An Online Reconfiguration Method for Time-Sensitive Networks

Abstract

This invention discloses an online reconfiguration method for time-sensitive networks, comprising the following steps: S1: Abstracting the network topology to form a directed graph; S2: Determining the update time of each link based on the given downtime and the old traffic transmission scheme, using this as the new baseline time for traffic inheritance; S3: Incremental scheduling, establishing deterministic constraints based on the already scheduled transmission scheme and current traffic attributes, and calculating the start transmission time using network calculus; S4: Waiting for all traffic scheduling to complete, if not, returning to S3; S5: Generating a Global Channel Log (GCL) based on the new scheduling scheme, and generating a configuration file through configuration software and distributing it to switches and terminals for activation. This invention can guarantee network consistency and low packet loss deterministic requirements during reconfiguration. The earliest update time selection ensures extremely low response time for this mechanism, thereby guaranteeing low latency deterministic requirements for traffic transmission; it can also guarantee low packet loss deterministic transmission requirements.

Description

Technical Field

[0001] This invention relates to the field of network data processing technology, and in particular to an online reconfiguration method for time-sensitive networks. Background Technology

[0002] With the rapid development of the Industrial Internet of Things (IIoT), industrial production lines are gradually transforming from mass production to flexible production lines with customized manufacturing. In flexible production lines, the frequent information interaction between different production equipment and the coordination requirements of operational processes place deterministic transmission requirements on the field communication network, including low latency, low jitter, and low packet loss rate. Traditional industrial Ethernet, due to its "best-effort" service characteristics, cannot meet these needs. To address this, the IEEE 802.1TSN working group proposed Time-Sensitive Networking (TSN) technology and developed a series of sub-standards as the basis for deterministic traffic transmission. Among them, IEEE 802.1Qbv defines a programmable gating mechanism—the Time-Aware Shaper (TAS)—which uses traffic gates and gate control lists (GCLs) to control the transmission of traffic in the switch port buffer queues; IEEE 802.1Qcc provides a configuration model for the deployment of TSN on-end devices and switches; and IEEE 802.1AS-Rev provides the basis for clock synchronization for device deployment and traffic forwarding on switches. Meanwhile, the constantly evolving production demands in flexible production lines, leading to network changes such as equipment additions and removals, topology shifts, and task traffic variations, place higher demands on the real-time reconfiguration capabilities of TSN networks. TSN network reconfiguration is the process of switching network devices from an old transmission scheme to a new one: at the moment the network receives a new traffic demand and shuts down, all end devices stop generating old traffic; the old traffic already sent (legacy traffic) continues to be transmitted according to the old scheme; the transmission time at each switch node is calculated based on the new traffic attributes, and a corresponding new GCL configuration scheme is obtained; the time when the overall scheme solution is completed is the new scheme generation time; finally, the GCL scheme is deployed to the devices, and after the update time, each node generates and transmits new traffic according to the new configuration. Therefore, achieving real-time reconfiguration of TSN networks while ensuring deterministic traffic transmission is a highly challenging problem.

[0003] For the TSN network reconfiguration problem, the main methods currently used are the two-stage reconfiguration method, the sequential reconfiguration method, and the conflict-free constraint solution method. The first two focus on designing the network reconfiguration mechanism and calculating the update time, while the conflict-free constraint solution method focuses on designing the algorithm for solving the network transmission scheme after reconfiguration. Specifically, the two-phase reconfiguration method seeks the update moment when the network globally transitions from the old state to the new state, but ignores the problem of conflicts between old and new traffic transmission schemes caused by residual traffic at the update moment, thus disrupting the consistency of network reconfiguration. To solve the consistency problem, the sequential update method sets the network update moment when all old traffic transmissions are completed, thus avoiding conflicts between residual traffic and new traffic. However, this introduces a long waiting time for the injection of new traffic, disrupting the real-time performance of reconfiguration. The conflict-free constraint solution method, considering the coexistence of old and new traffic in the network, models the conflict-free problem between the unknown new transmission scheme and the known old transmission scheme as an integer linear programming (ILP) problem, and uses an ILP solver to obtain the new transmission scheme. However, the extremely high time complexity of this solution algorithm causes a delay in the deployment of the new scheme, resulting in a large number of frame drops during the reconfiguration period, making it difficult to guarantee the deterministic requirements of traffic.

[0004] In the prior art, the patent application number is CN202110750997.2, entitled "A Token Update System and Method for Traffic Shaper in Time-Sensitive Networks". The specific approach is as follows: a token bucket update system is configured for the traffic shaper of the switch port buffer queue. When the buffer queue is empty, the number of tokens in the corresponding queue token bucket is configured and the queue is added to the empty queue pool until a new packet arrives in the queue and the token bucket is updated. When the buffer queue is not empty, the system updates the number of tokens in the relevant buffer queue token bucket when the packet stored in the buffer queue is scheduled, until the scheduling is completed. However, this method only involves updating the receiving port when traffic transmission packets arrive at the switch. It does not consider the network reconfiguration time node and the design of the traffic transmission scheme after reconfiguration when network parameters such as topology, routing, and traffic change. It cannot guarantee the real-time nature of network reconfiguration and the deterministic transmission of traffic in the network. The patent application number is CN202310917826.3, entitled "TSN Device Reconfiguration Method". The specific approach is as follows: when a new TSN device is connected, the new topology and service flow requirements are determined by obtaining the adjacency table of the old TSN device and the traffic information of the service controller. Then, a GCL is generated based on the scheduling period, time slot frame length, transmission order, and guard band, and the configuration is distributed using the YANG model and NETCONFIG protocol. However, this distributed configuration method simply establishes the configuration process without involving the design of specific update times, configuration effective times, and detailed design of each port's GCL. It cannot guarantee reconfiguration consistency and the deterministic transmission requirements of traffic after reconfiguration. The patent application number is 202210184833.2, entitled "A Network Configuration Management Method for TSN Switches". The specific approach is as follows: when a new TSN device is connected, the new device is discovered through broadcast and the new global topology and traffic information are obtained. The central controller uses the traffic information to achieve joint optimization of routing scheduling, determines each traffic transmission path and GCL, and distributes them to the switch. However, this method does not consider the conflict between new and legacy traffic during network reconfiguration. Furthermore, the GCL generation scheme does not consider frame isolation in terms of transmission time for data frames from different traffic sources, causing conflicts during the transmission of different data frames and failing to guarantee the deterministic transmission requirements of traffic during reconfiguration. The patent application number is CN201810661209.0, entitled "A Tabu-Based Virtual Network Reconfiguration Method for SDN," which specifically addresses the problem of virtual network updates with congested nodes. Based on network node priorities, it uses a tabu search algorithm to determine the mapping of virtual nodes to physical nodes after the network update, taking into account the remaining CPU resources. However, this method only establishes the mapping scheme for the updated nodes and does not consider the transmission path of the updated traffic or the configuration of the GCL for each node, thus failing to guarantee the deterministic requirements of traffic transmission after the update.

[0005] Therefore, how to design a reconfiguration mechanism to ensure network consistency and real-time performance during the reconfiguration process, and how to design relevant reconfiguration algorithms based on this mechanism to obtain the transmission scheme after network reconfiguration at a lower time cost to ensure the determinism of traffic transmission, is an urgent problem to be solved. Those skilled in the art are committed to developing an online reconfiguration method for time-sensitive networks. Summary of the Invention

[0006] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is as follows:

[0007] In time-sensitive networks, the lack of a reconfiguration mechanism that can guarantee network consistency and real-time performance leads to conflicts between newly registered traffic and legacy traffic in the network after reconfiguration. This results in data frame loss or extremely high queuing latency in the switching nodes, failing to meet the low latency and low packet loss requirements of traffic. The lack of real-time performance also causes traffic to suffer from extremely high waiting time when injected into the network, making it impossible to guarantee the low latency requirements of traffic transmission.

[0008] In time-sensitive networks, there is a lack of efficient and fast algorithms for solving transmission schemes that can guarantee deterministic traffic transmission after reconfiguration. The extremely high time complexity of existing algorithms means that the reconfiguration calculation time from the time of downtime to the time of new scheme generation is much longer than the downtime response time. As a result, the new transmission scheme cannot be issued and deployed at the time of update, causing a large number of data frames to be lost after the new traffic begins to be injected at the time of update, which undermines the deterministic transmission requirement of low packet loss.

[0009] To achieve the above objectives, the present invention provides an online reconfiguration method for time-sensitive networks, the method comprising the following steps:

[0010] S1: Abstract the network topology of terminal devices, switches and links in the network to form a directed graph;

[0011] S2: Determine the link update time of each switch based on the downtime given by the system and the old traffic transmission scheme, and use this as the new reference time to inherit traffic.

[0012] S3: Incremental scheduling establishes deterministic constraints based on the already scheduled transmission scheme and current traffic attributes, and uses network calculus to calculate the start transmission time;

[0013] S4: According to the determined scheduling order of the incremental algorithm, continue until all traffic is scheduled. If not, return to step S3.

[0014] S5: Generate GCL according to the new scheduling scheme, and generate configuration files through configuration software and send them to switches and terminals to wait for them to take effect;

[0015] Furthermore, in step S1, the terminal device and switch are abstracted as nodes, and the link is abstracted as a directed edge;

[0016] Furthermore, the difference between the shutdown time and the update time mentioned in step S2 is the current link shutdown response time;

[0017] Furthermore, the update time is expressed by the following formula:

[0018] First, using the old transmission scheme, calculate the sequence number of traffic frames remaining in the network after the shutdown of different switch links when ρ takes effect. These are frames that have been sent from the source node but have not yet reached the destination node, specifically satisfying the following formula:

[0019]

[0020] In the formula, F represents the set of flows. Indicates flow rate f i During the injection time at the source node, Indicates the current flow f i The sequence number of the last frame remaining in the network after the shutdown takes effect;

[0021] Link V1-V 00 There is a 1st frame with residual traffic f1, and no subsequent links need to wait for this frame to pass. Therefore, the update times of each switch link that this frame passes through are obtained:

[0022]

[0023] In the formula, ε represents the set of links. Indicates the current link τ k At the time of the update, l i B represents the length of the traffic frame packet, and B represents the bandwidth.

[0024] Furthermore, the scheme inheritance of the flow in step S2 is expressed by the following formula:

[0025]

[0026] In the formula, R i Indicates flow rate f i The set of links traversed;

[0027] It also includes calculating the network reconfiguration supercycle based on traffic cycle information, denoted by C, which is obtained by the least common multiple of all traffic cycles, i.e.:

[0028] C = LCM(p i );

[0029] Furthermore, the deterministic constraints in step S3 include deadline constraints, frame sequence constraints, and frame isolation constraints;

[0030] The deadline constraint, calculated from the time each traffic frame is sent from the source node, requires that the end-to-end delay to reach the destination node not exceed the maximum allowable delay, as expressed by the formula:

[0031]

[0032] In the formula, Indicates the start time of transmission on the last hop link;

[0033] The frame sequence constraint stipulates that the transmission time of the same data frame on the previous switch link should be less than the transmission time on the subsequent switch link, as expressed by the formula:

[0034]

[0035] In the formula, H i This indicates the total number of links through which the current traffic passes. This indicates that the source node sends the link. Indicates traffic on the link The start time of transmission;

[0036] The frame isolation constraint ensures that frames from any two streams do not overlap within the same link's transmission interval, as expressed by the formula:

[0037]

[0038] In the formula, C represents the supercycle, which is obtained from the least common divisor of all flow cycles, i.e., C = LCM(p i );

[0039] Furthermore, the scheduling order of the incremental algorithm is sorted according to the traffic cycle and deadline information:

[0040] First, sort according to period, that is, for flow f i and f i+1 ;

[0041] If p i >p i+1 Then the flow f i Sort in stream f i+1 Before, if f i <f i+1 Then the flow f i Sort in stream f i+1 after;

[0042] If p i =p i+1Then sort by deadline, if d i <d i+1 Then the flow f i Sort in stream f i+1 Before, if d i >d i+1 Then the flow f i Sort in stream f i+1 after;

[0043] If d i =d i+1 Then the sorting results are determined manually according to actual needs;

[0044] Furthermore, the incremental scheduling described in step S3 is to calculate the transmission scheme of new traffic that is not inherited in an incremental solution manner, one flow at a time and one link at a time, based on the constraints and the sorting results.

[0045] Furthermore, in a single round of scheduling, the incremental scheduling traverses all links from the source node to the destination node, and establishes a switch remaining service curve and a link cumulative service curve for the current link. The worst-case latency is calculated using the network calculus method and used as the start time for the next link. The switch remaining service curve is expressed by the following formula:

[0046]

[0047] In the formula, Indicates link τ at time t k The arrival curve, Indicates link τ at time t k The service curve, Indicates link τ at time t k Remaining service capacity curve;

[0048] Furthermore, the link τ k The worst-case delay is the maximum horizontal distance between the arrival curve and the service curve, expressed by the following formula:

[0049]

[0050] The start time of the next link is the sum of the start time of the current link and the worst-case delay of the current link, thus obtaining the transmission scheme for all new traffic.

[0051] This invention proposes an online reconfiguration method for time-sensitive networks. By designing an asynchronous network reconfiguration mechanism, the network update time can be determined based on the actual downtime of each link, thus achieving asynchronous network reconfiguration. By ensuring that legacy traffic is transmitted before the update time, packet loss is avoided, guaranteeing network consistency and low packet loss determinism during reconfiguration. The earliest update time is chosen to ensure extremely low response time for this mechanism, thereby guaranteeing low-latency determinism in traffic transmission. Simultaneously, the update time also provides a baseline time for solving subsequent new schemes. An incremental reconfiguration transmission scheme solving algorithm based on network calculus is designed. Using the update times of each link calculated by the aforementioned mechanism as the baseline time for solving new transmission schemes, an inheritance judgment method is first used to obtain transmission schemes for some new traffic, reducing the amount of data required to solve the new scheme. Then, an incremental scheduling method is used to schedule traffic one by one, using a gradual iteration from local optimum to global optimum, avoiding the high computational complexity of simultaneously solving all unknowns. Ultimately, this algorithm transforms the computationally complex inter-frame conflict determination into worst-case latency calculation based on network computation when solving the new scheme, achieving a solution with linear time complexity. This avoids the exponential time complexity caused by conflict determination, thereby reducing the computational complexity of the solution algorithm and achieving extremely low time complexity. This avoids the problem that the new transmission scheme still cannot be deployed after the downtime response time ends, ensuring the deterministic transmission requirement with low packet loss. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of the architecture of the online reconfiguration method of the present invention;

[0053] Figure 2 This is a topology diagram of the time-sensitive network system of the present invention;

[0054] Figure 3 The network topology is a directed graph modeled according to this invention;

[0055] Figure 4 This is a schematic diagram of the scheme inheritance strategy of the present invention;

[0056] Figure 5 This is a schematic diagram illustrating the worst-case latency calculation in network computation according to the present invention. Detailed Implementation

[0057] The following description, with reference to the accompanying drawings, illustrates several preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0058] In the accompanying drawings, components with the same structure are indicated by the same numerical designation, and components with similar structures or functions are indicated by similar numerical designations. The dimensions and thicknesses of each component shown in the drawings are arbitrary, and the present invention does not limit the dimensions and thicknesses of each component. To make the illustrations clearer, the thickness of some components has been appropriately exaggerated in the drawings.

[0059] This invention provides an online reconfiguration method for time-sensitive networks, and its architecture diagram is shown below. Figure 1 As shown, it includes the following steps:

[0060] S1: Abstract the network topology of terminal devices, switches and links in the network to form a directed graph;

[0061] S2: Determine the link update time of each switch based on the downtime given by the system and the old traffic transmission scheme, and use this as the new reference time to inherit traffic.

[0062] S3: Incremental scheduling establishes deterministic constraints based on the already scheduled transmission scheme and current traffic attributes, and uses network calculus to calculate the start transmission time;

[0063] S4: According to the determined scheduling order of the incremental algorithm, continue until all traffic is scheduled. If not, return to step S3.

[0064] S5: Generate GCL based on the new scheduling scheme, and generate configuration files through configuration software and distribute them to switches and terminals for them to take effect.

[0065] In a more specific embodiment, the present invention designs a network asynchronous reconfiguration mechanism and an incremental reconfiguration transmission scheme solution algorithm based on network calculus.

[0066] First, the asynchronous network reconfiguration mechanism calculates the earliest update time on the switching nodes by analyzing the distribution of legacy traffic in the network. Legacy traffic is transmitted before the update time to ensure reconfiguration consistency. Asynchronous update times are configured for different switching nodes to reduce the waiting time for new traffic injected into the network, ensuring real-time reconfiguration.

[0067] The asynchronous network reconfiguration mechanism includes the following steps:

[0068] 1.1: A directed graph is formed by abstracting the network topology of terminal devices, switches, and links in the network. All terminal devices and switches are abstracted as nodes, and unidirectional links between devices are abstracted as directed edges. The network system topology and data routing information in this embodiment are as follows: Figure 2As shown, the TSN network consists of 6 terminal devices and 4 switch nodes. Each terminal device can be used for sending and receiving traffic. Each switch connects terminal devices and other switches for traffic relay. All links are full-duplex links. The corresponding abstracted network topology directed graph and data flow transmission paths are shown below. Figure 3 As shown.

[0069] 1.2: Determine the update time of each switch link based on the system-given downtime and the old traffic transmission scheme. The difference between the update time and the downtime is the current link downtime response time. In this example, the given downtime ρ for all links is 0.5ms, the length of all data frames is 1250 bytes, and the bandwidth of all links is B = 100Mbps. The periods of the three traffic flows under the old transmission scheme are {p1 = 2ms, p2 = 4ms, p3 = 8ms}, and the start transmission time of each hop at the switch is {O1:{V 00 -V1:0,V1-V 00 :1ms},O2=:{V 01 -V1:1ms,V1-V2:2ms,V2-V 02 :3ms},O3:{V 02 -V2:1ms,V2-V3:2ms,V3-V 03 :3ms}}, as shown in Table 1:

[0070] Table 1 Old Transmission Scheme

[0071]

[0072] When the network downtime is 0.5ms, the sequence numbers of traffic frames remaining in the network after ρ takes effect at the downtime are first calculated using the old transmission scheme. These are the frames that have been sent from the source node but have not yet reached the destination node. Specifically, the following formula is satisfied:

[0073]

[0074] In the formula, F represents the set of flows. Indicates flow rate f i During the injection time at the source node, Indicates the current flow f i The sequence number of the last frame remaining in the network after the shutdown takes effect. In this example, it's link V1-V. 00 The first frame of the legacy traffic f1 exists, and no subsequent links need to wait for this frame to pass. Therefore, the update times of each switch link that this frame passes through are obtained:

[0075]

[0076] In the formula, ε represents the set of links. Indicates the current link τ k At the time of the update, l i Let B represent the length of the traffic frame and B represent the bandwidth. In this example, at 0.5ms, the legacy frame in the network is the first frame of f1. Therefore, the update times of each link traversed by f1 are the times when the current link completes transmission, and the update times of the other links are the times when they stop. The data frame takes 0.1ms to complete transmission from the start of transmission. Finally, except for the directed link V1-V... 01 Except for the 1.1ms update time, the update time for all other links is 0.5ms.

[0077] Secondly, the incremental reconfiguration transmission scheme solution algorithm based on network calculus, on the basis of the aforementioned reconfiguration mechanism, transforms the inter-frame conflict determination with exponential time complexity into delay calculation through network calculus, and obtains the reconfiguration transmission scheme with linear complexity, which greatly reduces the calculation time of the reconfiguration scheme, reduces data frame packet loss during the solution of the new transmission scheme, and ensures the deterministic transmission requirements.

[0078] The algorithm for solving incremental reconfiguration transmission schemes based on network calculus includes the following steps:

[0079] 2.1: Under the condition that the route remains unchanged, a transmission scheme is used to determine whether new traffic can inherit the old traffic based on the update time of each link. The strategy diagram is shown below. Figure 4 As shown. If, in the old transmission scheme, the start time of traffic on all the links it traverses is later than the link update time, then the traffic can inherit the old transmission scheme. The condition for scheme inheritance is:

[0080]

[0081] In the formula R i Indicates flow rate f i The set of links traversed. In this example, traffic 2 and 3 can both inherit the old scheme.

[0082] 2.2: Calculate the network reconfiguration time-lapse based on traffic cycle information. Let C represent the time-lapse, which is obtained by finding the least common multiple of all traffic cycles, i.e.:

[0083] C = LCM(p i )

[0084] In this example, the overcycle is the least common multiple of 2ms, 4ms, and 8ms, which is 8ms.

[0085] 2.3: Based on the update time of each link, establish a series of deterministic constraints on link transmission and terminal equipment when new traffic is transmitted, including deadline constraints, frame sequence constraints, and frame isolation constraints. The key constraints are:

[0086] 2.3.1: Deadline Constraint. The end-to-end delay of each traffic frame, calculated from the time it is sent from the source node, should not exceed the maximum allowable delay. In this example, the periods of the four new traffic frames are {p1} * =2ms,p2 * =4ms,p3 * =8ms,p4 * =8ms}, the maximum allowable delays are {d1} * =6ms,d2 * =12ms,d3 * =24ms,d4 * =24ms}, which means it should satisfy the following formula:

[0087]

[0088] In the formula This indicates the start time of transmission on the last hop link.

[0089] 2.3.2: Frame sequence constraint: The transmission time of the same data frame on the previous switch link should be less than the transmission time on the subsequent switch link, i.e., satisfying the following formula:

[0090]

[0091] In the formula H i This indicates the total number of links through which the current traffic passes. This indicates that the source node sends the link. Indicates traffic on the link The start time of transmission.

[0092] 2.3.3: Frame isolation constraint: Frames from any two streams must not overlap within the transmission interval of the same link, i.e., the following equation must be satisfied:

[0093]

[0094] In the formula, C represents the supercycle, which is obtained from the least common divisor of all flow cycles, i.e., C = LCM(p i In this example, the timeout is 8ms.

[0095] 2.4: Based on the flow cycle and deadline information, sort the set of flows to be scheduled to determine the scheduling order of the incremental algorithm. First, sort by cycle, that is, for flow f... i and f i+1 ,

[0096] 2.4.1: If p i >p i+1 Then the flow f i Sort in stream f i+1Before, if f i <f i+1 Then the flow f i Sort in stream f i+1 after;

[0097] 2.4.2: If p i =p i+1 Then sort by deadline. If d i <d i+1 Then the flow f i Sort in stream f i+1 Before, if d i >d i+1 Then the flow f i Sort in stream f i+1 after;

[0098] 2.4.3: If d i =d i+1 The sorting results are then determined manually according to actual needs.

[0099] 2.5: Based on the constraints in step 3 and the sorting results in step 4, the transmission schemes for new, uninherited traffic are calculated incrementally, flow by flow and link by link. During scheduling, the incremental algorithm extracts one data stream from the traffic set in each round, calculating only the transmission scheme for the current stream. For traffic with newly generated transmission schemes, these schemes are considered known and added to the frame isolation constraints of the current traffic scheduling. In each round of scheduling, all links are traversed from the source node to the destination node, and a switch remaining service curve and a link cumulative service curve are established for the current link. The worst-case delay is calculated using network calculus and used as the start time for the next link. The switch remaining service curve is thus obtained as follows:

[0100]

[0101] In the formula, Indicates link τ at time t k The arrival curve, Indicates link τ at time t k The service curve, Indicates link τ at time t k Remaining service capacity curve. A diagram illustrating the service curve, arrival curve, and worst-case latency is shown below. Figure 5 As shown. Link τ k The worst-case delay is the maximum horizontal distance between the arrival curve and the service curve, as shown in the following formula:

[0102]

[0103] The start time of the next link is the sum of the start time of the current link and the worst-case delay of the current link. This yields the transmission schemes for all new traffic. In this example, the new transmission schemes are shown in Table 2. Flow 1, which cannot inherit the old scheme, and the newly added Flow 4 both generate new schemes.

[0104] Table 2 New Transmission Scheme

[0105]

[0106] 2.6: Generate GCL based on the generated new traffic scheduling scheme, and generate corresponding configuration files for terminal devices and switch ports through relevant configuration software. Finally, distribute the configuration software to the devices and make it effective after the update time.

[0107] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for online reconfiguration for time sensitive networking, characterized in that, The method comprises the following steps: S1: abstracting the terminal device, switch and link in the network to form a directed graph; S2: determining the update time of each switch link according to the given downtime and old traffic transmission scheme, taking the update time as a new reference time, and performing traffic inheritance according to the reference time; the update time is expressed by the following formula: First, the old transmission scheme is used to calculate the different switch links at the downtime moment The traffic frame sequence number left in the network after taking effect, specifically satisfies the following formula: wherein, denotes a flow set, denotes a flow the injection time at the source node, denotes the current flow the last frame sequence number left in the network after the effect of the down time instant Link V1-V 00 There is the 1st frame of the legacy traffic fl, while no subsequent link needs to wait for the frame transmission to pass, and then the update time of each switch link through which the frame passes is obtained: In the formula, denotes a link set, denotes a current link update time, denotes a traffic frame packet length, denotes a bandwidth; The scheme inheritance of the traffic is expressed by the following formula: In the formula, represents the flow rate the set of links passed through; Also included is calculating a network reconfiguration super-period from the traffic period information, using to represent the super-period, which is the least common multiple of all the traffic periods, specifically: ; S3: incremental scheduling, determining a deterministic constraint condition according to the scheduled transmission scheme and current traffic attribute, and calculating the start transmission time by using network calculus; the deterministic constraint condition comprises a deadline constraint, a frame sequence constraint and a frame isolation constraint; The deadline constraint is calculated from the time when each traffic frame is sent from the source node to the end-to-end delay of the traffic frame arriving at the destination node, and the maximum allowed delay is expressed by the following formula: In the formula, denotes the start transmission time on the last hop link; The frame sequence constraint is that the transmission time of the same data frame on the previous switch link should be less than the transmission time of the next switch link, which is expressed by the following formula: In the formula, represents the total number of links through which the current flow passes, represents the link from which the source node sends out, represents the start transmission time of the flow on the link . The frame isolation constraint is that the transmission intervals of frames of any two flows on the same link do not overlap, which is expressed by the following formula: ； The incremental scheduling is to calculate the transmission scheme of the un-inherited new traffic in an incremental manner according to the constraint condition and a preset sorting result, and the transmission scheme is calculated flow by flow and link by link; The sorting result is determined according to the traffic period and deadline information, and the sorting rule is as follows: First, the period is sorted: when the stream ordering in the stream preceding, when the stream ordering in the stream following; when, in order of the cutoff time, when, the stream is sorted in the stream before, when, the stream is sorted in the stream after; In this case, the ranking result is determined artificially according to actual needs. The network calculus calculation process is as follows: in one round of scheduling, all links passed by the source node to the destination node are traversed, the switch residual service curve and the link cumulative service curve are established for the current link, the worst delay is calculated and used as the start transmission time of the next link; the switch residual service curve is expressed by the following formula: wherein denotes the arrival curve of the link at time t, denotes the service curve of the link at time t, denotes the residual service capacity curve of the link at time t. The link The worst latency of the link is the maximum horizontal distance between the arrival curve and the service curve, which is expressed by the formula as follows: The start transmission time of the next link is the sum of the start transmission time of the current link and the worst delay of the current link, and thus the transmission scheme of all new traffics is obtained; S4: according to the scheduling order of the incremental algorithm determined, the scheduling of all traffics is completed, and if the scheduling is not completed, the step S3 is returned to; S5: generating a GCL according to the new scheduling scheme, and generating a configuration file through a configuration software and downloading the configuration file to the switch and terminal to wait for taking effect.

2. The time-sensitive networking oriented online reconfiguration method of claim 1, wherein, In step S1, the terminal device and switch are abstracted as nodes, and the link is abstracted as a directed edge.

3. The time-sensitive networking oriented online reconfiguration method of claim 1, wherein, In step S2, the difference between the downtime and the update time is the current link downtime response time.

Images