A method for updating traffic in power communication network based on pre-routing
By using pre-routing and seven-tuple model to optimize traffic scheduling in the power communication network, the problem of slow scheduling update speed in the dynamic network is solved, and the rapid and deterministic transmission of data flows in the power communication network is realized, and the scheduling ability is improved and bandwidth utilization.
Patent Information
- Application Number
- CN202310053488.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-02-03
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2043-02-03
AI Technical Summary
The prior art is difficult to quickly update time-sensitive network scheduling in power automation scenarios in dynamic networks, especially when equipment is dynamically added or removed, and there are offline methods to calculate the time exponentially with the increase in the number of streams, which cannot meet the power service. Real-time requirements for control information.
Adopting the pre-routed power communication network traffic update method, by abstracting the network topology into a directed graph, defining the characteristics of real-time data flow and best-effort data, optimizing traffic scheduling using a seven-tuple model and constraints, combining IEEE802.1Qbv Standard, determine the over-period internal gate control list, optimize the delay and bandwidth utilization.
It realizes the update and effective routing of the new flow gate control list in a very short time, ensures the deterministic transmission of data flow, improves the scheduling ability and bandwidth utilization of time-sensitive networks, and is suitable for dynamic update scenarios of power communication networks. .
Smart Images

Figure QLYQS_2 
Figure QLYQS_4 
Figure QLYQS_5
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of time-sensitive networks, and in particular to a method and device for updating flow in a power communication network based on pre-routing. Background Art
[0002] With the development of the industrial Internet and the emergence of smart grids, the contradiction between large traffic and real-time performance has arisen in actual power networks due to the continuous growth of data traffic and the increasingly stringent requirements of power grids on the real-time performance of business control information. At present, the Time Sensitive Network (TSN) Task Group is studying the extension of the Ethernet standard and has proposed a series of standards and specifications for link layer enhancement mechanisms and traffic strategies, including time synchronization, traffic scheduling, reliable transmission, and network management standards. TSN is considered by the industry to be the most important technology for future industrial communications. TSN is an OSI Layer 2 technology that provides deterministic communication on standard Ethernet. It ensures that messages can be transmitted over the network within a fixed and predictable amount of time. The switches and devices in the TSN network are time synchronized. Global scheduling controls when devices and switches generate and transmit data streams. Therefore, switches can reserve resources in advance to ensure deterministic transmission of data streams.
[0003] The time-controlled gating mechanism of the Time-Aware Shaper (TAS) is the core of its traffic scheduling. It places a gate behind each output port queue of each switch. Determined by the Gate Control List (GCL), the gate has two states, "open" or "closed". If the gate is open, the data packet is eligible for transmission, and if the gate is closed, transmission is prohibited.
[0004] However, at present, few works focus on the online scheduling update problem in dynamic networks, and existing methods cannot solve the multicast communication problem. Although different offline scheduling methods have been proposed to provide near-optimal solutions, they are not suitable for power automation scenarios that require dynamic addition or removal of devices and dynamic change of locations. In addition, the calculation time of some offline methods increases exponentially with the increase in the number of flows, which is unbearable for the real-time requirements of power business control information. Therefore, we pay more attention to the speed of problem solving and update scenarios. The pre-routing algorithm provides more optional paths for the scheduling stage based on load balancing to increase schedulability. At the same time, the optimized algorithm further speeds up the algorithm execution time, making it easier to quickly obtain the updated schedule. In the update stage, we dynamically arrange the schedule for the network without using a solver to determine whether new time-sensitive flows can be deployed in the current network state to accelerate traffic scheduling.
[0005] In order to understand the development status of the prior art, we searched, compared and analyzed the existing papers and patents, and screened out the following technical information that is highly relevant to the present invention:
[0006] Reference Solution 1: In "Heuristic List Scheduler for Time Triggered Traffic in Time Sensitive Networks", in order to solve the interdependence problem of routing and scheduling constraints, the authors introduced a heuristic list scheduler (HLS) that uses the combined route and scheduling constraints to generate an effective schedule in one step. Compared with the scheduler using a fixed path, this method significantly improves the ability to find a feasible schedule. In addition, HLS also supports multicast communication, distributed real-time applications, and inter-flow dependencies.
[0007] Reference Solution 2: In "Adaptive Group Routing and Scheduling in Multicast Time-Sensitive Networks", the authors proposed an adaptive group routing and scheduling (AGRS) method in multicast time-sensitive networks. Its framework includes three stages: pre-processing, scheduling synthesis, and post-processing. In the pre-processing stage, unnecessary links are first pruned, and qualified complete subgraphs are adaptively grouped to obtain a simplified topological structure. Then, based on the simplified topological structure, the scheduling synthesis stage jointly considers the routing and scheduling processes to improve schedulability. Finally, the post-processing stage processes the scheduling results of the previous stages to generate a feasible schedule.
[0008] In the paper "Adaptive Scheduling for Multicluster Time-Triggered Train Communication Networks", the authors proposed an algorithm based on mixed integer linear programming and counterexample guidance method to adaptively generate schedules based on changes at the system cluster level, targeting the scenario where all affected flows need to be rescheduled again when the traffic in the network changes. Adaptive scheduling can respond quickly to changes at runtime, and the algorithm enhances the incremental scheduler, allowing the scheduler to respond to changes at the flow level and cluster level.
[0009] The solution 1 in the literature improves the schedulability of the algorithm and also considers the application to large-scale networks and multicast situations, but lacks theoretical analysis of the key factors that cause TSN flow delay, and does not further analyze the bandwidth utilization of best-effort flows. The solution 2 in the literature jointly considers the routing and scheduling processes to improve schedulability and supports multicast communication environments, but the solution process method for optimizing the execution time in large-scale networks is relatively simple, only a rough pruning method is used, and the topology of the simulation experiment environment is simple, and it has not been verified in large-scale networks. The calculation time of the algorithm offline method increases exponentially with the increase in the number of flows. The solution 3 in the literature uses the isomorphic characteristics of multi-cluster systems to speed up the scheduling process, ensuring that the network adapts to the changes in the system cluster level online, while effectively reducing the offline scheduling time, but the scope of application is small and is only applicable to network environments with simple coupling processes within the train. Summary of the invention
[0010] The purpose of the present invention is to provide a method and device for updating flow in a power communication network based on pre-routing, so as to solve the problems raised in the above-mentioned background technology.
[0011] In order to solve the above technical problems, the present invention provides the following technical solutions:
[0012] A method for updating power communication network traffic based on pre-routing comprises the following steps:
[0013] S100. Abstractly model the network topology as a directed graph G(V,E), where V is a node set and E is an edge set:
[0014] Node set V = (S∪H), where S and H are the sets of switches and terminal hosts respectively;
[0015] E is a set of tuples representing links in the network. i ,v j )|v i ,v j ∈V,i≠j}, and at node V i and V j There are links between them, and each element in the set represents a unidirectional link;
[0016] S200, build TSN time-scale mesh diagram based on full-duplex Ethernet link, link node V i and V j The physical link between them corresponds to two directed links, namely [V i ,V j ] and [V j ,V i ];
[0017] S300, real-time data TT flow for system control and best-effort data for auxiliary services flow in the directed link;
[0018] S400, modeling the real-time data TT stream characteristics;
[0019] S500, define a unidirectional link [V i ,V j The link attribute on ] is a triple h ij Indicates link [V i ,V j ] super period, and h ij = lcm{fp:fsζ = [V i ,V j ]},d ij represents the propagation delay of the link, n ij Indicates the number of queues on the sending switch egress port;
[0020] S600. To improve the schedulability of all types of traffic in the TSN network, an overall cost function is defined.
[0021] According to the above technical solution, the real-time data TT flow and the best effort data of S300 are specifically as follows:
[0022] The real-time data TT stream has a fixed period and size and must reach all destinations within the specified delay and meet the jitter limit constraints;
[0023] Best-effort data does not require timing guarantees and has loose latency constraints.
[0024] According to the above technical solution, the specific modeling method of S400 is:
[0025] F TT The scheduled and configured flow F config And the newly generated business flow F new Each TT data stream f∈F TT We can use the seven-tuple f=<p,w,d,q,s,D,S> Definition: fp represents the cycle of the flow, fw represents the frame length, that is, the window size required to send the flow, fd represents the maximum tolerable end-to-end delay, fq represents the outbound port queue occupied by the flow, fs represents the source node of the flow, fD represents the set of receiving nodes of the flow, and fS represents the set of routing nodes it contains, including the scheduling schedule sent by each node on it.
[0026] According to the above technical solution, the specific method of S500 is:
[0027] Based on the IEEE802.1Qbv standard, the egress link of the terminal system ij=1, n on the TSN switch egress link ij =8, each switch has multiple input ports and output ports, which is responsible for forwarding the frames of the input port to the corresponding output port according to the destination address. The flow update scheduling mechanism needs to determine a gate control list within a super period, which can be executed cyclically to control the transmission of each TT flow.
[0028] According to the above technical solution, the real-time data TT stream needs to be constrained to ensure that the frames of the newly added or modified TT stream are transmitted in the network in a real, time-determined, sequential and non-overlapping manner. The updated constraint conditions are as follows:
[0029] (1) Frame Constraint
[0030] The basic frame sending time constraint, for flow f∈F new On each routing link s i Each instance on .ζ has an offset s i .o j ,1≤j≤s i .ι should satisfy the constraints:
[0031] s i .o j =s i .o j mod fp,if s i .o j >fp (1)
[0032] 0≤s i .o j ≤fp-fw (2)
[0033] Where: mod is the modulus, that is, the sending window must satisfy that if the sending offset of the instance is greater than the flow period, it should be modulo the flow period, and the sending offset should be greater than or equal to 0 and less than or equal to the period minus the transmission delay;
[0034] (2) Delay Constraint
[0035] For a flow f∈F new , for each instance on any receiving end d∈fD link, its offset s d .o j , 1≤j≤s d .ι, should reach the destination within the maximum tolerable end-to-end delay time, Identify the link at the receiving end of the stream, which should satisfy the following constraints:
[0036]
[0037] That is, for any flow and any sink d, the sending offset plus the transmission delay and propagation delay of any instance on the receiving link is Subtract the sending offset s of the instance on the sending link f.s .o j , this time interval is the end-to-end delay, which must be greater than zero and less than or equal to the maximum end-to-end delay tolerance time fd of the flow. The flow that meets the delay constraint is called a schedulable flow, otherwise it is unschedulable;
[0038] (3) Sending order constraints
[0039] Specifies the order that frames should follow when passing through the links on the routing path. For any flow f∈F new , any link s i .ζ=[v a ,v x ] and any other link s j .ζ=[v x ,v b ], where v a 、v b 、v x For different nodes, s i .ι identifies the number of instances of the routing node, which should satisfy the following constraints:
[0040]
[0041] 1≤j≤s.ι
[0042]
[0043] The start time of the instance transmission on the successor link must be greater than or equal to the completion time of the instance transmission on the predecessor link, that is, greater than the instance sending offset s on the predecessor link. i .o j , streaming time f i .w, link propagation delay and the sum of the maximum value of the clock deviation of the entire network, δ, where δ is used to ensure the timing of link transmission when there is an error in clock synchronization;
[0044] (4) No sending conflict constraints
[0045] The core constraint of transmission on the link is that there is no transmission time overlap conflict, that is, any two data flows passing through the same link should not have any transmission overlap in time. n ∈F new Any routing node f n .s i The xth instance of sends offset f n .s i .ox , with any flow f having the same routing path c ∈F config The yth instance Send offset:
[0046]
[0047] [f c .s i .o y +f c .p*(y-1)≤f n .s i .o x +f n .w+f n .p*(x-1)]
[0048] V[f c .s i .o y +f c .w+f c .p*(y-1)≥f n .s i .o x +f n .p*(x-1)] (5)
[0049] This constraint states that for instance f n .s i .o x , the time when the transmission of any configured scheduled instance starts is less than or equal to f n The instance transfer completion time or f n The instance transmission start time is less than or equal to the transmission completion time of any configured scheduled instance;
[0050] (5) Jitter Constraint
[0051] TT flow rate c ∈F config The delay jitter constraints should be met:
[0052] Indicates that the flow is on link f c .s i The ideal waiting time on ζ is the difference between the earliest transmittable time and the flow arrival time when there is no jitter, and is expressed as Indicates the waiting time of the link from the sending terminal to all receiving terminals, using Indicates the actual waiting time of the flow on the link, Indicates the total actual waiting time from the stream sending terminal to all receiving end links. Indicates the maximum delay jitter tolerance of the flow, with the following constraints:
[0053]
[0054] Among them, f c .s i .o n Identify the flow f c At routing node f c .s i Send offset, f c .s i+1 .o n Identify the flow f c At the next routing node f c .s i+1 The nth instance of sends the offset, is the link propagation delay. At the same time, the algorithm considers whether zero receiving jitter is configured, that is, whether the receiving terminal receives all frame instances of the stream without jitter at all times. If all are received when the offset is 0, it is called zero receiving jitter ZRJ. If it is configured as RJ, that is, receiving jitter is allowed, the receiver can receive frame instances at different time points. Otherwise, the receiver will receive the frame at a constant rate equal to the period.
[0055] (6) Frame Isolation Constraints
[0056] The frame isolation constraint is implemented in the algorithm. It does not allow multiple flows to arrive in the same queue at the same time to prevent the uncertain order of entering the queue from causing the uncertain transmission order of the egress link. That is, the constraint stipulates that only frames from one flow can be stored in one queue at the same time. Flow f n ∈F new Any routing node f n .s i The xth instance of sends offset f n .s i .o x , with any flow f having the same routing path c ∈F config The yth instance Send the offset. If a conflict occurs when arranging a new flow, its outbound port queue is arranged in the next queue:
[0057]
[0058] Constraint 1:
[0059] [f c .s i .o y +f c .p*(y-1)>f n .s i .o x +f n.p*(x-1)]
[0060] ∨[f c .s i+1 .o y′ +f c .p*(y′-1) <f n .s i+1 .o x′ +f n .p*(x′-1)] (10)
[0061] Constraint 2:
[0062] [f c .s i .o y +f c .p*(y-1) <f n .s i .o x +f n .p*(x-1)]
[0063] ∨[f c .s i+1 .o y′ +f c .p*(y′-1)>f n .s i+1 .o x′ +f n .p*(x′-1)] (11).
[0064] According to the above technical solution, the element s is defined as follows:
[0065] Any element s∈fS can be represented by a quaternion Definition, where s.ζ represents the node egress link, sh identifies the node's height in the multicast tree, s.ι represents the number of instances of the flow in each cycle, and the set Indicates the offset of each instance of the flow at the outbound port of this node, that is, flow f i From the transmission link f i .s1.ζ passes through the intermediate node and reaches the receiving link f i .s d .ζ,f i .s d ∈fD.
[0066] According to the above technical solution, the function definition method of S600 is as follows:
[0067] When a link in the network is full of allocated time slots, it becomes a bottleneck link, which greatly reduces the schedulability of the network. At this time, MaxAlloSlots is defined as the maximum number of allocable time slots in the outbound port of each TSN network link i. No network link is allowed to be occupied by more than the maximum number of allocated time slots.
[0068] Defining AlloSlots i The number of time slots allocated to link i satisfies the time slot allocation constraint:
[0069]
[0070] MaxAlloSlots is used to weight the weight of the link, flow f i The specific weighting formula of the routing link load weight is:
[0071]
[0072] LW fi The total number of weighted routing paths accounts for the total number of links in the network and the load of each link, where α is greater than zero and is used to adjust the number of time slots. i .S|Identify stream f i The number of routing nodes is nonlinearly weighted. If the load on a link is close to the maximum number of allocated time slots, the weight will become extremely large, which will make the link unavailable, reducing the probability of the flow using the path, and thus more inclined to choose links with light loads for routing and scheduling;
[0073] Therefore, our goal is to minimize MaxAlloSlots. In the update algorithm, this value can be increased only when all pre-routing paths cannot be scheduled. It is defined as follows:
[0074]
[0075] Compared with the prior art, the beneficial effects achieved by the present invention are:
[0076] The present invention studies the routing and scheduling update method of time-sensitive network data flow in the power communication business network, and proposes a pre-routing algorithm, a heuristic queue incremental update algorithm and a time slot scheduling algorithm based on the gate control scheduling method of the time-aware shaper. The new flow gate control list is solved in a very short time and updated to each switch output port, ensuring the deterministic transmission of data flow in the power communication network control update scenario. Quickly solve the effective port gate control list and effective routing to accommodate more time-critical traffic to improve bandwidth utilization and schedulability. DETAILED DESCRIPTION
[0077] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0078] The present invention provides a technical solution: a method for updating flow of a power communication network based on pre-routing, characterized in that it comprises the following steps:
[0079] S100. Abstractly model the network topology as a directed graph G(V,E), where V is a node set and E is an edge set:
[0080] Node set V = (S∪H), where S and H are the sets of switches and terminal hosts respectively;
[0081] E is a set of tuples representing links in the network. i ,v j )|v i ,v j ∈V,i≠j}, and at node V i and V j There are links between them, and each element in the set represents a unidirectional link;
[0082] S200, build TSN time-scale mesh diagram based on full-duplex Ethernet link, link node V i and V j The physical link between them corresponds to two directed links, namely [V i ,V j ] and [V j ,V i ];
[0083] S300, real-time data TT flow for system control and best-effort data for auxiliary services flow in the directed link;
[0084] S400, modeling the real-time data TT stream characteristics;
[0085] S500, define a unidirectional link [V i ,V j The link attribute on ] is a triple h ij Indicates link [V i ,V j ] super period, and h ij = lcm{fp:fsζ = [V i ,V j ]},d ijrepresents the propagation delay of the link, n ij Indicates the number of queues on the sending switch egress port;
[0086] S600. To improve the schedulability of all types of traffic in the TSN network, an overall cost function is defined.
[0087] Specifically, the real-time data TT flow and the best effort data of S300 are as follows:
[0088] The real-time data TT stream has a fixed period and size and must reach all destinations within the specified delay and meet the jitter limit constraints;
[0089] Best-effort data does not require timing guarantees and has loose latency constraints.
[0090] Specifically, the specific modeling method of S400 is:
[0091] F TT The scheduled and configured flow F config And the newly generated business flow F new Each TT data stream f∈F TT We can use the seven-tuple f=<p,w,d,q,s,D,S> Definition: fp represents the cycle of the flow, fw represents the frame length, that is, the window size required to send the flow, fd represents the maximum tolerable end-to-end delay, fq represents the outbound port queue occupied by the flow, fs represents the source node of the flow, fD represents the set of receiving nodes of the flow, and fS represents the set of routing nodes it contains, including the scheduling schedule sent by each node on it.
[0092] Specifically, the specific method of S500 is:
[0093] Based on the IEEE802.1Qbv standard, the egress link of the terminal system ij =1, n on the TSN switch egress link ij =8, each switch has multiple input ports and output ports, which is responsible for forwarding the frames of the input port to the corresponding output port according to the destination address. The flow update scheduling mechanism needs to determine a gate control list within a super period, which can be executed cyclically to control the transmission of each TT flow.
[0094] Specifically, the real-time data TT stream needs to be constrained to ensure that the frames of the newly added or modified TT stream are transmitted in the network in a real, deterministic, sequential and non-overlapping manner. The updated constraint conditions are as follows:
[0095] (1) Frame Constraint
[0096] The basic frame sending time constraint, for flow f∈Fnew On each routing link s i Each instance on .ζ has an offset s i .o j ,1≤j≤s i .ι should satisfy the constraints:
[0097] s i .o j =s i .o j mod fp,if s i .o j >fp (1)
[0098] 0≤s i .o j ≤fp-fw (2)
[0099] Where: mod is the modulus, that is, the sending window must satisfy that if the sending offset of the instance is greater than the flow period, it should be modulo the flow period, and the sending offset should be greater than or equal to 0 and less than or equal to the period minus the transmission delay;
[0100] (2) Delay Constraint
[0101] For a flow f∈F new , for each instance on any receiving end d∈fD link, its offset s d .o j , 1≤j≤s d .ι, should reach the destination within the maximum tolerable end-to-end delay time, Identify the link at the receiving end of the stream, which should satisfy the following constraints:
[0102]
[0103] That is, for any flow and any sink d, the sending offset plus the transmission delay and propagation delay of any instance on the receiving link is Subtract the sending offset s of the instance on the sending link f.s .o j , this time interval is the end-to-end delay, which must be greater than zero and less than or equal to the maximum end-to-end delay tolerance time fd of the flow. The flow that meets the delay constraint is called a schedulable flow, otherwise it is unschedulable;
[0104] (3) Sending order constraints
[0105] Specifies the order that frames should follow when passing through the links on the routing path. For any flow f∈F new , any link s i .ζ=[v a ,v x ] and any other link sj .ζ=[v x ,v b ], where v a 、v b 、v x For different nodes, s i .ι identifies the number of instances of the routing node, which should satisfy the following constraints:
[0106]
[0107] 1≤j≤s.ι
[0108]
[0109] The start time of the instance transmission on the successor link must be greater than or equal to the completion time of the instance transmission on the predecessor link, that is, greater than the instance sending offset s on the predecessor link. i .o j , streaming time f i .w, link propagation delay and the sum of the maximum value of the clock deviation of the entire network, δ, where δ is used to ensure the timing of link transmission when there is an error in clock synchronization;
[0110] (4) No sending conflict constraints
[0111] The core constraint of transmission on the link is that there is no transmission time overlap conflict, that is, any two data flows passing through the same link should not have any transmission overlap in time. n ∈F new Any routing node f n .s i The xth instance of sends offset f n .s i .o x , with any flow f having the same routing path c ∈F config The yth instance Send offset:
[0112]
[0113] [f c .s i .o y +f c .p*(y-1)≤f n .s i .o x +f n .w+f n .p*(x-1)]
[0114] ∨[f c.s i .o y +f c .w+f c .p*(y-1)≥f n .s i .o x +f n .p*(x-1)] (5)
[0115] This constraint states that for instance f n .s i .o x , the time when the transmission of any configured scheduled instance starts is less than or equal to f n The instance transfer completion time or f n The instance transmission start time is less than or equal to the transmission completion time of any configured scheduled instance;
[0116] (5) Jitter Constraint
[0117] TT flow rate c ∈F config The delay jitter constraints should be met:
[0118] Indicates that the flow is on link f c .s i The ideal waiting time on ζ is the difference between the earliest transmittable time and the flow arrival time when there is no jitter, and is expressed as Indicates the waiting time of the link from the sending terminal to all receiving terminals, using Indicates the actual waiting time of the flow on the link, Indicates the total actual waiting time from the stream sending terminal to all receiving end links. Indicates the maximum delay jitter tolerance of the flow, with the following constraints:
[0119]
[0120] Among them, f c .s i .o n Identify the flow f c At routing node f c .s i Send offset, f c .s i+1 .o n Identify the flow f c At the next routing node f c .s i+1 The nth instance of sends the offset, is the link propagation delay. At the same time, the algorithm considers whether zero receiving jitter is configured, that is, whether the receiving terminal receives all frame instances of the stream without jitter at all times. If all are received when the offset is 0, it is called zero receiving jitter ZRJ. If it is configured as RJ, that is, receiving jitter is allowed, the receiver can receive frame instances at different time points. Otherwise, the receiver will receive the frame at a constant rate equal to the period.
[0121] (6) Frame Isolation Constraints
[0122] The frame isolation constraint is implemented in the algorithm. It does not allow multiple flows to arrive in the same queue at the same time to prevent the uncertain order of entering the queue from causing the uncertain transmission order of the egress link. That is, the constraint stipulates that only frames from one flow can be stored in one queue at the same time. Flow f n ∈F new Any routing node f n .s i The xth instance of sends offset f n .s i .o x , with any flow f having the same routing path c ∈F config The yth instance Send the offset. If a conflict occurs when arranging a new flow, its outbound port queue is arranged in the next queue:
[0123]
[0124] Constraint 1:
[0125] [f c .s i .o y +f c .p*(y-1)>f n .s i .o x +f n .p*(x-1)]
[0126] ∨[f c .s i+1 .o y′ +f c .p*(y′-1) <f n .s i+1 .o x′ +f n .p*(x′-1)](10)
[0127] Constraint 2:
[0128] [f c .s i .o y +fc .p*(y-1) <f n .s i .o x +f n .p*(x-1)]
[0129] ∨[f c .s i+1 .o y′ +f c .p*(y′-1)>f n .s i+1 .o x′ +f n .p*(x′-1)](11).
[0130] Specifically, the element s is defined as follows:
[0131] Any element s∈fS can be represented by a quaternion Definition, where s.ζ represents the node egress link, sh identifies the node's height in the multicast tree, s.ι represents the number of instances of the flow in each cycle, and the set Indicates the offset of each instance of the flow at the outbound port of this node, that is, flow f i From the transmission link f i .s1.ζ passes through the intermediate node and reaches the receiving link f i .s d .ζ,f i .s d ∈fD.
[0132] Specifically, the function definition method of S600 is as follows:
[0133] When a link in the network is full of allocated time slots, it becomes a bottleneck link, which greatly reduces the schedulability of the network. At this time, MaxAlloSlots is defined as the maximum number of allocable time slots in the outbound port of each TSN network link i. No network link is allowed to be occupied by more than the maximum number of allocated time slots.
[0134] Defining AlloSlots i The number of time slots allocated to link i satisfies the time slot allocation constraint:
[0135]
[0136] MaxAlloSlots is used to weight the weight of the link, flow f i The specific weighting formula of the routing link load weight is:
[0137]
[0138] The total number of weighted routing paths accounts for the total number of links in the network and the load of each link, where α is greater than zero and is used to adjust the number of time slots. i .S|Identify stream f i The number of routing nodes is nonlinearly weighted. If the load on a link is close to the maximum number of allocated time slots, the weight will become extremely large, which will make the link unavailable, reducing the probability of the flow using the path, and thus more inclined to choose links with light loads for routing and scheduling;
[0139] Therefore, our goal is to minimize MaxAlloSlots. In the update algorithm, this value can be increased only when all pre-routing paths cannot be scheduled. It is defined as follows:
[0140]
[0141] In this application, the pseudo codes in Table 1, Table 2 and Table 3 are used to implement the scheduling of the algorithm:
[0142]
[0143] Table 1
[0144]
[0145] Table 2
[0146]
[0147] Table 3
[0148] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0149] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for updating power communication network traffic based on pre-routing, characterized in that: The following steps are involved: S100. Abstractly model the network topology as a directed graph G(V,E), where V is a node set and E is an edge set: Node set V = (S∪H), where S and H are the sets of switches and terminal hosts respectively; E is a set of tuples representing links in the network. i ,v j )|v i ,v j ∈V,i≠j}, and at node V i and V j There are links between them, and each element in the set represents a unidirectional link; S200, build TSN time-scale mesh diagram based on full-duplex Ethernet link, link node V i and V j The physical link between them corresponds to two directed links, namely [V i ,V j ] and [V j ,V i ]; S300, real-time data TT flow for system control and best-effort data for auxiliary services flow in the directed link; S400: Modeling is performed based on the real-time data TT stream characteristics. The specific modeling method is as follows: F TT The scheduled and configured flow F config And the newly generated business flow F new Each TT data stream f∈F TT We can use the seven-tuple f=<p,w,d,q,s,D,S> Definition: fp represents the period of the flow, fw represents the frame length, i.e., the window size required to send the flow, fd represents the maximum tolerable end-to-end delay, fq represents the outbound port queue occupied by the flow, fs represents the source node of the flow, fD represents the set of receiving nodes of the flow, and fS represents the set of routing nodes it contains, including the scheduling schedule sent by each node on it; The real-time data TT stream needs to be constrained to ensure that the frames of the newly added or modified TT stream are transmitted in the network in a real, deterministic, sequential and non-overlapping manner. The updated constraint conditions are as follows: (1) Frame Constraint The basic frame sending time constraint, for flow f∈F new On each routing link s i Each instance on .ζ has an offset s i .o j ,1≤j≤s i .ι should satisfy the constraints: s i .o j =s i .o j mod f.p,if s i .o j >f.p (1) 0≤s i .o j ≤fp-fw (2) Where: mod is the modulus, that is, the sending window must satisfy that if the sending offset of the instance is greater than the flow period, it should be modulo the flow period, and the sending offset should be greater than or equal to 0 and less than or equal to the period minus the transmission delay; (2) Delay Constraint For a flow f∈F new , for each instance on any receiving end d∈fD link, its offset s d .o j , 1≤j≤s d .ι, should reach the destination within the maximum tolerable end-to-end delay time, Identify the link at the receiving end of the stream, which should satisfy the following constraints: That is, for any flow and any sink d, the sending offset plus the transmission delay and propagation delay of any instance on the receiving link is Subtract the sending offset s of the instance on the sending link f.s .o j , this time interval is the end-to-end delay, which must be greater than zero and less than or equal to the maximum end-to-end delay tolerance time fd of the flow. The flow that meets the delay constraint is called a schedulable flow, otherwise it is unschedulable; (3) Sending order constraints Specifies the order that frames should follow when passing through the links on the routing path. For any flow f∈F new , any link s i .ζ=[v a ,v x ] and any other link s j .ζ=[v x ,v b ], where v a 、v b 、v x For different nodes, s i .ι identifies the number of instances of the routing node, which should satisfy the following constraints: 1≤j≤s.ι The start time of the instance transmission on the successor link must be greater than or equal to the completion time of the instance transmission on the predecessor link, that is, greater than the instance sending offset s on the predecessor link. i .o j , streaming time f i .w, link propagation delay and the sum of the maximum value of the clock deviation of the entire network, δ, where δ is used to ensure the timing of link transmission when there is an error in clock synchronization; (4) No sending conflict constraints The core constraint of transmission on the link is that there is no transmission time overlap conflict, that is, any two data flows passing through the same link should not have any transmission overlap in time. n ∈F new Any routing node f n .s i The xth instance of sends offset f n .s i .o x , with any flow f having the same routing path c ∈F config The yth instance Send offset: [f c .s i .o y +f c .p*(y-1)≤f n .s i .o x +f n .w+f n .p*(x-1)] V[f c .s i .o y +f c .w+f c .p*(y-1)≥f n .s i .o x +f n .p*(x-1)] (5) This constraint states that for instance f n .s i .o x , the time when the transmission of any configured scheduled instance starts is less than or equal to f n The instance transfer completion time or f n The instance transmission start time is less than or equal to the transmission completion time of any configured scheduled instance; (5) Jitter Constraint TT flow rate c ∈F config The delay jitter constraints should be met: Indicates that the flow is on link f c .s i The ideal waiting time on ζ is the difference between the earliest transmittable time and the flow arrival time when there is no jitter, and is expressed as Indicates the waiting time of the link from the sending terminal to all receiving terminals, using Indicates the actual waiting time of the flow on the link, Indicates the total actual waiting time from the stream sending terminal to all receiving end links. Indicates the maximum delay jitter tolerance of the flow, with the following constraints: Among them, f c .s i .o n Identify the flow f c At routing node f c .s i Send offset, f c .s i+1 .o n Identify the flow f c At the next routing node f c .s i+1 The nth instance of sends the offset, is the link propagation delay; at the same time, the algorithm considers whether zero receive jitter is configured, that is, whether the receiving terminal receives all frame instances of the stream without jitter at all times. If all are received when the offset is 0, it is called zero receive jitter ZRJ. If it is configured as RJ, that is, receive jitter is allowed, the receiver can receive frame instances at different time points. Otherwise, the receiver will receive the frame at a constant rate equal to the period. (6) Frame Isolation Constraints The frame isolation constraint is implemented in the algorithm. It does not allow multiple flows to arrive in the same queue at the same time to prevent the uncertain order of entering the queue from causing the uncertain transmission order of the egress link. That is, the constraint stipulates that only frames from one flow can be stored in one queue at the same time. Flow f n ∈F new Any routing node f n .s i The xth instance of sends offset f n .s i .o x , with any flow f having the same routing path c ∈F config The yth instance Send the offset. If a conflict occurs when arranging a new flow, its outbound port queue is arranged in the next queue: Constraint 1: [f c .s i .o y +f c .p*(y-1)>f n .s i .o x +f n .p*(x-1)] V[f c .s i+1 .o y′ +f c .p*(y ′ -1)<f n .s i+1 .o x′ +f n .p*(x′-1)](10) Constraint 2: [f c .s i .o y +f c .p*(y-1)<f n .s i .o x +f n .p*(x-1)] V[f c .s i+1 .o y′ +f c .p*(y ′ -1)>f n .s i+1 .o x′ +f n .p*(x′-1)](11); S500, define a unidirectional link [V i ,V j The link attribute on ] is a triple h ij Indicates link [V i ,V j ] super period, and h ij = lcm{fp:fsζ = [V i ,V j ]},d ij represents the propagation delay of the link, n ij Indicates the number of queues on the sending switch egress port; S600. In order to improve the schedulability of all types of traffic in the TSN network, an overall cost function is defined, and the definition method is as follows: When a link in the network is fully allocated with time slots, it becomes a bottleneck link, which greatly reduces the schedulability of the network. At this time, MaxAlloSlots is defined as the maximum number of allocatable time slots in the outbound port of each TSN network link i. No network link is allowed to be occupied by more than the maximum number of time slots allocated. Defining AlloSlots i The number of time slots allocated to link i satisfies the time slot allocation constraint: MaxAlloSlots is used to weight the weight of the link, flow f i The specific weighting formula of the routing link load weight is: The total number of weighted routing paths accounts for the total number of links in the network and the load of each link, where α is greater than zero and is used to adjust the number of time slots. i .S|Identify stream f i The number of routing nodes is nonlinearly weighted. If the load on a link is close to the maximum number of allocated time slots, the weight will become extremely large, which will make the link unavailable, reducing the probability of the flow using the path, and thus more inclined to choose links with light loads for routing and scheduling; Therefore, our goal is to minimize MaxAlloSlots. In the update algorithm, this value can be increased only when all pre-routing paths cannot be scheduled. It is defined as follows:
2. The method for updating the flow of a power communication network based on pre-routing according to claim 1, characterized in that: The real-time data TT flow and best effort data of S300 are as follows: The real-time data TT stream has a fixed period and size and must reach all destinations within the specified delay and meet the jitter limit constraints; Best-effort data does not require timing guarantees and has loose latency constraints.
3. The method for updating the flow of a power communication network based on pre-routing according to claim 1, characterized in that: The specific method of S500 is: Based on the IEEE802.1Qbv standard, the egress link of the terminal system ij =1, n on the TSN switch egress link ij =8, each switch has multiple input ports and output ports, which is responsible for forwarding the frames of the input port to the corresponding output port according to the destination address. The flow update scheduling mechanism needs to determine a gate control list within a super period, which can be executed cyclically to control the transmission of each TT flow.
4. The method for updating the flow of a power communication network based on pre-routing according to claim 1, characterized in that: The element s is defined as follows: Any element s∈fS can be represented by a quaternion Definition, where s.ζ represents the node egress link, sh identifies the node's height in the multicast tree, s.ι represents the number of instances of the flow in each cycle, and the set Indicates the offset of each instance of the flow at the outbound port of this node, that is, flow f i From the transmission link f i .s1.ζ passes through the intermediate node and reaches the receiving link f i .s d .ζ,f i .s d ∈fD.
