higher; and when the number of the subnet fault units and the number of the edge-level fault units are the same, a subnet with higher priority is assigned by default or randomly.
Further, the switch unit includes an upper input port, a lower input port, and an upper output port, where the two subnets are A, B subnets, and the step S2 specifically includes:
establishing input-output association relations of 00, 01, 10 and 11 between an input port of an input layer switch unit and an output port of an output layer switch unit according to the connection relation between the input port of an input layer and the output port of the output layer, wherein 0 or 1 of a first bit corresponds to an upper input port or a lower input port of the input layer switch unit, and 0 or 1 of a second bit corresponds to an upper output port or a lower output port of the output layer switch unit; and establishing a priority association relation among A, B, AB and BA through an input-output association relation according to the priorities of the two subnets, wherein A represents that the input-output association relation is connected through an A subnet, B represents that the input-output association relation is connected through a B subnet, AB represents that two input ports of one input layer switch unit are completely corresponding to two output ports of one output layer switch unit, and at the moment, the input-output association relation is connected through the A subnet with higher priority and then connected through the B subnet, and BA is opposite. In the filling process of the switch unit table, the problem of path conflict backtracking is avoided by designing a novel filling scheme, and the algorithm complexity is reduced.
Further, in step S2, a switch unit table is established to record the association relationship between the input port of the input layer, the output port of the output layer, and the connection priority of the subnet, where a row indicates an input layer switch unit and a column indicates an output layer switch unit, and in the process of searching for a next switch unit to be filled, it is first determined whether a corresponding port of the current input layer switch unit needs to be filled; then judging whether the current output layer switch unit has a corresponding port to be filled; and if no port needing to be filled in the row and the column corresponding to the switch unit exists, selecting the next switch unit to be filled in according to the port input sequence.
Further, in the association relationship, 00+ a, 01+ a, 10+ B, 11+ B, 10+ AB, 11+ AB, 00+ BA, 01+ BA correspond to the parallel state of the input layer switch unit, and 10+ a, 11+ a, 00+ B, 01+ B, 00+ AB, 01+ AB, 10+ BA, 11+ BA correspond to the cross state of the input layer switch unit; wherein 00+ A, 10+ A, 01+ B, 11+ B, 01+ AB, 11+ AB, 00+ BA, 10+ BA corresponds to the parallel state of the switch unit of the output layer; 01+ A, 11+ A, 00+ B, 10+ B, 00+ AB, 10+ AB, 01+ BA, 11+ BA corresponds to the cross state of the switching unit of the output layer.
A low-complexity obstacle avoidance routing device facing a Benes network comprises the Benes network, a fault unit analysis module, a switch unit table filling module, a switch state determination module and a subnet path updating module, wherein the Benes network is an array formed by a group of switch units, a plurality of arrays are sequentially nested, each array comprises an input layer, an output layer and two subnets between the input layer and the output layer, the input layer of each array is respectively connected with the input ends of the two subnets of the array, the output layer of each array is respectively connected with the output ends of the two subnets of the array, each input layer, each output layer and each subnet respectively comprise a plurality of switch units arranged in an array, each switch unit is a 2 input-output switch unit, each subnet of the array at the current layer is respectively nested with a next layer of array, and for each array, the input layer of the array is used as an edge-level input layer of the array, the output layer of the array is used as an edge-level output layer of the array;
and (3) sequentially and circularly calling the following modules for each array from the outer layer of the nesting to the center until the innermost intermediate-level array of the nesting is reached:
the fault unit analysis module is used for analyzing the fault units of the array and judging the connection priority of the two subnets according to the analysis result;
the switch unit table filling module is used for constructing an input port of an input layer, an output end of an output layer and an incidence relation of connection priorities of two subnets according to the connection relation of the input port of the input layer and the output port of the output layer and the connection priorities of the two subnets;
the switch state determining module is used for determining the switch state of each switch unit in the input layer serving as the edge level and the switch state of each switch unit in the output layer serving as the edge level according to the association relationship, wherein the switch states are in a cross state or a parallel state;
and the subnet path updating module is used for updating paths of the two subnets according to the switch state of the input layer serving as the edge level and the switch state of the output layer serving as the edge level so as to update the connection relation of the input-output ports in the two subnets and determine the state of a subsequent internal switch unit.
Further, the switch unit includes an upper input port, a lower input port, an upper output port, and a lower output port, where the two subnets are A, B subnets, the switch unit table filling module establishes a connection relationship between the input port of the input layer and the output port of the output layer and a connection priority of the two subnets by constructing a switch unit table, where a row of the switch unit table indicates an input layer unit switch and a column indicates an output layer unit switch, and the switch unit table includes an input-output port table and a subnet connection condition table.
Further, the input-output port table establishes input-output association relations of 00, 01, 10 and 11 between the input port of the input layer switch unit and the output port of the output layer switch unit according to the connection relation between the input port of the input layer and the output port of the output layer, and fills in the row-column intersection of the input-output port table, wherein 0 or 1 of a first bit corresponds to the upper input port or the lower input port of the input layer switch unit, and 0 or 1 of a second bit corresponds to the upper output port or the lower output port of the output layer switch unit.
Further, according to the priority and the input-output association relationship of the two subnets, the subnet connection condition table fills corresponding priority association relationships a, B, AB and BA at the row-column intersection, wherein a represents that the input-output association relationship is connected through a subnet a, B represents that the input-output association relationship is connected through a subnet B, AB represents that two input ports of an input layer switch unit and two output ports of an output layer switch unit completely correspond, at this time, the input-output association relationship is connected through a subnet a with higher priority and then through a subnet B, and BA is opposite.
Further, the array formed by the switch units is a cross switch matrix, the switch state determining module determines the cross/parallel state of the cross switch matrix according to a switch unit table, and in the association relationship, 00+ a, 01+ a, 10+ B, 11+ B, 10+ AB, 11+ AB, 00+ BA, 01+ BA corresponds to the parallel state of the switch units of the input layer, and 10+ a, 11+ a, 00+ B, 01+ B, 00+ AB, 01+ AB, 10+ BA, 11+ BA corresponds to the cross state of the switch units of the input layer; wherein 00+ A, 10+ A, 01+ B, 11+ B, 01+ AB, 11+ AB, 00+ BA, 10+ BA corresponds to the parallel state of the switch unit of the output layer; 01+ A, 11+ A, 00+ B, 10+ B, 00+ AB, 10+ AB, 01+ BA, 11+ BA corresponds to the cross state of the switching unit of the output layer.
The invention has the advantages and beneficial effects that:
according to the low-complexity obstacle avoidance routing method and device for the Benes network, based on the separation routing algorithm, through designing a novel switch unit filling scheme, the conflict backtracking problem of the separation routing algorithm is solved, and the complexity of the routing algorithm is greatly reduced; aiming at the problem of the switching unit fault, the subnet fault unit analysis module is used for judging the priority of subnet connection, so that the switching unit with damage or unsatisfactory performance can be effectively avoided, the blocking rate of a routing scheme is reduced, and the performance and the stability of network switching are improved.
Drawings
Fig. 1 is a flowchart of an obstacle avoidance routing method according to an embodiment of the present invention.
Fig. 2 is a 16x16 Benes network array diagram in an embodiment of the present invention.
Fig. 3 is a table of switch cells for 16x16 in an embodiment of the present invention.
Fig. 4 is a structural diagram of an obstacle avoidance routing apparatus according to an embodiment of the present invention.
Fig. 5 is a diagram of 16x16 Benes network input-output port connections and faulty unit locations in an embodiment of the present invention.
Fig. 6 is a routing diagram under a non-obstacle avoidance routing algorithm in an embodiment of the present invention.
Fig. 7 is a table of switch unit input-output ports in an embodiment of the present invention.
Fig. 8 is a table of subnet connections of switch cells in an embodiment of the invention.
Fig. 9 is a state diagram of an edge level switch cell in an embodiment of the invention.
Fig. 10 is a routing diagram under the obstacle avoidance routing algorithm in the embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present invention, are given by way of illustration and explanation only, not limitation.
As shown in fig. 1, a low-complexity obstacle avoidance routing method facing a Benes network, where the Benes network is an array formed by a group of switch units, as shown in fig. 2, a plurality of arrays are sequentially nested, each array includes an input layer, an output layer, and A, B subnets between the input layer and the output layer, the input layer of each array is connected to the input ends of the two subnets of the array, the output layer of each array is connected to the output ends of the two subnets of the array, each input layer, each output layer, and each subnet respectively includes a plurality of switch units arranged in an array, each switch unit is a 2 × 2 input-output switch unit, each subnet of a layer array is nested with a next layer of array, for each array, the input layer of the array serves as an edge level input layer of the array, and the output layer of the array serves as an edge level output layer of the array, A. b, the two subnets are determined by the state of a switch unit at an edge level (comprising an input layer and an output layer), and the method comprises the following steps:
step S1, analyzing the array for fault units, and determining the connection priority of two sub-networks according to the analysis result;
determining the number of subnet fault units and the number of edge level fault units in the array, wherein the subnet fault units comprise failed switch units in a corresponding subnet, and the edge level fault units comprise failed switch units in an input layer of the array and failed switch units in an output layer of the array; according to the number of the subnet fault units and the number of the edge level fault units in the array, judging the connection priority of the two subnets as the priority filled in a subnet connection condition table, wherein the priority is higher for subnets with few subnet fault units; when the number of the subnet fault units is the same, the priority of the subnet with fewer edge fault units is higher; and when the number of the subnet fault units and the number of the edge-level fault units are the same, a subnet with higher priority is assigned by default or randomly.
Specifically, the filling priority of the sub-network connection condition table in the next step is determined according to the statistics of the sub-network fault units and the statistics of the sub-network edge level fault units. When the sub-networks with fewer fault units fill in the sub-network connection condition table, the priority is higher, and when the fault units in the sub-networks are the same, the priority is higher when the sub-network connection condition table is filled in the sub-networks with fewer edge-level fault units. If both are the same, the default A subnet is higher in priority than the B subnet.
Step S2, according to the connection relationship between the input port of the input layer and the output port of the output layer and the connection priority of the two subnets, constructing the input port of the input layer, the output end of the output layer and the association relationship between the connection priority of the two subnets;
the switch unit comprises an upper input port, a lower input port, an upper output port and a lower output port, wherein the two subnets are A, B subnets respectively, and establish input-output association relations of 00, 01, 10 and 11 between the input port of the input layer switch unit and the output port of the output layer switch unit according to the connection condition of the input-output ports, wherein 0 or 1 of a first bit corresponds to the upper input port or the lower input port of the input layer switch unit, and 0 or 1 of a second bit corresponds to the upper output port or the lower output port of the output layer switch unit; establishing a priority association relation among A, B, AB and BA through an input-output association relation according to the priority of the subnets, wherein A represents that the input-output association relation is connected through an A subnet, B represents that the input-output association relation is connected through a B subnet, AB represents that two input ports of an input layer switch unit are completely corresponding to two output ports of an output layer switch unit, and at the moment, the input-output association relation is connected through the A subnet with higher priority and then connected through the B subnet, and BA is opposite.
Establishing a switch unit table for recording the incidence relation of input ports of an input layer, output ends of an output layer and subnet connection priorities, wherein rows represent input layer unit switches, columns represent output layer unit switches, and whether a corresponding port of a current input layer switch unit needs to be filled is judged firstly in the process of searching a next filled switch unit; then judging whether the current output layer switch unit has a corresponding port to be filled; and if no port needing to be filled in the row and the column corresponding to the switch unit exists, selecting the next switch unit to be filled in according to the port input sequence.
Specifically, an edge-level switch cell input-output port table and a subnet connection condition table, such as the switch cell table filled in the 16 × 16 case shown in fig. 3, are respectively filled according to the input-output port connection condition, and the filled form positions indicate the connection condition of the input-output layer switch cells. Each switch unit as shown in fig. 2 includes two input ports and two output ports, A, B two subnets are connected respectively, and 00, 01, 10, 11 are filled in the input-output port table as shown in fig. 3 according to the input-output ports being the upper ports or the lower ports of the switch unit. Wherein the first bit 0 or 1 corresponds to the upper port or the lower port of the input layer switch unit, and the second bit 0 or 1 corresponds to the upper port or the lower port of the output layer switch unit. Similarly, according to the subnet priority condition in step S1, filling a, B, AB, BA in the subnet connection condition table as shown in fig. 3, it should be noted that since the input and output of the edge level switch unit are respectively connected to A, B two subnets, there is no two a or two B condition in each row and column in the subnet connection condition table, wherein if two ports of one input level switch unit and two ports of one output level switch unit correspond to each other, an AB or BA element appears, AB indicates that between the input and output switches corresponding to a certain grid in the subnet connection condition table, the subnet a with higher priority is selected first, since the input and output switches both have two ports, and the other port is respectively connected to A, B two subnets, therefore, the same switch unit can only select the subnet B, and vice versa. It should be noted that, in the process of searching for the next switch unit to be filled, it is first determined whether there is a corresponding port to be filled in (i.e. one row in fig. 3) in the current switch unit of the input layer; then, whether a corresponding port of the current output layer switch unit needs to be filled in (namely, a column in fig. 3) is judged; and if no port needing to be filled in the row and the column corresponding to the switch unit exists, selecting the next switch unit to be filled in according to the port input sequence.
Step S3, determining the switch state of each switch unit in the input layer as the edge level and the switch state of each switch unit in the output layer as the edge level according to the association relationship, wherein the switch states are a cross state or a parallel state;
the switch states include a Cross/parallel (Cross/bar) state, wherein 00+ A, 01+ A, 10+ B, 11+ B, 10+ AB, 11+ AB, 00+ BA, 01+ BA corresponds to the parallel state of the switch units of the input layer, and 10+ A, 11+ A, 00+ B, 01+ B, 00+ AB, 01+ AB, 10+ BA, 11+ BA corresponds to the Cross state of the switch units of the input layer; wherein 00+ A, 10+ A, 01+ B, 11+ B, 01+ AB, 11+ AB, 00+ BA, 10+ BA corresponds to the parallel state of the switch unit of the output layer; 01+ A, 11+ A, 00+ B, 10+ B, 00+ AB, 10+ AB, 01+ BA, 11+ BA corresponds to the cross state of the switching unit of the output layer.
Specifically, the state of the edge-level switch unit is determined according to the input-output port table and the subnet connection condition table of the edge-level switch unit obtained in the previous step. Where 00+ a, 01+ a, 10+ B, 11+ B, 10+ AB, 11+ AB, 00+ BA, 01+ BA correspond to the input-level switch unit Bar state, where only the input port, i.e. the first bit 0 or 1, and the following selected sub-networks are considered for the input-level switch unit, e.g.: the first bit is 0, corresponding to sub-network a, and the first bit is 1, the following sub-network corresponds to B, and the status of AB, BA is also referred to, for example: 11+ AB, because the data of the subnet connection condition table is filled in an overlay mode, the state A is overlaid by the state B, so the state 11+ AB is equivalent to the state 11+ B, and the condition of the state BA is similar; the 10+ A, 11+ A, 00+ B, 01+ B, 00+ AB, 01+ AB, 10+ BA and 11+ BA correspond to the Cross state of the switch unit of the input layer; wherein 00+ A, 10+ A, 01+ B, 11+ B, 01+ AB, 11+ AB, 00+ BA, 10+ BA corresponds to the state of the Bar of the switching unit of the output layer; 01+ A, 11+ A, 00+ B, 10+ B, 00+ AB, 10+ AB, 01+ BA, 11+ BA corresponds to the Cross state of the switching unit of the output layer.
The bar state means that the transverse input is connected to the longitudinal output, and the longitudinal input is connected to the transverse output; cross states refer to horizontal input connected to horizontal output and vertical input connected to vertical output.
Step S4, according to the switch state of the input layer as the edge level and the switch state of the output layer as the edge level, updating the paths of the two subnets to update the connection relationship of the input-output ports in the two subnets to determine the state of the subsequent internal switch unit;
specifically, the subnet path condition, i.e., the input-output port connection condition in the subnet is updated according to the input-output port connection condition and the state of the edge level switch unit obtained in step S3.
Step S5, judging whether the input-output port connection condition reaches the middle level, if yes, determining the state of the middle level switch unit and ending the process; if not, the next edge level is set as the latest edge level, and the input-output port connection status at this time is analyzed, returning to step S1.
As shown in fig. 4, a Benes network-oriented low-complexity obstacle avoidance routing device includes a Benes network, a fault unit analysis module, a switch unit table filling module, a switch state determination module, and a subnet path update module, where the Benes network is an array formed by a group of switch units, and includes an input layer, an output layer, and two subnets between the input layer and the output layer, the input layer is connected to input ends of the two subnets, and the output layer is connected to output ends of the two subnets; the switch unit comprises an upper input port, a lower input port, an upper output port and a lower output port, and the two subnets are A, B subnets respectively; the array formed by the switch units is a cross switch matrix.
And the fault unit analysis module is used for judging the connection priority of the sub-network.
The switch unit table filling module is used for constructing an incidence relation between an input layer input port, an output layer output port and the subnet connection priority according to the input-output port connection condition and the subnet connection priority; and the switch unit table filling module is used for establishing the incidence relation among the input port of the input layer, the output port of the output layer and the connection priority of the subnets by constructing a switch unit table, wherein the row of the switch unit table represents the unit switch of the input layer, the column represents the unit switch of the output layer, and the switch unit table comprises an input-output port table and a subnet connection condition table. And the input-output port table fills corresponding input- output association relations 00, 01, 10 and 11 at the row-column intersection according to the connection condition of the input-output ports, wherein 0 or 1 of a first bit corresponds to an upper input port or a lower input port of the input-layer switch unit, and 0 or 1 of a second bit corresponds to an upper output port or a lower output port of the output-layer switch unit. And a network connection condition table, wherein corresponding priority association relations A, B, AB and BA are filled in the row-column intersection according to the priority of the subnet and the input-output association relation, A represents that the input-output association relation is connected through the subnet A, B represents that the input-output association relation is connected through the subnet B, AB represents that two input ports of an input layer switch unit and two output ports of an output layer switch unit are completely corresponding, at the moment, the input-output association relation is connected through the subnet A with higher priority and then connected through the subnet B, and BA is opposite.
The switch state determining module is used for determining the switch states of the input layer and the output layer which are used as edge levels according to the incidence relation; the switch state determining module determines the cross/parallel state of the cross switch matrix according to the switch unit table, wherein 00+ A, 01+ A, 10+ B, 11+ B, 10+ AB, 11+ AB, 00+ BA, 01+ BA corresponds to the parallel state of the switch units of the input layer, and 10+ A, 11+ A, 00+ B, 01+ B, 00+ AB, 01+ AB, 10+ BA, 11+ BA corresponds to the cross state of the switch units of the input layer; wherein 00+ A, 10+ A, 01+ B, 11+ B, 01+ AB, 11+ AB, 00+ BA, 10+ BA corresponds to the parallel state of the switch unit of the output layer; 01+ A, 11+ A, 00+ B, 10+ B, 00+ AB, 10+ AB, 01+ BA, 11+ BA corresponds to the output layer switch cell crossing state.
The subnet path updating module updates the subnet path according to the switch state, and updates the input-output port connection condition in the subnet so as to determine the state of a subsequent internal switch unit; if the input-output port connection condition reaches the middle stage, determining the state of a middle stage switch unit; if not, the secondary edge stage is taken as the latest edge stage, the input-output port connection condition at the moment is analyzed, and the fault unit analysis module is returned.
In the embodiment of the present invention, an implementation is provided for a case where a 16x16 Benes network is not in full configuration and has a specific fault, as shown in fig. 5, the input-output port connection conditions ((0, 3), (2,14), (3,1), (13, 5)) are marked when the location of the fault unit is not in full configuration. When the failure unit analysis module is not deployed, that is, a route obtained by using a non-obstacle avoidance routing algorithm is as shown in fig. 6, because two failure units exist on a routing path, the performance and stability of network switching are greatly affected.
When the low-complexity obstacle avoidance routing method is adopted, the number of the subnet fault units and the number of the subnet edge level fault units are analyzed to determine the subnet connection priority. As shown in fig. 5, when determining the edge level switch status, A, B shows the same number of fault cells in the subnet, and since the a subnet has a larger number of corresponding edge level fault cells, the B subnet has higher priority than the a subnet when filling the switch cell table. The switch unit input-output port table and the subnet connection condition table obtained according to the edge level switch unit table filling scheme are respectively shown in fig. 7 and 8, and the corresponding row and column filling elements in the tables indicate that the edge level (input and output layer) switch units at the corresponding positions are used. After that, the states of the edge-level (first and seventh levels) switch units shown in fig. 9 can be obtained according to the element filling conditions of the input-output port table and the subnet connection condition table, where 0 indicates that the switch unit is not used, 1 indicates that the switch unit is Bar state, and 2 indicates that the switch unit is Cross state. And obtaining the input-output port conditions of the secondary edge stages (the second stage and the sixth stage) according to the input-output port conditions and the edge stage switch states. At this time, since the connection condition of the input-output ports does not reach the middle level (fourth level), the secondary edge level needs to be regarded as the edge level and the above steps need to be repeated for different sub-networks, and finally, the routing diagram obtained by the obstacle avoidance routing method of the present invention is shown in fig. 10.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A low-complexity obstacle avoidance routing method for a Benes network is provided, the Benes network is an array formed by a group of switch units, a plurality of arrays are nested in sequence, each array comprises an input layer and an output layer, and two subnets between the input layer and the output layer, the input layer of each array is connected with the input ends of the two subnets of the array, the output layer of each array is connected with the output ends of the two subnets of the array, each input layer, each output layer and each subnet respectively comprise a plurality of switch units arranged in an array, each switch unit is a 2 x 2 input-output switch unit, each subnet of the current layer array is respectively nested with a next layer array, for each array, the input layer of the array serves as an edge level input layer of the array, and the output layer of the array serves as an edge level output layer of the array, wherein the method comprises the following steps:
and sequentially executing the following operations on each array from the outer layer of the nest to the center until the innermost intermediate-level array of the nest is reached:
step S1, analyzing the array for fault units, and determining the connection priority of two sub-networks according to the analysis result;
step S2, according to the connection relationship between the input port of the input layer and the output port of the output layer and the connection priority of the two subnets, constructing the input port of the input layer, the output port of the output layer and the association relationship between the connection priority of the two subnets;
step S3, determining the switch state of each switch unit in the input layer as the edge level and the switch state of each switch unit in the output layer as the edge level according to the association relationship, wherein the switch states are a cross state or a parallel state;
step S4, updating the paths of the two sub-networks according to the switch state of the input layer as the edge level and the switch state of the output layer as the edge level, so as to update the connection relation of the input-output ports in the two sub-networks;
step S5, judging whether the input-output port connection condition reaches the middle level, if yes, determining the state of the middle level switch unit and ending the process; if not, the next edge level is set as the latest edge level, and the input-output port connection status at this time is analyzed, returning to step S1.
2. The Benes network-oriented low-complexity obstacle avoidance routing method according to claim 1, wherein: the step S1 specifically includes:
determining the number of subnet fault units and the number of edge level fault units in the array, wherein the subnet fault units comprise failed switch units in a corresponding subnet, and the edge level fault units comprise failed switch units in an input layer of the array and failed switch units in an output layer of the array;
judging the connection priority of the two subnets according to the number of subnet fault units and the number of edge level fault units in the array;
the sub-networks with few sub-network fault units have higher priority; when the number of the subnet fault units is the same, the priority of the subnet with fewer edge fault units is higher; and when the number of the subnet fault units and the number of the edge-level fault units are the same, a subnet with higher priority is assigned by default or randomly.
3. The Benes network-oriented low-complexity obstacle avoidance routing method according to claim 1, wherein: the switch unit includes an upper input port, a lower input port, and an upper output port and a lower output port, the two subnets are A, B subnets, and the step S2 specifically includes:
establishing input-output association relations of 00, 01, 10 and 11 between an input port of an input layer switch unit and an output port of an output layer switch unit according to the connection relation between the input port of an input layer and the output port of the output layer, wherein 0 or 1 of a first bit corresponds to an upper input port or a lower input port of the input layer switch unit, and 0 or 1 of a second bit corresponds to an upper output port or a lower output port of the output layer switch unit;
and establishing a priority association relation among A, B, AB and BA through an input-output association relation according to the priority of the two subnets, wherein A represents that the input-output association relation is connected through an A subnet, B represents that the input-output association relation is connected through a B subnet, AB represents that two input ports of an input layer switch unit and two output ports of an output layer switch unit are completely corresponding, at the moment, the input-output association relation is connected through the A subnet with higher priority and then connected through the B subnet, and BA is opposite.
4. The method for low-complexity obstacle avoidance routing for the Benes network as claimed in claim 1, wherein: in step S2, a switch unit table is established to record the association relationship between the input port of the input layer, the output port of the output layer, and the connection priority of the subnet, where a row indicates an input layer switch unit and a column indicates an output layer switch unit, and in the process of searching for a next switch unit to be filled, it is first determined whether a corresponding port of the current input layer switch unit needs to be filled; then judging whether the current output layer switch unit has a corresponding port to be filled; and if no port which needs to be filled in is arranged in the row and the column corresponding to the switch unit, selecting the next switch unit to be filled in according to the port input sequence.
5. The Benes network-oriented low-complexity obstacle avoidance routing method according to claim 3, wherein: in the association relationship, 00+ A, 01+ A, 10+ B, 11+ B, 10+ AB, 11+ AB, 00+ BA, 01+ BA corresponds to the parallel state of the switch units of the input layer, and 10+ A, 11+ A, 00+ B, 01+ B, 00+ AB, 01+ AB, 10+ BA, 11+ BA corresponds to the cross state of the switch units of the input layer; wherein 00+ A, 10+ A, 01+ B, 11+ B, 01+ AB, 11+ AB, 00+ BA, 10+ BA corresponds to the parallel state of the switch unit of the output layer; 01+ A, 11+ A, 00+ B, 10+ B, 00+ AB, 10+ AB, 01+ BA, 11+ BA corresponds to the cross state of the switching unit of the output layer.
6. A low-complexity obstacle avoidance routing device facing a Benes network comprises the Benes network, a fault unit analysis module, a switch unit table filling module, a switch state determination module and a subnet path updating module, wherein the Benes network is an array formed by a group of switch units, a plurality of arrays are sequentially nested, each array comprises an input layer, an output layer and two subnets between the input layer and the output layer, the input layer of each array is respectively connected with the input ends of the two subnets of the array, the output layer of each array is respectively connected with the output ends of the two subnets of the array, each input layer, each output layer and each subnet respectively comprise a plurality of switch units arranged in an array, each switch unit is a 2 input-output switch unit, each subnet of the array at the current layer is respectively nested with a next layer of array, and for each array, the input layer of the array is used as an edge-level input layer of the array, the output layer of the array is used as an edge-level output layer of the array, and is characterized in that:
and (3) sequentially and circularly calling the following modules for each array from the outer layer of the nesting to the center until the innermost intermediate-level array of the nesting is reached:
the fault unit analysis module is used for analyzing the fault units of the array and judging the connection priority of the two subnets according to the analysis result;
the switch unit table filling module is used for constructing an input port of an input layer, an output port of an output layer and an incidence relation of the connection priorities of the two subnets according to the connection relation of the input port of the input layer and the output port of the output layer and the connection priorities of the two subnets;
the switch state determining module is used for determining the switch state of each switch unit in the input layer serving as the edge level and the switch state of each switch unit in the output layer serving as the edge level according to the association relationship, wherein the switch states are in a cross state or a parallel state;
the subnet path updating module is used for updating paths of the two subnets according to the switch state of the input layer serving as the edge level and the switch state of the output layer serving as the edge level so as to update the connection relation of the input-output ports in the two subnets; the subnet route updating module updates the subnet route according to the switch state, and updates the input-output port connection condition in the subnet so as to determine the state of a subsequent internal switch unit; if the input-output port connection condition reaches the middle level, determining the state of a middle level switch unit; if not, the secondary edge stage is taken as the latest edge stage, the input-output port connection condition at the moment is analyzed, and the fault unit analysis module is returned.
7. The Benes network-oriented low-complexity obstacle avoidance routing device according to claim 6, wherein: the switch unit comprises an upper input port, a lower input port, an upper output port and a lower output port, the two subnets are A, B subnets respectively, the switch unit table filling module is used for establishing the connection relation between the input port of the input layer and the output port of the output layer and the connection priority of the two subnets by constructing a switch unit table, wherein the row of the switch unit table represents the unit switch of the input layer, the column represents the unit switch of the output layer, and the switch unit table comprises an input-output port table and a subnet connection condition table.
8. The Benes network-oriented low-complexity obstacle avoidance routing device according to claim 7, wherein: the input-output port table establishes input-output association relations of 00, 01, 10 and 11 between the input port of the input layer switch unit and the output port of the output layer switch unit according to the connection relation of the input port of the input layer and the output port of the output layer, and fills in the row-column intersection of the input-output port table, wherein 0 or 1 of a first bit corresponds to the upper input port or the lower input port of the input layer switch unit, and 0 or 1 of a second bit corresponds to the upper output port or the lower output port of the output layer switch unit.
9. The Benes network-oriented low-complexity obstacle avoidance routing device according to claim 8, wherein: and according to the priority and the input-output association relation of the two subnets, filling corresponding priority association relations A, B, AB and BA at the row-column intersection, wherein A represents that the input-output association relation is connected through the subnet A, B represents that the input-output association relation is connected through the subnet B, AB represents that two input ports of an input layer switch unit are completely corresponding to two output ports of an output layer switch unit, and at the moment, the input-output association relation is connected through the subnet A with higher priority and then connected through the subnet B, and BA is opposite.
10. The Benes network-oriented low-complexity obstacle avoidance routing device according to claim 9, wherein: the array formed by the switch units is a cross switch matrix, the switch state determining module determines the cross/parallel state of the cross switch matrix according to a switch unit table, and in the incidence relation, 00+ A, 01+ A, 10+ B, 11+ B, 10+ AB, 11+ AB, 00+ BA, 01+ BA corresponds to the parallel state of the switch units of the input layer, and 10+ A, 11+ A, 00+ B, 01+ B, 00+ AB, 01+ AB, 10+ BA, 11+ BA corresponds to the cross state of the switch units of the input layer; wherein 00+ A, 10+ A, 01+ B, 11+ B, 01+ AB, 11+ AB, 00+ BA, 10+ BA corresponds to the parallel state of the switch unit of the output layer; 01+ A, 11+ A, 00+ B, 10+ B, 00+ AB, 10+ AB, 01+ BA, 11+ BA corresponds to the cross state of the switching unit of the output layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210275748.7A CN114363251B (en) | 2022-03-21 | 2022-03-21 | Low-complexity obstacle avoidance routing method and device for Benes network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210275748.7A CN114363251B (en) | 2022-03-21 | 2022-03-21 | Low-complexity obstacle avoidance routing method and device for Benes network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114363251A CN114363251A (en) | 2022-04-15 |
| CN114363251B true CN114363251B (en) | 2022-07-08 |
Family
ID=81094437
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210275748.7A Active CN114363251B (en) | 2022-03-21 | 2022-03-21 | Low-complexity obstacle avoidance routing method and device for Benes network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114363251B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115379316B (en) * | 2022-08-03 | 2024-04-05 | 无锡芯光互连技术研究院有限公司 | Pipelined BENES network route solving hardware accelerating device |
| CN115379318B (en) * | 2022-08-03 | 2024-04-05 | 无锡芯光互连技术研究院有限公司 | BENES network route speculative solving method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102347898A (en) * | 2010-07-26 | 2012-02-08 | 丛林网络公司 | Methods and apparatus related to route selection within a network |
| CN107431535A (en) * | 2015-04-27 | 2017-12-01 | 华为技术有限公司 | The transmission path method for building up and device exchanged for optical signal |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5940389A (en) * | 1997-05-12 | 1999-08-17 | Computer And Communication Research Laboratories | Enhanced partially self-routing algorithm for controller Benes networks |
| CN100350292C (en) * | 2004-06-24 | 2007-11-21 | 北京邮电大学 | Two-dimensional micro-electromechanical optical switch using shuffle Benes interstage join rule and method thereof |
| CN101014012B (en) * | 2007-02-02 | 2010-08-25 | 华为技术有限公司 | Benes network and variable length packet distributing method and equipment |
| CN100440847C (en) * | 2007-07-13 | 2008-12-03 | 清华大学 | P2i interconnection structure based extendable router |
| US9916274B2 (en) * | 2015-07-23 | 2018-03-13 | Cavium, Inc. | Apparatus and method for on-chip crossbar design in a network switch using benes network |
| CN111698583B (en) * | 2020-05-28 | 2022-07-01 | 重庆邮电大学 | On-chip optical switching network architecture and routing algorithm for wavelength-mode hybrid multiplexing |
| CN113938772B (en) * | 2021-09-29 | 2022-06-14 | 电子科技大学 | Multicast control method of optical switch array based on Benes structure |
-
2022
- 2022-03-21 CN CN202210275748.7A patent/CN114363251B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102347898A (en) * | 2010-07-26 | 2012-02-08 | 丛林网络公司 | Methods and apparatus related to route selection within a network |
| CN107431535A (en) * | 2015-04-27 | 2017-12-01 | 华为技术有限公司 | The transmission path method for building up and device exchanged for optical signal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114363251A (en) | 2022-04-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114363251B (en) | Low-complexity obstacle avoidance routing method and device for Benes network | |
| EP2186268B1 (en) | Balanced redundancy cross-connect communications interconnect and method | |
| US5093824A (en) | Distributed protocol for improving the survivability of telecommunications trunk networks | |
| US10021025B2 (en) | Distributed determination of routes in a vast communication network | |
| CA2636834C (en) | Large-scale packet switch | |
| JPH0638256A (en) | Method for selection of connecting route of cross connection network | |
| CN102740177B (en) | Non-blocking expandable multistage photoswitch array and working method thereof | |
| CN113938772B (en) | Multicast control method of optical switch array based on Benes structure | |
| US6567576B2 (en) | Optical switch matrix with failure protection | |
| US8391282B1 (en) | Systems and methods for overlaid switching networks | |
| JPH07322311A (en) | Switching network,its exchange device and operation of switching network | |
| US5343194A (en) | Methods and system for immediately connecting and reswitching digital cross-connect networks | |
| US20190379953A1 (en) | Intelligence- defined optical tunnel network system controller and control method thereof | |
| US20140064144A1 (en) | Switching System Employing Independent Data Switches Connecting Orthogonal Sets of Nodes | |
| CN107079205B (en) | PIC type optical switch matrix routing configuration method and device | |
| US6901071B2 (en) | Row upgrade for a scalable switching network | |
| JP2003009194A (en) | Optical crossconnect system and its controller and controlling method | |
| CN115379318A (en) | BENES network route speculative solution method and device | |
| Varsou et al. | Waveband protection mechanisms in hierarchical optical networks | |
| JPH0630455A (en) | Matrix retrieval network | |
| KR100262474B1 (en) | Methods and system for immediately connecting and reswitching digital cross-connect networks | |
| CN109151623A (en) | A kind of large-scale strict non-blockage light-crossing connection matrix structure and its control method | |
| US20080143473A1 (en) | Digital Cross-Connect Path Selection Method | |
| EP2285046B1 (en) | Method and apparatus for realizing interaction of optical channel data unit protection tangency rings | |
| US20230198256A1 (en) | Redundancy matrix-based system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |