Heuristic traffic scheduling method based on delay tolerance

Abstract

The invention discloses a heuristic traffic scheduling method based on delay tolerance, which relates to the technical field of time-sensitive networks, and comprises the following steps of: S1, calculating a load balancing route, and relatively uniformly distributing traffic to all available links through a load balancing strategy so as to avoid network congestion and improve the overall bandwidth utilization rate; s2, delay tolerant scheduling is calculated, a scheduling strategy is set, and delay existing in an intermediate node is tolerated, so that a larger-scale flow can be scheduled under the same condition; and S3, calculating compression, and improving the utilization efficiency of network resources by reducing the number of guard bands and optimizing the use of network bandwidth. According to the method, the traffic schedulability is improved, and more time-sensitive flows can be scheduled through a delay tolerance mechanism under the same scheduling condition; meanwhile, the bandwidth utilization rate is improved, the system can distribute the network bandwidth more reasonably through the load balancing routing and compression algorithm, and waste of the bandwidth is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of time-sensitive networks, and particularly to a heuristic traffic scheduling method based on delay tolerance. Background Art

[0002] With the rapid development of in-vehicle network technology, the requirements for real-time performance and reliability of in-vehicle communication systems are increasing day by day. As the current advanced communication technology, Time-Sensitive Network (TSN) has been gradually introduced into in-vehicle communication systems due to its advantages in time determinacy, and TSN is gradually becoming the core infrastructure to support the transmission of critical service data. As an enhanced evolution of traditional Ethernet, TSN provides reliable technical support for in-vehicle communication by introducing a deterministic transmission mechanism. The core idea of TSN traffic scheduling is to allocate definite transmission time slots for various data flows in the network through pre-computation. This scheduling method can ensure that critical service data is transmitted within a strictly specified time window, thereby meeting the delay requirements of microseconds or even nanoseconds. For example, in an intelligent driving system, environmental perception data must be transmitted within a specified time; otherwise, it will affect the decision-making response speed of the vehicle.

[0003] From the perspective of technical implementation, traffic scheduling needs to solve the following problems: First is the generation algorithm of the scheduling table, how to achieve the optimal allocation of network resources on the premise of meeting the time limit constraints of all time-sensitive flows; second is the scalability of the scheduling strategy, when the network scale expands, how to maintain the scheduling efficiency from not decreasing or decreasing slowly. Traditional traffic scheduling algorithms face problems such as low bandwidth utilization, high latency, and poor scalability. In time-sensitive networks, how to ensure the priority transmission of time-sensitive traffic while making full use of network bandwidth, reducing latency, and improving the overall efficiency of the system has become an urgent problem to be solved. Existing traffic scheduling methods still have low scheduling efficiency and resource utilization when the traffic scale is large or the network topology is complex, and cannot meet the applications with high real-time and high bandwidth requirements. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a heuristic traffic scheduling method based on delay tolerance.

[0005] The purpose of the present invention is achieved by the following technical solutions: The present invention discloses a heuristic traffic scheduling method based on delay tolerance, including the following steps: S1. Calculate the load-balanced routing, and distribute the traffic relatively evenly to all available links through the load-balancing strategy; S2. Calculate the delay-tolerant scheduling, set the scheduling strategy, and tolerate the existence of delays in intermediate nodes; S3. Calculate compression, optimize the use of network bandwidth by reducing the number of guard bands, and improve the utilization efficiency of network resources.

[0006] Further, the computing load balancing routing described in step S1 includes sorting and heuristic path selection, specifically including the following sub-steps: S11. Based on the traffic busy degree function sort all the traffic to be scheduled from large to small. The formula of the traffic busy degree function is , where represents the payload of the flow, and represents the period of the flow; S12. Perform heuristic routing based on the sorted traffic in step S11. Determine the heuristic evaluation function according to the already generated delay and the estimated subsequent delay; by searching all adjacent nodes except the upstream node in the current node, select the node with the minimum overall delay value as the routing node, and search sequentially until the destination node: Calculate the conflict function between different flows through the formula , where represents the transmission path that the flow i has passed through, represents the number of path overlaps between the flow and the flow and the flow , represents the frame size of the flow , represents the period of the flow , represents the frame size of the flow , represents the period of the flow , represents the maximum value of conflicts that occur between any two flows with overlapping paths among all flows; Calculate the reserved delay through the formula , where represents the propagation delay of the flow ; When any flow reaches an intermediate node, assume that the subsequent transmission path reaches the destination node using the shortest path first, and generate an estimated delay . Calculate the heuristic delay prediction function through the formula ; Finally, calculate the total estimated delay of large-scale periodic time-sensitive traffic through ; Finally, calculate the total estimated delay of large-scale periodic time-sensitive traffic through ; .

[0007] Preferably, step S2 specifically includes the following steps: S21. Input parameters, input the description information of each frame, including the period, the first offset, and the transmission path. S22. Calculate the start transmission time of each node in the path within the first period of each frame, and start scheduling in sequence. During the scheduling process, use the conflict detection algorithm to judge whether the currently detected frame conflicts with the already scheduled frames frame by frame. If there is no conflict, continue with the scheduling of the next frame; if there is a conflict, execute step S23. S23. Scheduling in the first stage, call the conflict handling algorithm in the first stage, try to allocate a second offset for this frame so that it does not conflict with the already scheduled frames. If the scheduling is successful, that is, there is no conflict with other frames, add the second offset to the timestamps of this frame on all path nodes to complete the final time positioning; if the scheduling fails, execute step S24. S24. Scheduling in the second stage, for the frames that have not been successfully scheduled, call the conflict handling algorithm in the second stage, avoid the conflict time slots, calculate the third offset and the cache delay , and update the transmission time of the frame on each node.

[0008] Preferably, the conflict detection algorithm described in step S22 specifically includes the following steps: S221. Traverse the already scheduled frames , calculate the earliest possible transmission time interval of the currently detected frame within this period, where represents the first start time, represents the first end time; represents the first end time; S222. Make a conflict-free judgment. If or , where represents the second start time, represents the second end time, then the currently detected frame does not overlap with the already scheduled frame in this period, jump out of the current period loop, and continue to judge the next frame; S223. Judge the conflict situation. If and, then the start time of the currently detected frame is within the interval of the already scheduled frame , use the cache delay to resolve the conflict: , and return ; where represents whether there is a conflict, represents the already scheduled frame that conflicts with the currently detected frame, represents the value of the adjustable offset, Indicates adjustable delay; If and , then the current detection frame will block the scheduled frame , and through the fourth offset let the current detection frame avoid the scheduled frame : , and return , where indicates yes; If the above conditions are not met, then return , indicating failure, None indicating non - existence.

[0009] Preferably, the first - stage conflict handling algorithm described in step S23 specifically includes the following steps: S231. Traverse each switching node passed by the current detection frame , and through the conflict detection algorithm described in step S22, return the conflict judgment result and the fourth offset ; S232. Judge the conflict occurrence situation. If a conflict occurs, if the fourth offset exceeds the maximum allowable offset , return , scheduling fails; otherwise, use the fourth offset to modify the scheduling time of the current detection frame , and restart the whole - path detection; if no conflict occurs, the fourth offset has met the scheduling requirements of all nodes, return the fourth offset .

[0010] Preferably, the second - stage conflict handling algorithm described in step S24 specifically includes the following steps: S241. Traverse each switching node on the frame path within the period, and obtain the scheduling time of this node in the current period , where indicates the third start time, indicates the third end time, and detect whether there is a conflict with the scheduled frame through the conflict detection algorithm described in step S22, and return ; If there is no conflict, continue to detect the next node; if a conflict occurs, preferably use the cache delay to solve it through step S243; S242. If the returned adjustable delay And if the cumulative cache delay does not exceed the maximum allowable time delay of the detection frame, then add the adjustable delay to the transmission times of all nodes starting from the current node in the path ; If step S242 cannot solve the problem, then use the cache delay through step S243 to jointly solve it with the time offset; S243. If the conflict cannot be solved only by the cache delay , then use the offset to adjust the overall start time of the frame; judge , where represents the maximum offset allowed for the frame. If the judgment is yes, then exit the program with failure and return ; Otherwise, adjust to , and reset the path scheduling time; exit the current cycle detection, return to the outermost layer to re-detect the entire path. If no conflict occurs in the current detection, it means the scheduling is successful, and return the adjusted scheduling time and the final offset .

[0011] Preferably, step S3 specifically includes the following steps: S31. Traverse all operations for each flow and calculate the delayable time of this operation; S32. Try to delay the operation, traverse each flow again, traverse each forwarding operation therein, and analyze whether there is an effective compression opportunity within the delayable range: if it is judged that there is a compression opportunity, then delay this operation for a period of time; when no operation is compressed after a full round of traversal, the algorithm terminates.

[0012] The beneficial effects of the present invention are: 1) The present invention improves the traffic schedulability. Under the same scheduling conditions, through the delay tolerance mechanism, more time-sensitive flows can be scheduled.

[0013] 2) The present invention improves the bandwidth utilization rate. Through the load balancing routing and compression algorithms, the system can more reasonably allocate the network bandwidth and avoid the waste of bandwidth. Description of the Drawings

[0014] Figure 1 is a step schematic diagram of a delay-tolerant heuristic traffic scheduling method according to an embodiment of the present invention; Figure 2 is a principle schematic diagram of a delay-tolerant heuristic traffic scheduling method according to an embodiment of the present invention; Figure 3 is a relationship topology diagram between nodes of a delay-tolerant heuristic traffic scheduling method according to an embodiment of the present invention. Detailed Embodiment

[0015] The technical solution of the present invention will be clearly and completely described below in conjunction with embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0016] The present invention proposes a traffic scheduling method based on a heuristic algorithm. By introducing load balancing routing, delay tolerance mechanism and compression algorithm, the use of network resources is optimized, and the schedulability of traffic and bandwidth utilization rate are significantly improved. The schematic diagram of its principle is as Figure 2 shown. Among them, sorting is a pre-step of the entire algorithm process, so it is classified in computing routing. The traffic scheduling algorithm of the present invention has efficient load balancing, high schedulability and bandwidth utilization rate, and can meet the requirements of time-sensitive applications. The schematic diagram of its steps is as Figure 1 shown, and specifically includes the following steps: S1. Calculate the load balancing routing. Through the load balancing strategy, the traffic is relatively evenly distributed to all available links; it is used to avoid network congestion and improve the overall bandwidth utilization rate; S2. Calculate the delay tolerance scheduling. Set the scheduling strategy to tolerate delays in intermediate nodes; enable a larger scale of traffic to be scheduled under the same conditions; S3. Calculate compression. By reducing the number of guard bands, the use of network bandwidth is optimized, and the utilization efficiency of network resources is improved.

[0017] Exemplarily, the calculation of the load balancing routing described in step S1 includes sorting and heuristic path selection, and specifically includes the following sub-steps: S11. Based on the traffic busyness function sort all the traffic to be scheduled from large to small. Specifically, the busier the traffic is in a single cycle, the more forward its sorting position is. The traffic busyness function is defined as the ratio of the payload of the flow to the period; its formula is , where represents the payload of the flow, represents the period of the flow; S12. Perform heuristic routing based on the sorted traffic in step S11. Determine the heuristic evaluation function with the already generated latency and the estimated upcoming latency as specific metrics. By searching all adjacent nodes of the current node except the upstream node, select the node with the minimum overall latency value as the routing node, and search sequentially until the destination node. The specific algorithm flow is as follows, where the node set V represents the switches and terminal nodes in the network. Switches are responsible for data forwarding and scheduling, while terminal nodes act as the source or destination of data. The edge set E represents the physical connections between nodes, and each directed edge represents a communication direction. Therefore, a pair of reverse edges, such as and , jointly represent the full-duplex physical connection between node and node b. The specific algorithm flow is as follows: Input: Source node s , destination node t , graph G=(V,E) Output: The optimal path from s to t 1) Initialize parameters: open list open←{s} , closed list close←∅ ; 2) while open≠∅ do 3) Select the node with the minimum cost from open u ; 4) Remove u from open , and add it to close ; 5) if u=t then 6) Return the path; 7) end 8) Traverse all adjacent nodes of u v∈Neighbors(u) {if v∉close then 9) Add v to open ; 10) end 11)end 12)return Path does not exist The key point is the specific calculation method of the node latency cost. Let the total estimated latency of the heuristic routing algorithm be , and its expression is ; where, ​​​Denote the delay function that has been generated currently, i.e., the delay D generated from the source node to the current node n; Denote the heuristic delay prediction function, which represents the delay that is expected to be generated from the current node n to the destination node. The specific calculation method of the predicted delay is as follows: the delay generated by using the shortest path first routing from the current node n to the destination node. The routing calculation should, as much as possible, disperse the time-sensitive traffic to different time slots on different paths for transmission to avoid conflicts between flows. According to the transmission characteristics of the time-sensitive traffic and the similarity between paths, a conflict function between different flows is defined, and through the formula Calculate the conflict function between different flows , where Denote the transmission path that flow i has passed through, Denote flow and flow The number of overlapping paths between them, Denote flow The frame size of, Denote flow The period of, Denote flow The frame size of, Denote flow The period of, Denote the maximum value of conflicts that occur between any two flows with overlapping paths among all flows; The final heuristic function evaluation value of the routing algorithm is the delay. Therefore, delay reservation is performed according to the possible size of conflicts. In order to make the reserved delay closer to the actual value, define the reserved delay As the product of the maximum conflict degree between any two flows and their propagation delay, the expression of the reserved delay is ; where, Denote flow The propagation delay of.

[0018] When any flow reaches an intermediate node, assuming that the subsequent transmission path reaches the destination node using the shortest path first, a predicted delay is generated. Through the formula Calculate the heuristic delay prediction function ; Finally, through the formula Calculate the total predicted delay of the large-scale periodic time-sensitive traffic .

[0019] Exemplarily, in step S2, during the scheduling process, a delay tolerance mechanism is introduced, allowing a certain delay to occur at the intermediate node without affecting the transmission timeliness of the entire flow. The proposed traffic scheduling algorithm is called a two-stage conflict handling algorithm, which specifically includes the following steps: S21. Input parameters, input the description information of each frame, including the period, the first offset, and the transmission path. S22. Calculate the start transmission time of each node in the path within the first period of each frame, and start scheduling in order. During the scheduling process, use the conflict detection algorithm to determine frame by frame whether the currently detected frame conflicts with the scheduled frames. If there is no conflict, continue to schedule the next frame; if there is a conflict, execute step S23. S23. Scheduling in the first stage, call the conflict handling algorithm in the first stage, try to allocate a second offset for this frame so that it does not conflict with the scheduled frames. If the scheduling is successful, that is, there is no conflict with other frames, add the second offset to the timestamps of this frame on all path nodes to complete the final time positioning; if the scheduling fails, execute step S24. S24. Scheduling in the second stage, for the frames that have not been successfully scheduled, call the conflict handling algorithm in the second stage, avoid the conflict time slots, calculate the third offset and the buffer delay , and update the transmission time of the frame on each node.

[0020] The overall process of the algorithm is as follows: Input: Information of each frame: (ignoring the switch processing time), period T , initial offset Output: Transmission time of the frame in each switch node 1) Initialization: 2) Each frame f:f←{} 3) Set of frames in the switch node s : F s ←∅ 4) Set of scheduled frames: F deployed ←∅ 5) Scheduling cycle time: T cycle 6) For each frame f do 7) Calculate the path, and calculate the initial transmission time of each switch node in order according to offset t start ,t end (only in the first period) 8) if f Pass through the switch node​s then 9) Add f to F s 10) end 11)end 12) / / First stage: Scheduling without buffer delay 13) For each switch node s∈S sorted do 14) For each frame f∈F s do 15) Execute the first-stage conflict handling algorithm to obtain the adjusted offset1 16) If deployable then 17) Add f to F deployed 18) Update all time parameters of f : t start ←t start +offset 1 ,t end ←t end +offset 1 19) end 20) end 21)End 22) / / Second stage: Adding frames with buffer delay 23) For each frame f∈F s and f∉F deployed do 24) Execute the second-stage conflict handling algorithm to obtain offset2 and buffer_delay 25)end Exemplarily, the conflict detection algorithm described in step S22 specifically includes the following steps: S221. Traverse the scheduled frames , and calculate the earliest possible transmission time interval of the current detected frame in this cycle, where represents the first start time, represents the first end time; S222. Perform non - conflict judgment. If or , where represents the second start time, represents the second end time, then the current detected frame and the scheduled frame do not overlap in this cycle. Jump out of the current cycle loop and continue to judge the next frame; S223. Conflict situation judgment. If and , then the start time of the current detected frame is within the interval of the scheduled frame . Use cache delay to resolve the conflict: , and return ; where represents whether a conflict occurs, represents the scheduled frame that conflicts with the current detected frame, represents the value of the adjustable offset, represents the adjustable delay; If and , then the current detected frame will block the scheduled frame . Let the current detected frame avoid the scheduled frame through the fourth offset : , and return , where represents yes; If the above conditions are not met, then return , represents failure, None represents non - existence. The specific process of the conflict detection algorithm is as follows: Input: Deployed frame set Deployed, current frame f curr time window of t start ,t end , cycle T Output: Conflict flag is_conflict, conflict frame f conflict , offset to be adjusted offset, cache delay to be increased delay 1) For each frame f∈Deployed and f∈F s do 2) For k←0to ⌊ ⌋ do 3) if t start (f curr )+kT≥t end (f) or t end (f curr )+kT≤t start (f) then 4) continue without time overlap, skip detection 5) end 6) else 7) if t start (f curr )+kT∈[t start (f),t end (f)] then 8) delay← t end (f)−(t start (f curr )+kT) ; 9) return True,f,0,delay Need cache delay 10) end 11) else if t start (f)∈[t start (f curr )+kT,t end (f curr )+kT] then 12) offset’← t end (f curr )+kT−t start (f) ; 13) return True, f , offset’, 0 Need to adjust the offset 14) end 15) end 16) end 17) end 18) return False, None, None, None Conflict-free Exemplarily, the first-stage conflict handling algorithm described in step S23 specifically includes the following steps: S231. Traverse each switching node passed by the current detection frame and return the conflict judgment result and the fourth offset through the conflict detection algorithm described in step S22 ; S232. Judge the occurrence of conflicts. If a conflict occurs and the fourth offset exceeds the maximum allowable offset , return , scheduling fails; otherwise, use the fourth offset to modify the scheduling time of the current detection frame and restart the detection of the entire path; if no conflict occurs, the fourth offset has met the scheduling requirements of all nodes, return the fourth offset .

[0021] The specific process of the first-stage conflict handling algorithm is as follows: Input: Set of frames F , Set of switch nodes S , Initial time parameter Output: Adjusted offset offset’ (if successful) or false (if the conflict is unsolvable) 1) conflict_free ← false; 2) while true do 3) conflict_free ← true; 4) for each switch node s∈S do 5) if a conflict is detected then 6) conflict_free ← false; 7) break out of the current loop and re-detect 8) end 9) Record s 's transmission time t start ,t end ; 10) while there are periodic overlaps do 11) Update offset’ and adjust it to the next cycle 12) Calculate the new time parameter: t start ←t start +T,t end ←t end +T ; 13) if the conflict still exists then 14) if offset’>benchmark then 15) return false Conflict exceeds the threshold and cannot be resolved 16) end 17) break Re - detect 18) end 19) end 20) end 21) if conflict_free then 22) return offset’ Return the offset without conflict 23) end 24)end Exemplarily, the second - stage conflict handling algorithm described in step S24 specifically includes the following steps: S241. Traverse each switching node (output port) on the frame path within the cycle, and obtain the scheduling time of this node in the current cycle , where represents the third start time, represents the third end time, and detect whether there is a conflict with the scheduled frame through the conflict detection algorithm described in step S22, and return ; if there is no conflict, continue to detect the next node; if a conflict occurs, preferentially use the buffer delay through step S243 to solve it; S242. If the adjustable delay returned and the cumulative buffer delay does not exceed the maximum allowable time delay of the detected frame, then add the adjustable delay to the transmission time of all nodes starting from the current node in the path ; if step S242 cannot solve it, then use the buffer delay and time offset to jointly solve it; S243. If the conflict cannot be solved only by the buffer delay , then use the offset to adjust the overall start time of the frame; judge , where Indicates the maximum offset allowed for a frame. If the judgment is yes, exit with failure and return ; Otherwise, adjust to , and reset the path scheduling time; Exit the current cycle detection, return to the outermost layer to re-detect the entire path. If no conflict occurs in the current detection, it means the scheduling is successful, and return the adjusted scheduling time and the final offset .

[0022] The specific process of the second-stage conflict handling algorithm is as follows: Input: Deployed frame set, initial time parameters of the current frame f , switch node set S , cycle T , conflict threshold benchmark Output: adjusted offset, cache delay of each switch node 1) while True do 2) for each switch node s∈S do 3) Record s 's initial transmission time t start ,t end ; 4) delay←0; 5) while True do 6) if conflict is detected then 7) local_adjusted←False; 8) Reset the time parameters to restore the initial settings 9) end 10) else 11) Update to the next cycle: t start ←t start +T,t end ← tend +T; 12) end 13) is_conflict, f c, offset’, delay ← ConflictDetection(Deployed, f , s ); 14) if is_conflict and not global_adjusted then 15) if offset’>benchmark then 16) return false Beyond the threshold, cannot be resolved 17) end 18) Adjust f The time parameter only affects the current and subsequent nodes 19) local_adjusted ← True; 20) end 21) else if is_conflict and global_adjusted then 22) if 23) delay>benchmark then 24) return false Beyond the threshold, cannot be resolved 25) end 26) Adjust f The global time parameter affects all nodes global_adjusted ← True; 27) break Break out of the current loop 28) end 29) end 30) if not local_adjusted then 31) global_adjusted ← False; 32) break Require global adjustment 33) end 34) end 35) if not(global_adjusted or local_adjusted) then 36) return offset’, delay Conflict successfully resolved 37) end 38)end Exemplarily, step S3 specifically includes the following steps: S31. Traverse all operations (such as the forwarding action of a certain node) for each flow, and calculate the delayable time of this operation; that is, without affecting subsequent scheduling, how much time can the current operation be delayed at most; S32. Try to delay the operation. Traverse each flow again, traverse each forwarding operation in it, and analyze whether there is an effective compression opportunity within the delayable range: If it is determined that there is a compression opportunity, delay this operation for a period of time; when no operation is compressed after a full round of traversal, the algorithm terminates.

[0023] The specific process of compression calculation is as follows: Input: Scheduled time-sensitive flow set Output: Optimized scheduling result 1) Initialize the compression flag compress←True 2) while compress do 3) Reset the compression flag compress←False 4) For each data flow flow do 5) For each node node do 6) Calculate the time slack computeSlack(sched,flow,oper) 7) end 8) end 9) For each data flow flow do 10) For each node node do 11) Calculate the delayable time computeDelays(slack) 12) If delay is allowed then 13) Apply the delay adjustment applyDelays(flow,oper,sched,delay) 14) Set the continue compression flag compress←True 15) end 16) end 17) end 18)end 19) Return the final scheduling result return sched Exemplarily, in a vehicle network environment, the traffic scheduling method of the present invention is applied to data transmission between multiple vehicle communication nodes. The relationship topology diagram between each node is as Figure 3As shown, it includes 16 terminal nodes and 8 switch nodes. In the figure, Es0 - Es15 represent 16 terminal nodes, and Sw0 - Sw7 (Switch0 - Switch7) represent 8 switch nodes. The terminal nodes contain various types of traffic flows. The traffic flow parameters are shown in Table 1.

[0024] Table 1: Traffic Flow Parameters

[0025] It is set that the maximum available bandwidth of the link is 100 Mbps. The entire scheduling process can be divided into four steps: sorting, routing, scheduling, and compression. According to the periods and load sizes of the above parameters, the busyness of each type of traffic can be obtained, and the traffic sequences are sorted from small to large according to the busyness function to obtain the sorted traffic sequence. For example: [Command and Control 1, Command and Control 3, Command and Control 2, Command and Control 5, Command and Control 4, Command and Control 6, Command and Control 7, Command and Control 8, Command and Control 9, Command and Control 10]. Heuristic routing calculation is performed on the sorted traffic sequence to obtain the paths of all types of traffic. The entire path consists of a node sequence. For example, after calculation, the routing of the Command and Control 5 traffic flow is [ES3, SW1, SW2, SW3, ES6]. Based on the above routing calculation, the forwarding paths of all flows are determined. Now, according to the order arranged in 1, the traffic is scheduled in order, and the transmission time windows of each flow passing through each node of the path are calculated. For example, the time windows of the Command and Control 5 type traffic at each node of the path are [[533, 540], [555, 562], [577, 584]]. Based on the above time windows, the time transmission windows of each node port can be obtained, and the guard band quantity is compressed in units of a single port. For example, the time window calculated at a certain port of the intermediate node SW1 is [[233, 238], [267, 276], [332, 339]]. At this port, the traffic belonging to [233, 238] can be offset as far back as possible within the range of [239, 260] without exceeding the maximum allowable delay of the traffic.

[0026] The above are only the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein, should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications, and environments, and can be changed within the scope of the concept described herein through the above teachings or the techniques or knowledge in related fields. And the changes and alterations made by those skilled in the art that do not depart from the spirit and scope of the present invention should all be within the protection scope of the appended claims of the present invention.

Claims

1. A heuristic traffic scheduling method based on delay tolerance, characterized in that It includes the following steps: S1. Calculate the load balancing route, and distribute the traffic relatively evenly to all available links through the load balancing policy; S2. Calculate the delay-tolerant scheduling, set the scheduling policy to tolerate delays in intermediate nodes; S3. Calculate compression, optimize the use of network bandwidth by reducing the number of guard bands, and improve the utilization efficiency of network resources.

2. The heuristic traffic scheduling method based on delay tolerance according to claim 1, characterized in that: The calculation of the load balancing route described in step S1 includes sorting and heuristic path selection, specifically including the following sub-steps: S11. Based on the traffic busy degree function Sort all the traffic to be scheduled from large to small. The formula of the traffic busy degree function is , where represents the payload of the flow, represents the period of the flow; S12. Perform heuristic routing based on the sorted traffic in step S11, determine the heuristic evaluation function according to the generated delay and the estimated subsequent delay; by searching all neighbor nodes except the upstream node in the current node, select the node with the minimum overall delay value as the routing node, and search sequentially until the destination node: Calculate the conflict function between different flows through the formula where represents the transmission path that flow i has passed through, represents the number of path overlaps between flow and flow and flow ; represents the frame size of flow ; represents the period of flow ; represents the frame size of flow ; represents the period of flow ; represents the maximum value of conflicts that occur between any two flows with overlapping paths among all flows; Calculate the reserved time delay through the formula where the propagation delay of the flow is calculated by is represented by ; When any stream reaches the intermediate node, assuming that the subsequent transmission path uses the shortest path first to reach the destination node, an estimated delay is generated , through the formula calculate the heuristic delay estimation function ; Finally, by calculating the total estimated delay of large-scale periodic time-sensitive traffic .

3. A heuristic traffic scheduling method based on delay tolerance according to claim 2, characterized in that, Step S2 specifically includes the following steps: S21. Input parameters, input the description information of each frame, including the period, the first offset and the transmission path; S22. Calculate the start transmission time of each node in the path of each frame in the first period, and start scheduling in sequence. During the scheduling process, use the conflict detection algorithm to judge frame by frame whether the currently detected frame conflicts with the scheduled frames. If there is no conflict, continue the scheduling of the next frame; if there is a conflict, execute step S23; S23. The scheduling in the first stage, call the conflict handling algorithm in the first stage, try to allocate a second offset for this frame to make it non-conflicting with the scheduled frames. If the scheduling is successful, that is, there is no conflict with other frames, add the second offset to the timestamps of this frame on all path nodes to complete the final time positioning; if the scheduling fails, execute step S24; S24. Scheduling in the second stage. For frames that have not been successfully scheduled, call the conflict handling algorithm in the second stage to avoid conflict time slots, calculate the third offset offset2 and the buffer delay , and update the transmission time of the frame on each node.

4. A heuristic traffic scheduling method based on delay tolerance according to claim 3, characterized in that The conflict detection algorithm described in step S22 specifically includes the following steps: S221. Traverse the scheduled frames , and calculate the earliest possible transmission time interval of the current detection frame within this cycle , where represents the first start time and represents the first end time; S222. Perform a conflict-free judgment. If or , where represents the second start time, represents the second end time, then the current detected frame and the scheduled frame do not overlap in this cycle, jump out of the current cycle loop, and continue to judge the next frame; S223. Conflict situation judgment. If and , then the start time of the current detection frame is within the interval of the scheduled frame . Use buffer delay to resolve the conflict: , and return ; where indicates whether a conflict occurs, indicates the scheduled frame that conflicts with the current detection frame, indicates the value of the adjustable offset, indicates the adjustable delay; If and , then the current detection frame will block the scheduled frame , and through the fourth offset make the current detection frame avoid the scheduled frame : , and return , where means yes; If the above conditions are not met, then return , indicating failure, None indicating non - existence.

5. A heuristic traffic scheduling method based on delay tolerance according to claim 4, characterized in that The conflict handling algorithm in the first stage described in step S23 specifically includes the following steps: S231. Traverse the current detection frame For each switching node passed through, return the conflict judgment result and the fourth offset through the conflict detection algorithm described in step S22 ; S232. Determine the conflict occurrence situation. If a conflict occurs and the fourth offset exceeds the maximum allowable offset , return , and the scheduling fails; otherwise, use the fourth offset to modify the scheduling time of the current detection frame and restart the entire path detection; if no conflict occurs, the fourth offset has met the scheduling requirements of all nodes, return the fourth offset .

6. A heuristic traffic scheduling method based on delay tolerance according to claim 5, characterized in that The conflict handling algorithm in the second stage described in step S24 specifically includes the following steps: S241. Traverse each switching node on the frame path within the traversal period, and obtain the scheduling time of this node in the current period , where represents the third start time, represents the third end time, and detect whether there is a conflict with the scheduled frame through the conflict detection algorithm described in step S22, and return ; if there is no conflict, continue to detect the next node; if a conflict occurs, preferentially use cache delay through step S243 to solve; S242. If the adjustable delay returned and the cumulative buffer delay does not exceed the maximum allowable time delay of the detection frame, add the adjustable delay to the transmission times of all nodes starting from the current node in the path ; if step S242 cannot solve the problem, use the buffer delay through step S243 and jointly solve it with the time offset; S243. If only the cache latency cannot resolve the conflict, use the offset to adjust the overall start time of the frame; judge , where represents the maximum offset allowed for the frame. If the judgment is yes, exit with failure and return ; otherwise, adjust to , and reset the path scheduling time; exit the current cycle detection, return to the outermost layer to re-detect the entire path. If no conflict occurs during the current detection, it means the scheduling is successful, and return the adjusted scheduling time and the final offset .

7. A heuristic traffic scheduling method based on delay tolerance according to claim 6, characterized in that Step S3 specifically includes the following steps: S31. Traverse all operations for each flow and calculate the delayable time of this operation; S32. Try to delay the operation, traverse each flow again, traverse each forwarding operation therein, and analyze whether there is an effective compression opportunity within the delayable range: if it is judged that there is a compression opportunity, delay this operation for a period of time; when no operation is compressed after a full round of traversal, the algorithm terminates.

Images