A control method, device and system of a time-sensitive network switch

Abstract

The application discloses a kind of control method, equipment and system of time sensitive network switch, method applied to time sensitive network TSN switch control plane includes: receiving control strategy issued by central controller, the control strategy includes constraint set, including performance constraint, resource constraint, behavior constraint;Based on constraint in control strategy, according to service demand and switch software and hardware capability determination flow configuration parameter;The flow configuration parameter is issued to the forwarding plane of TSN switch, to indicate that forwarding plane generates and loads TSN stream related hardware table item, executes TSN stream forwarding;Abstract information is reported to central controller, abstract information includes the processing result of forwarding plane execution TSN stream forwarding and / or the preset performance index aggregated based on processing result.This application can solve the technical problems, such as high single-point failure risk, large network response delay and poor scalability, existing in the current TSN centralized management and control mode.

Description

Technical Field

[0001] This invention belongs to the field of Time-Sensitive Networking (TSN) technology, and relates to TSN control methods, TSN switching equipment, and TSN control systems. Background Technology

[0002] With the continued implementation of industries such as connected vehicles, traditional Ethernet, lacking the ability to guarantee time-sensitive services, can no longer meet the requirements for high-reliability, low-latency data transmission. Time-Sensitive Networking (TSN), as a next-generation network technology based on Ethernet, provides deterministic transmission guarantees for time-sensitive services in the network through the IEEE 802.1 series of standards, becoming a key technological direction for solving the above problems.

[0003] In current TSN network architectures, a centralized management and control model is commonly adopted. This means that a single centralized controller manages all TSN switching devices in the network. Specifically, the centralized controller first needs to obtain the network topology information and service requirements, then perform network resource planning based on this information, and distribute the planning results to each TSN switching device. Finally, the forwarding plane of the switching device executes the corresponding policies.

[0004] However, this centralized management model has the following significant drawbacks:

[0005] 1. High risk of single point of failure: The centralized controller is the core of the entire TSN network architecture. Once the controller fails, all TSN switching devices will lose control, which may result in time-sensitive services in the network not being guaranteed, or even cause the entire network to be paralyzed.

[0006] 2. Large network response latency: Since all control commands need to be generated and issued through a centralized controller, when the network scale expands or business needs change dynamically, the computing pressure on the controller and the command transmission latency will increase significantly, resulting in control response time in the hundreds of milliseconds, which cannot meet the real-time requirements of milliseconds or even microseconds.

[0007] 3. Poor network scalability: In a centralized architecture, each new TSN switching device needs to be included in the management scope of the centralized controller, and the entire network resources need to be re-planned and configured. This is not only more complicated to operate, but may also cause the controller's management capabilities to reach a bottleneck, making it difficult to adapt to the expansion needs of large networks.

[0008] 4. Low resource utilization: Centralized controllers typically reserve resources based on the maximum demand of the entire network services, resulting in some switching devices having their port bandwidth and time slice resources idle for a long time, while other switching devices cannot meet service demands due to insufficient resources, resulting in a waste of network resources and low resource utilization.

[0009] 5. Poor heterogeneous compatibility: Centralized controllers need to issue precise hardware-level operation commands to TSN switching equipment from different manufacturers and models. This makes the implementation of the controller extremely complex and is not conducive to achieving plug-and-play and unified management of multi-vendor equipment. Summary of the Invention

[0010] The purpose of this invention is to provide a control method, device and system for time-sensitive network switches to solve the technical problems of high single-point failure risk, large network response delay and poor scalability in the existing centralized management and control mode of TSN.

[0011] To achieve the above-mentioned objectives, in a first aspect, the present invention provides a control method for a time-sensitive network (TSN) switch, applied to the control plane of a TSN switch, comprising:

[0012] The system receives control strategies issued by the central controller. The control strategies include a set of constraints, which includes one or more of performance constraints, resource constraints, and behavioral constraints.

[0013] Based on the constraints in the control strategy, the flow configuration parameters are determined according to business requirements and the hardware and software capabilities of the switch.

[0014] The flow configuration parameters are sent to the forwarding plane of the TSN switch to instruct the forwarding plane to generate and load TSN flow-related hardware entries and perform TSN flow forwarding.

[0015] The summary information is reported to the central controller. The summary information includes the processing result of the forwarding plane performing TSN flow forwarding and / or the aggregated preset performance index obtained based on the processing result.

[0016] Furthermore, the determination of flow configuration parameters based on constraints in the control strategy, according to service requirements and the hardware and software capabilities of the switch, includes:

[0017] Extract the performance constraints, resource constraints, and / or behavioral constraints contained in the control strategy;

[0018] Based on the extracted constraints, with business requirements as the target and the switch's hardware and software capabilities as the physical feasible domain, the executable flow configuration parameters of the switch's forwarding plane are determined. The flow configuration parameters include one or more of the following: flow identification rules, scheduling gating list, bandwidth threshold, queue mapping, priority, flow supervision parameters, and redundant path table.

[0019] Furthermore, the method also includes:

[0020] In the event of a detected link interruption, port failure, or device malfunction, the flow configuration parameters are adjusted based on the constraints in the control policy. The adjustment of the flow configuration parameters enables path switching, bandwidth reallocation, and / or gating scheduling adjustments. The adjustment results are reported to the central controller as part of the summary information.

[0021] Furthermore, the method also includes:

[0022] After detecting that a new neighboring switch has joined the network and completed link negotiation, the control policy is synchronized to the new neighboring switch; the summary information includes the switch's network entry result.

[0023] Furthermore, the method also includes:

[0024] When a switch needs to coordinate across nodes, it negotiates with neighboring switches using a distributed negotiation algorithm based on constraints in the control policy. The negotiation includes bandwidth allocation, path coordination, or priority mapping. The negotiation results are then reported to the central controller as part of the summary information.

[0025] Secondly, the present invention provides a control method for a time-sensitive network switch, applied to a central controller, comprising:

[0026] A control policy is issued to the control plane of a Time-Sensitive Network (TSN) switch. The control policy includes a set of constraints, which includes one or more of performance constraints, resource constraints, and behavioral constraints. The control policy is used to constrain the switch control plane to determine flow configuration parameters. The flow configuration parameters are determined based on the constraints in the control policy, according to service requirements and the switch's hardware and software capabilities. They are used to instruct the switch forwarding plane to generate and load TSN flow-related hardware entries and execute TSN flow forwarding.

[0027] The receiver receives summary information reported by the switch control plane. The summary information includes the processing results of the forwarding plane performing TSN flow forwarding and / or the aggregated preset performance indicators obtained based on the processing results.

[0028] Furthermore, the method also includes: performing global optimization based on the summary information to update the control strategy.

[0029] Furthermore, the global optimization includes:

[0030] The central controller summarizes and analyzes the summary information reported by all switches in the network. The data analysis includes one or more of the following: multi-node data normalization and alignment, statistical analysis of network performance indicators, anomaly and bottleneck identification, load balancing analysis, failure mode and reliability analysis, and resource utilization efficiency analysis.

[0031] Based on the data analysis results and combined with the preset global optimization objectives, a new round of control strategies adapted to the current state of the entire network are generated through intelligent algorithms. The global optimization objectives consider one or more of the following: network resource utilization, end-to-end latency, network load balancing, and redundancy cost. The intelligent algorithms include one or more of the following: constraint satisfaction problem solving algorithms, Bayesian network probabilistic reasoning algorithms, fault propagation graph deduction algorithms, genetic algorithms, and particle swarm optimization algorithms.

[0032] Thirdly, the present invention provides a time-sensitive network switch device, including a control plane and a forwarding plane;

[0033] The control plane is used to receive control policies issued by the central controller. The control policies include a set of constraints, which includes one or more of performance constraints, resource constraints, and behavioral constraints. Based on the constraints in the control policies, the control plane determines flow configuration parameters according to service requirements and the hardware and software capabilities of the switch. The control plane issues the flow configuration parameters to the forwarding plane. The control plane also reports summary information to the central controller. The summary information includes the processing results of the forwarding plane performing TSN flow forwarding and / or the aggregated preset performance indicators obtained based on the processing results.

[0034] The forwarding plane is used to generate and load hardware entries related to Time Sensitive Network (TSN) flows based on the flow configuration parameters issued by the control plane, perform TSN flow forwarding, and feed back the processing results of TSN flow forwarding to the control plane.

[0035] Fourthly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method described in the first or second aspect.

[0036] Fifthly, the present invention provides a control system for a time-sensitive network switch, including a central controller and a time-sensitive network (TSN) switch control plane, wherein the switch control plane is used to implement the method described in the first aspect, and the central controller is used to implement the method described in the second aspect.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The present invention provides a control method for a time-sensitive network switch (TSN), which offloads management and control functions from the central controller to the TSN switch control plane. The TSN switch control plane receives control policies from the central controller, determines flow configuration parameters based on constraints within the control policies and the switch's hardware and software capabilities, and sends these parameters to the forwarding plane. This instructs the forwarding plane to generate and load TSN flow-related hardware entries, execute TSN flow forwarding, and achieve autonomous decision-making without multiple round-trip communications with the central controller. This method reduces management and control response time from hundreds of milliseconds in traditional centralized architectures to microseconds, meeting the stringent real-time requirements of time-sensitive services. Simultaneously, it eliminates the single point of failure risk of the central controller. Even if the central controller is disconnected or fails, the TSN switch can still operate autonomously according to the received control policies, improving the overall reliability and resilience of the network. Furthermore, the central controller only needs to issue control policies and does not need to concern itself with the underlying implementation details of different manufacturers and models of switches, reducing the complexity of network deployment and maintenance and improving network scalability.

[0039] 2. In the control method for a time-sensitive network switch provided by this invention, the processing results of TSN flow forwarding and / or the aggregated preset performance indicators obtained based on the processing results are reported as summary information to the central controller. This helps the central controller to perform global optimization to update the control strategy, forming a closed-loop control of distribution, execution, feedback, and optimization. This can fully leverage the real-time advantages of TSN switch autonomy and ensure that the overall behavior of the entire network always evolves towards the global optimization goal. Thus, it takes into account both the flexibility of local autonomy and the synergy of global control, and can achieve optimal allocation of global resources and long-term stable operation.

[0040] 3. The time-sensitive network switch device provided by the present invention includes a control plane and a forwarding plane. The control plane receives control policies and determines flow configuration parameters, and the forwarding plane performs TSN flow forwarding according to the flow configuration parameters. Its beneficial effects are consistent with the control method of the first aspect.

[0041] 4. The beneficial effects of the electronic device provided by the present invention are consistent with the control methods of the first and second aspects.

[0042] 5. The control system for the time-sensitive network switch provided by the present invention has the same beneficial effects as the control methods in the first and second aspects. Attached Figure Description

[0043] Figure 1 A schematic flowchart of a control method for a time-sensitive network switch provided in an embodiment of the present invention (applied to the control plane of a TSN switch).

[0044] Figure 2A schematic flowchart of a control method for a time-sensitive network switch provided in an embodiment of the present invention (applied to a central controller).

[0045] Figure 3 This is a schematic diagram of a switching device in a time-sensitive network provided in an embodiment of the present invention.

[0046] Figure 4 A schematic diagram of the functional modules of a time-sensitive network switch provided in an embodiment of the present invention.

[0047] Figure 5 This is a schematic diagram illustrating the functional module settings and interaction flow of the control plane in a switch device provided in an embodiment of the present invention.

[0048] Figure 6 This is a schematic diagram illustrating the functional module settings and interaction process of the forwarding plane in a switch device provided in an embodiment of the present invention.

[0049] Figure 7 This is a schematic diagram of the interaction process between the control plane and the forwarding plane in a switch device provided in an embodiment of the present invention.

[0050] Figure 8 This is a schematic diagram of the electronic device structure provided in an embodiment of the present invention.

[0051] Figure 9 This is a schematic diagram of the control system structure of a time-sensitive network switch provided in an embodiment of the present invention. Detailed Implementation

[0052] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0053] To facilitate understanding and explanation of the technical solutions of the embodiments of the present invention, the relevant prior art and terms will be explained below.

[0054] Time-Sensitive Networking (TSN) is a set of Ethernet extension protocols developed by the IEEE 802.1 working group. Its core objective is to endow traditional "best-effort" Ethernet with deterministic time guarantees, low latency, and high reliability, enabling standard Ethernet to meet the real-time requirements of critical scenarios. Its predecessor was the TSN working group, renamed from the Audio / Video Bridge working group in November 2012. Essentially, it is a derivative extension of the IEEE 802.1Q (Virtual Local Area Network) standard, focusing on achieving ultra-low latency and high availability in transmission.

[0055] Centralized management refers to a management model that unifies the management authority, core resources, key decisions, and operation and maintenance functions that are scattered across various levels and units into a central node (management entity). This entity then coordinates, schedules, approves, and supervises these functions. The core of centralized management is centralized coordination, unified standardization, and efficient collaboration, making it suitable for management scenarios that require strong consistency and compliance.

[0056] High-availability Seamless Redundancy (HSR) is a ring network redundancy protocol for industrial Ethernet and critical communication scenarios. Its core objective is to achieve zero interruption, zero latency, and zero packet loss in communication links and data transmission, providing reliable communication assurance for scenarios with extremely high business continuity requirements.

[0057] Parallel Redundancy Protocol (PRP) is a highly available, seamless redundancy protocol based on a dual-network parallel architecture. Its core logic is to achieve communication redundancy through "dual-network independent parallelism and dual data transmission and single reception" to ensure uninterrupted communication when a single link or node fails.

[0058] IEEE 802.1 Qci is a fine-grained management standard for Ethernet data streams, with its core technology being Per-Stream Filtering and Policing (PSFP). This standard distinguishes each independent data stream in the network, configuring it with dedicated filtering rules, bandwidth limits, and priority policies to ensure stable transmission of critical flows and guarantee network determinism and reliability.

[0059] The control method applied to the control plane of a TSN switch according to the embodiments of the present invention will be further described below.

[0060] like Figure 1 As shown in the figure, an embodiment of the present invention provides a control method for a time-sensitive network switch, which is applied to the control plane of a TSN switch. The method may include the following steps S11 to S14.

[0061] S11. Receive the control strategy issued by the central controller. The control strategy includes a set of constraints, which includes one or more of performance constraints, resource constraints, and behavioral constraints.

[0062] S12. Based on the constraints in the control strategy, determine the flow configuration parameters according to business requirements and the hardware and software capabilities of the switch.

[0063] S13. Send the flow configuration parameters to the forwarding plane of the TSN switch to instruct the forwarding plane to generate and load TSN flow-related hardware entries and perform TSN flow forwarding;

[0064] S14. Report summary information to the central controller. The summary information includes the processing result of the forwarding plane performing TSN flow forwarding and / or the aggregated preset performance indicators obtained based on the processing result.

[0065] In this embodiment, the TSN switch control plane receives control policies from the central controller. Based on the constraints in the control policies, it determines flow configuration parameters according to service requirements and the switch's hardware and software capabilities, and sends these parameters to the forwarding plane. This instructs the forwarding plane to generate and load TSN flow-related hardware entries, execute TSN flow forwarding, and achieve autonomous decision-making without requiring multiple round-trip communications with the central controller. This embodiment reduces the control response time from hundreds of milliseconds in traditional centralized architectures to microseconds, meeting the stringent real-time requirements of time-sensitive services. Simultaneously, it eliminates the single point of failure risk of the central controller. Even if the central controller is disconnected or fails, the TSN switch can still operate autonomously according to the received control policies, improving the overall reliability and resilience of the network. The central controller only needs to issue control policies and does not need to concern itself with the underlying implementation details of different manufacturers and models of switches, reducing the complexity of network deployment and maintenance and improving network scalability. Furthermore, in this embodiment, the processing results of TSN flow forwarding and / or the aggregated preset performance indicators obtained based on the processing results are reported to the central controller as summary information. This helps the central controller to perform global optimization to update the control strategy, forming a closed-loop control of distribution, execution, feedback, and optimization. This can fully leverage the real-time advantages of TSN switch autonomy and ensure that the overall behavior of the entire network always evolves towards the global optimization goal. Thus, it takes into account both the flexibility of local autonomy and the synergy of global management and control, and can achieve optimal allocation of global resources and long-term stable operation.

[0066] In some possible implementations, step S12, based on the constraints in the control policy, determines the flow configuration parameters according to service requirements and the switch's hardware and software capabilities. This may include: extracting performance constraints, resource constraints, and / or behavioral constraints contained in the control policy; and based on the extracted constraints, taking service requirements as the target and the switch's hardware and software capabilities as the physical feasible domain, determining the flow configuration parameters executable by the switch's forwarding plane. The flow configuration parameters include one or more of the following: flow identification rules, scheduling gating lists, bandwidth thresholds, queue mappings, priorities, flow policing parameters, and redundant path tables.

[0067] In this embodiment, constraint extraction transforms the global intent of the central controller into quantifiable constraints that can be locally computed and mapped on the control plane. Based on the real-time status of the local network, and combined with service requirements and the hardware and software capabilities of the switches, flow configuration parameters are jointly determined. Service requirements determine the mapping target, including service type (e.g., time-triggered streams, audio / video streams, critical control streams, surveillance video streams, etc.), traffic characteristics (rate, frame length, periodicity, bursts, etc.), and transmission quality requirements (e.g., latency sensitivity, reliability sensitivity, isolation sensitivity, etc.). These service requirements determine the core indicators and implementation goals that the flow configuration parameters must guarantee. The hardware and software capabilities of the switches determine the physical feasible domain of the mapping, including the switch network location (e.g., core layer, aggregation layer, edge layer, etc.), port hardware capabilities (e.g., line speed, support for scheduling mechanisms, support for frame preemption, etc.), queue size, time synchronization accuracy, and buffer size. These hardware and software capabilities determine the achievable range and accuracy of the flow configuration parameters. For example, the network location of a switch can be used to determine a redundant path table, port hardware capabilities can be used to determine flow identification rules, the number of queues can be used to determine priorities and queue mappings, time synchronization accuracy can be used to determine a scheduling gating list, and buffer size can be used to determine flow policing parameters and bandwidth thresholds.

[0068] In this embodiment, by extracting performance constraints, resource constraints, and / or behavioral constraints from the control policy and combining them with the switch's local hardware and software capabilities, the constraints are mapped to flow configuration parameters executable in the forwarding plane, such as flow identification rules, scheduling gating lists, and bandwidth thresholds. This achieves decoupling and transformation from global abstract intent to local concrete execution. The conversion process from control policy to flow configuration parameters is applicable to TSN flow configuration in different scenarios and for different services, demonstrating strong versatility.

[0069] The following example illustrates the process of determining flow configuration parameters in the control plane of a TSN switch and sending flow configuration parameters to the forwarding plane, using a specific application scenario as an example.

[0070] In practical applications, the central controller does not issue specific register configurations, queue gating times, port rates, or other local operation commands for the TSN switch. Instead, it issues control strategies that include a set of constraints. For example, constraints could include: end-to-end latency for critical services not exceeding a preset limit; jitter not exceeding a preset limit; link bandwidth utilization exceeding a preset threshold (e.g., 50%); device fault recovery time not exceeding a preset limit; packet loss rate approaching 0%; control flow having the highest priority, video flow having the next highest priority; maximum reserved bandwidth for each flow not exceeding a preset percentage; at least one redundant path provided between any two points; local devices not allowed to monopolize more than a preset percentage of backplane bandwidth; and new devices must undergo topology authentication upon access.

[0071] The TSN switch control plane completes policy localization. Based on the constraints in the control policy, combined with local access service types, service real-time requirements, service traffic characteristics and other service requirements, as well as the switch's network location, port hardware capabilities, queue quantity, time synchronization accuracy, buffer size and other switch hardware and software capabilities, specific flow configuration parameters that can be executed locally are determined.

[0072] The process of autonomous decision-making in the control plane involves determining and issuing flow configuration parameters, which can include three steps: policy parsing, parameter mapping, and execution.

[0073] In the strategy parsing step, the performance constraints, resource constraints, and / or behavioral constraints contained in the control strategy are extracted.

[0074] In the parameter mapping step, based on the extracted constraints, with business requirements as the target and the switch's hardware and software capabilities as the physical feasible domain, the executable flow configuration parameters of the switch's forwarding plane are determined. These parameters include flow identification rules, scheduling gating lists, bandwidth thresholds, queue mapping, priorities, flow policing parameters, and one or more of the redundant path table. In practical applications, targeted mapping is performed based on the parsed constraint types, combined with business requirements and the switch's hardware and software capabilities. For example, this may include the following aspects:

[0075] Flow classification mapping: Based on business priority and behavioral constraints, priority code points (PCP) / differentiated services code points (DSCP), queue priorities, and PSFP rules are mapped according to business level.

[0076] Scheduling parameter generation: Based on performance constraints such as latency and jitter, and combined with the time synchronization accuracy and queue scheduling capability of the switch, a gate control list (GCL), time slot window, and queue scheduling mode are generated according to the latency target.

[0077] Bandwidth parameter calculation: Based on performance constraints such as bandwidth utilization and port utilization, combined with service traffic characteristics and port hardware capabilities, the Committed Information Rate (CIR), Excess Information Rate (EIR), and queue bandwidth allocation are generated.

[0078] Path and redundancy configuration: Based on performance constraints such as hop count and fault recovery time, combined with the switch topology location and path routing capabilities, generate primary and backup paths, failover mechanisms, fault switching thresholds and redundancy group configurations.

[0079] During the execution step, the flow configuration parameters are sent to the forwarding plane, which then writes the corresponding hardware entries, such as flow tables, gating lists, PSFP tables, and port configuration tables.

[0080] In scenarios with high system compliance and reliability requirements, before the controller issues the flow configuration parameters to the forwarding plane, it also includes: verifying whether the flow configuration parameters meet the constraints in the control policy, ensuring that the locally generated flow configuration parameters do not violate the constraints in the control policy issued by the central controller, ensuring that the local autonomous behavior always operates within the constraints of the global control policy, and avoiding the flow configuration parameters generated due to local decision-making bias from violating the upper or lower limits of the constraints in the global policy. For example, it ensures that the gating time slot length allocated to a certain flow will not cause its end-to-end latency to exceed the upper limit specified by the policy.

[0081] Through the above steps, the decoupling and transformation from global abstract intent to local specific execution is achieved. The switch control plane can autonomously combine local real-time status to dynamically adjust flow configuration parameters, which not only meets the control policy constraints issued by the central controller, but also adapts to changes in local hardware and software resources and service load, thereby improving the network's autonomous operation and maintenance capabilities.

[0082] For the constraints in the control strategies of the aforementioned examples, the determination of flow configuration parameters can be as follows: For latency and jitter upper limit constraints, these can be determined as queue scheduling gating period, maximum queue length, and frame preemption threshold; for bandwidth utilization threshold constraints, these can be determined as flow bandwidth reservation upper limit, port shaping rate, traffic shaping scheduling parameters, and time-aware scheduling parameters; for fault recovery time upper limit constraints, these can be determined as fault detection period, path switching trigger threshold, and redundancy group switching latency; for packet loss rate constraints, these can be determined as buffer reservation strategy, retransmission enablement, and high-priority flow preemption mechanism; for service priority constraints, these can be determined as 802.1Q priority mapping, queue number, and scheduling weight; for single-flow maximum reserved bandwidth constraints, these can be determined as redundancy group configuration, fault failover strategy, and multi-path routing; for backplane bandwidth exclusive limitation constraints, these can be determined as device-level resource scheduling threshold and flow aggregation admission rules; for new device topology authentication constraints, these can be determined as access port enable strategy and topology legality verification rules.

[0083] More specifically, regarding the constraints in the control strategy issued by the central controller: "The upper limit of the latency of all monitoring video streams is 10ms, and the priority is 7", the performance constraints "end-to-end latency of monitoring video streams ≤ 10ms" and the behavioral constraints "the priority of monitoring video streams is 7" are extracted first. Combined with the business requirements "the monitoring video stream is a periodic constant rate service with a single stream base bandwidth of 15Mbps. It requires continuous transmission, stable latency, no blocking, and no interruption. It belongs to the critical TSN service", the local hardware and software capabilities of the switch are: "(1) Supports 8 output queues Queue 0-7, with Queue 7 having the highest hardware priority; (2) Supports Qbv time-aware scheduling with a minimum gating granularity of 1μs and supports a scheduling cycle of 125μs; (3) Supports minimum bandwidth reservation for queues, guaranteed bandwidth type, and Qci frame preemption function; (4) Supports port latency collection and path constraint control, and can configure routing blacklists and whitelists; (5) The native single-hop forwarding processing latency is ≤ 100μs. Queue 7 has sufficient dedicated buffer, and the hardware itself will not introduce additional latency loss." According to the real-time status of the local network, "current path hop count ≤ 5 hops, single-hop link propagation delay ≤ 1ms, port queue current queuing delay ≤ 200μs", the local executable flow configuration parameters can be determined as follows: (1) Queue mapping: TSN flow priority 7 → switch local Queue 7 (highest priority queue), corresponding to the control policy "priority 7". Direct mapping, the priority is a hierarchical relationship: TSN flow priority is an abstract requirement, and Queue 7 is a local specific queue implementation; (2) Gating list configuration: Queue 7 is opened once every 125μs, and the opening time is ≥ 80μs. The core corresponds to "delay limit 10ms" to ensure that the single-hop queuing delay is ≤ 200μs, and the propagation delay and forwarding delay of the 5-hop path are superimposed to finally meet the end-to-end delay ≤ 10ms; (3) Bandwidth allocation: Queue 7 reserves a minimum bandwidth of 20Mbps. The single-stream base bandwidth of the monitoring video stream is 15Mbps to avoid queuing delay increase due to insufficient bandwidth, thus indirectly ensuring "delay limit of 10ms"; (4) Qci configuration: Queue 7 is configured as "guaranteed bandwidth" type, enabling frame preemption function, and is not preempted by low-priority streams (Queue 0~6), thus avoiding delay increase due to preemption. (5) Path constraint: The stream is prohibited from passing through ports with current queuing delay > 300μs. Those skilled in the art can understand that the constraint of "delay limit of 10ms for all monitoring video streams" in the control strategy does not correspond to a single stream configuration parameter, but is jointly implemented by gating list configuration, bandwidth allocation, Qci configuration, etc. Among them, gating list configuration is the core instruction, which directly controls single-hop queuing delay, thereby ensuring that the total end-to-end delay does not exceed 10ms.

[0084] In some possible implementations, the control method further includes: adjusting flow configuration parameters based on constraints in the control strategy when a link interruption, port failure, or device abnormality is detected; implementing path switching, bandwidth reallocation, and / or gating scheduling adjustment based on the adjustment of flow configuration parameters; and reporting the adjustment results to the central controller as part of the summary information.

[0085] For example, the link status of the main link (e.g., port 1 → neighboring port A) can be monitored in real time with a fault detection cycle of 100μs. Upon detecting a link interruption, actions are adjusted under the guidance of a control policy. For instance, the core constraints of the control policy guiding the adjustment include: critical flows use 1+1 path protection, with redundant protection paths only activated during faults and not occupying bandwidth under normal conditions; the end-to-end latency of critical flows is ≤10ms, and the single-hop forwarding latency is ≤200μs; fault failover time is ≤5ms; any link utilization is <65%; flow priority is given to lightly loaded links to avoid hotspot links.

[0086] Based on the above control strategy, the specific steps for adjusting the flow configuration parameters are as follows:

[0087] (1) Path switching: Immediately enable the backup path (port 2 → neighboring port B) of the monitoring video stream, migrate the flow configuration command of the original main path (port 1 → neighboring port A) to port 2 corresponding to the backup path, and at the same time disable the flow forwarding function of port 1 of the main path to ensure that the fault switching time is ≤5ms, which meets the control policy constraints of "fault switching time ≤5ms" and "redundant protection path is only enabled in case of fault".

[0088] (2) Bandwidth reallocation: Query the backup path (port 2 → neighboring machine port B) and find that the current bandwidth utilization rate is 45% (lower than the utilization rate of any link constrained by the control policy < 65%). Reallocate a minimum bandwidth of 20Mbps (consistent with the original main path bandwidth) to the monitoring video stream to ensure sufficient bandwidth and avoid delay exceeding the standard due to insufficient bandwidth, which meets the constraint of "end-to-end delay limit of critical flow ≤ 10ms".

[0089] (3) Gating scheduling adjustment: Based on the current queuing delay (150μs) of the backup path port 2, adjust the gating opening time of Queue 7 (the local queue corresponding to the priority 7 flow) to 85μs (the original main path was 80μs) to ensure that the single-hop queuing delay is ≤200μs. With the propagation delay of the backup path added, the end-to-end delay is still ≤10ms, which meets the delay boundary constraint.

[0090] (4) Status reporting: After the adjustment is completed, the fault status, path switching results, and adjusted flow configuration parameters are reported to the central controller as summary information so that the central controller can grasp the status of the entire network.

[0091] This embodiment enables rapid detection and autonomous recovery from local faults, enhancing the network's dynamic adaptability and fault self-healing capabilities. The fault detection and recovery process is completed entirely locally, without waiting for instructions from the central controller, compressing fault recovery time to the millisecond level and ensuring the continuity and stability of service transmission. Furthermore, the adjustment results are reported as part of the summary information, enabling the central controller to synchronously grasp the fault handling status of the entire network and support subsequent global optimization.

[0092] In some possible implementations, the control method further includes: after detecting that a new neighboring switch has joined the network and completed link negotiation, synchronizing the control policy to the new neighboring switch; the summary information includes the switch joining result.

[0093] In this embodiment, when a new switch joins the network, there is no need for the central controller to issue control policies one by one. Adjacent switches can automatically complete policy synchronization, enabling the newly joined switch to quickly join the TSN network based on the constraints in the control policy. This further reduces the management burden on the central controller and improves the network's plug-and-play capability and expansion efficiency. Simultaneously, the switch joining result is included in the summary information and reported to the central controller, allowing the central controller to monitor network topology changes in real time and provide accurate topology information for subsequent global optimization.

[0094] In some possible implementations, the control method further includes: when the switch needs to coordinate across nodes, negotiating with neighboring switches based on constraints in the control policy using a distributed negotiation algorithm, the negotiation including bandwidth allocation, path coordination, or priority mapping; and reporting the negotiation results as part of the summary information to the central controller.

[0095] In this embodiment, adjacent TSN switches can complete cross-node bandwidth allocation, path coordination, or priority mapping based on constraints in the control policy through a distributed negotiation algorithm, further improving network scalability and reducing global communication overhead. For cross-switch service flow coordination requirements, relevant nodes can directly complete resource coordination through distributed negotiation, only needing to report the negotiation results to the central controller. This avoids the central controller participating in every local negotiation process, sharing the computational load of the central controller and making the deployment of large-scale TSN networks possible.

[0096] The following example demonstrates distributed negotiation for cross-node bandwidth allocation, under the global objective of maximizing the utilization of network resources.

[0097] Negotiation Trigger Scenario: Non-critical flows of switch A (local) need to borrow TSN reserved bandwidth, but its local reserved bandwidth is insufficient (only 10Mbps), while the TSN reserved bandwidth of the adjacent switch B is 60Mbps. Therefore, cross-node bandwidth borrowing negotiation with switch B is required through a distributed negotiation algorithm.

[0098] The constraints in the control strategy are as follows: non-critical flows can dynamically borrow bandwidth reserved by the TSN, but must not affect critical flows; the bandwidth utilization of each switch port must not be lower than 60% and not higher than 80% of the threshold; the maximum bandwidth limit for non-critical flows is ≤80Mbps. The distributed negotiation process strictly following the control strategy can be as follows:

[0099] (1) Initiate negotiation: The control plane of switch A sends a bandwidth borrowing request to the control plane of switch B, specifying the bandwidth borrowing requirement (20Mbps), the borrowing duration and the current bandwidth occupancy of its own critical flow;

[0100] (2) Constraint verification negotiation: The control plane of switch B, combined with the local status, verifies whether it complies with the control policy constraints. It can be: querying that the local critical flow bandwidth occupancy is 300Mbps (400Mbps reserved bandwidth), the remaining 100Mbps can be borrowed, and after borrowing 20Mbps, the local port bandwidth utilization rate is 80% (80%≥X≥60%), which does not violate any constraints;

[0101] (3) Negotiation reached: The two parties determined the details of bandwidth borrowing through a distributed fair allocation algorithm: Switch B borrowed 20Mbps of reserved bandwidth from Switch A, and agreed that when Switch B’s critical flow bandwidth demand increased, the borrowed bandwidth would be immediately reclaimed;

[0102] (4) Result reporting and execution: Both parties’ control planes adjust the local bandwidth allocation instructions based on the negotiation results, and report the negotiation results as summary information to the central controller.

[0103] The control method applied to the central controller according to the embodiments of the present invention will be further described below.

[0104] like Figure 2 As shown in the figure, an embodiment of the present invention provides a control method for a time-sensitive network switch, which is applied to a central controller. The method may include the following steps S21 to S22.

[0105] S21. Issue a control policy to the control plane of the Time-Sensitive Network (TSN) switch. The control policy includes a set of constraints, which includes one or more of performance constraints, resource constraints, and behavioral constraints. The control policy is used to constrain the switch control plane to determine flow configuration parameters. The flow configuration parameters are determined based on the constraints in the control policy, according to service requirements and the switch's hardware and software capabilities, and are used to instruct the switch forwarding plane to generate and load TSN flow-related hardware entries and execute TSN flow forwarding.

[0106] S22. Receive summary information reported by the switch control plane, the summary information including the processing result of the forwarding plane performing TSN flow forwarding and / or the aggregated preset performance index obtained based on the processing result.

[0107] In this embodiment, the central controller only needs to issue control policies to the switch control plane to constrain the switch control plane to determine flow configuration parameters. It does not need to care about the underlying implementation details of switches from different manufacturers and models. As long as different TSN switches have a unified control policy parsing interface in their control plane, they can seamlessly access the same TSN network, which significantly reduces the complexity of network deployment and operation and maintenance and improves the scalability of the network.

[0108] Preferably, the control method applied to the central controller further includes: S23, performing global optimization based on the summary information to update the control strategy. By performing global optimization based on the summary information reported by the switches to update the control strategy, the central controller can discover and resolve potential global problems in distributed autonomous operation, and generate a new round of control strategies to calibrate and guide the TSN switches. This forms a closed-loop control of distribution, execution, feedback, and optimization, which can fully leverage the real-time advantages of TSN switch autonomy and ensure that the overall behavior of the entire network always evolves towards the global optimization goal. This balances the flexibility of local autonomy with the synergy of global control, enabling optimal allocation of global resources and long-term stable operation.

[0109] In some possible implementations, step S23, the global optimization, includes:

[0110] The central controller summarizes and analyzes the summary information reported by all switches in the network. The data analysis includes one or more of the following: multi-node data normalization and alignment, statistical analysis of network performance indicators, anomaly and bottleneck identification, load balancing analysis, failure mode and reliability analysis, and resource utilization efficiency analysis.

[0111] Based on the data analysis results and combined with the preset global optimization objectives, a new round of control strategies adapted to the current state of the entire network are generated through intelligent algorithms. The global optimization objectives consider one or more of the following: network resource utilization, end-to-end latency, network load balancing, and redundancy cost. The intelligent algorithms include one or more of the following: constraint satisfaction problem solving algorithms, Bayesian network probabilistic reasoning algorithms, fault propagation graph deduction algorithms, genetic algorithms, and particle swarm optimization algorithms.

[0112] In this embodiment, the central controller summarizes and analyzes the summary information reported by all switches across the network, providing a reliable and unified input of the network-wide status for the intelligent algorithm. Based on the data analysis results and combined with the preset global optimization target, the intelligent algorithm generates a new round of control strategies adapted to the current network-wide status. This leverages the real-time advantages of the local autonomy of the TSN switches while ensuring that the overall behavior of the network always evolves towards the global optimization target. It balances the flexibility of local autonomy with the synergy of global control, achieving optimal allocation of global resources and long-term stable operation.

[0113] Specifically, the global optimization objective serves as the top-level guide for the central controller's network-wide scheduling. This can include maximizing network resource utilization, minimizing end-to-end latency, achieving network-wide load balancing, and minimizing redundancy costs, or a combination of multiple objectives such as network resource utilization, end-to-end latency, network-wide load balancing, and redundancy costs. The control strategy is a set of constraints issued by the central controller to each TSN switch after data analysis and intelligent algorithms, excluding specific port numbers, bandwidth values, and other flow parameters. Examples of control strategies related to the global optimization objective are shown in Table 1.

[0114] Table 1 Examples of Global Optimization Objectives and Corresponding Control Strategies

[0115]

[0116] The data analysis is used to normalize, statistically analyze, identify bottlenecks, perform load analysis, fault analysis, and trend prediction on the summary information reported by multiple switches, providing a reliable and unified network-wide status input for subsequent intelligent algorithms. Data analysis may include one or more of the following: multi-node data normalization and alignment, statistical analysis of network-wide performance indicators, anomaly and bottleneck identification, load balancing analysis, fault mode and reliability analysis, and resource utilization efficiency analysis.

[0117] For example, multi-node data normalization and alignment can include unifying units and aligning orders of magnitude for latency, bandwidth, utilization, and queue length reported by different switches; and aligning timestamps based on a global clock (IEEE 802.1AS) to ensure the entire network is on the same time base. Network performance metric statistical analysis can include calculating the average latency, maximum latency, jitter, and packet loss rate for each flow and link; statistically analyzing the average, peak, and volatility of bandwidth utilization for each switch port / link; and analyzing the hop distribution of flow paths, bottleneck link distribution, and rerouting frequency. Anomaly and bottleneck identification can include identifying links with long-term high loads, periodically congested nodes, and frequently failing links; identifying flows with excessive latency, areas with severe bandwidth contention, and points where local resources are exhausted. Load balancing analysis can include calculating the variance, range, and balancing coefficient of the entire network link utilization; and identifying backbone links and hotspots with excessively concentrated traffic. Fault mode and reliability analysis can include statistically analyzing fault types, fault frequencies, and fault recovery times; and analyzing the frequency of redundant path activation and the actual occupancy of redundant resources. Resource utilization efficiency analysis can include statistics on TSN reserved bandwidth idle rate, queue idle time slot ratio, flow table / PSFP resource utilization rate; and assessment of resource reuse degree and resource waste.

[0118] The intelligent algorithm includes one or more of the following: Constraint Satisfaction Problem (CSP) solving algorithm, Bayesian network probabilistic inference algorithm, Fault Propagation Graph (FPG) inference algorithm, genetic algorithm, and particle swarm optimization algorithm.

[0119] The following examples illustrate single-objective and multi-objective global optimization.

[0120] Example 1: In an optional end-to-end latency minimization single-objective global optimization, the central controller can collect summary information reported by each switch, such as port queue latency, jitter, flow path hop count, link propagation latency, flow priority, bandwidth usage, congestion status, buffer usage, topology, and available bandwidth.

[0121] The variables defined in the constructed optimization problem include:

[0122] F: The set of all TSN service flows in the network;

[0123] f∈F: A certain TSN stream;

[0124] T f The end-to-end total delay of stream f;

[0125] L: The set of all links in the network;

[0126] (u,v)∈L: A link from node u to node v;

[0127] d f (u,v): The single-hop delay generated by flow f on link (u,v).

[0128] The end-to-end delay of flow f is the sum of the single-hop delays of all links it traverses:

[0129] Tf=∑ (u,v)∈Pf d f (u,v)

[0130] Where Pf is the forwarding path of flow f.

[0131] The objective function in the constructed optimization problem is:

[0132] Min J=∑ f∈F Tf

[0133] The objective function described above aims to minimize the total end-to-end latency of all TSN streams in the entire network.

[0134] The constraints in the constructed optimization problem include:

[0135] ① Single-hop delay upper limit constraint:

[0136] d f (u,v)≤D max , f, (u,v)∈Pf

[0137] Where D max This represents the upper limit of single-hop delay.

[0138] ② Hard constraints on flow priority, including: the upper limit of latency for high-priority flows is strictly less than that for low-priority flows:

[0139] T f,high ≤T f,low max

[0140] Where T f,high T is the actual end-to-end delay for high-priority streams. f,low max This is the upper limit threshold for the delay of low-priority streams.

[0141] ③ Link aggregation load does not exceed link capacity, and service bandwidth is non-negative:

[0142] ∑ f∈F(u,v) b f ≤B (u,v) , b f ≥0

[0143] Where bf is the bandwidth required by service flow f, F(u,v) is the set of service flows passing through link (u,v), and B (u,v) Let (u,v) be the capacity of the link.

[0144] ④ Scheduling timing non-conflict constraints, including: Within the same port at the same time, gating scheduling windows do not overlap.

[0145] G f1 ∩G f2 =

[0146] Among them G f1 G f2 These are the scheduling time windows for flows f1 and f2, respectively.

[0147] ⑤ Path hop count constraint (optional):

[0148] |P f |≤H max

[0149] Where H max This represents the maximum allowed number of hops.

[0150] The optimization problem described above can be solved by combining the CSP (Constant Shortest Path) algorithm with an improved shortest path algorithm. First, priority, bandwidth, hop count, and scheduling non-conflict are all converted into CSP variables and domain constraints. Then, each link is modeled as a delay-weighted edge, where the weight equals the propagation delay plus the estimated queuing delay. Finally, the delay-weighted Dijkstra's algorithm or an improved Bellman-Ford algorithm is used to calculate the minimum delay path for each flow.

[0151] The final control policy issued to each switch can be an abstract rule extracted from the global optimal solution that does not depend on the specific topology and does not contain local flow tables. For example: 1) The end-to-end latency of critical control flows is ≤2ms; 2) The queuing latency of priority 7 single hops is ≤200μs, and that of priority 6 is ≤500μs; 3) Queuing for more than 2 cycles is prohibited on congested ports; 4) Paths with fewer hops and higher link idle time are given priority.

[0152] In terms of specific distribution strategies, all switch entries can have the same structure, while edge switches can be subject to stricter thresholds.

[0153] Example 2: In a multi-objective global optimization that combines optional end-to-end latency minimization with network-wide load balancing, the central controller collects information reported by each switch. Based on Example 1, it can add link bandwidth utilization, port traffic share, peak traffic, device CPU / backplane bandwidth utilization, hot links and congestion warnings, flow routing distribution, etc.

[0154] The variable definitions in the constructed optimization problem also include:

[0155] util link Link utilization;

[0156] The objective function in the constructed optimization problem is:

[0157] Min (∑ f∈F Tf, max(util link )-min(util link ))

[0158] The multi-objective optimization model simultaneously minimizes the total end-to-end latency of the entire network flow and the link utilization error.

[0159] The constraints in the constructed optimization problem include:

[0160] ① The critical flow delay hard constraint remains unchanged;

[0161] ② The maximum single-link bandwidth utilization rate is ≤70%;

[0162] ③Prohibit three ports from being under high load simultaneously;

[0163] ④ Flow migration does not trigger frequent rerouting.

[0164] In this example, the optimization problem can be solved using a multi-objective genetic algorithm (such as NSGA-II) or a particle swarm optimization (PSO) algorithm combined with CSP constraints. The steps can be as follows: first, perform a global search using NSGA-II or PSO to obtain a set of Pareto optimal solutions; then, use a CSP constraint solver to filter out solutions that do not meet the requirements of priority, bandwidth, and scheduling conflicts; and finally, select the compromise optimal solution (acceptable latency and most balanced load) from the set of optimal solutions.

[0165] The generated control policies may include: 1) end-to-end latency of critical control flows ≤ 10ms; 2) link bandwidth utilization between 30% and 65%; 3) high-priority flows are prohibited from being mapped to the same backbone link; 4) when the utilization of a link exceeds 65%, the flow is automatically triggered to be diverted to the second-best but low-load path; 5) the backplane bandwidth utilization of a single device does not exceed 80%.

[0166] In terms of specific distribution strategies, a consistent entry framework can be adopted for all switches, but the link utilization threshold and traffic offloading trigger conditions will be distributed differently based on the switch's location in the topology. The utilization threshold is more stringent for backbone switches, while the traffic offloading conditions are more lenient for edge switches.

[0167] Through the above global optimization process, the central controller can discover and resolve potential global problems in distributed autonomous operation, and generate a new round of control strategies to calibrate and guide the TSN switches. This fully leverages the real-time advantages of the TSN switches' local autonomy, while ensuring that the overall behavior of the entire network evolves towards global optimization goals such as maximizing resource utilization and minimizing end-to-end latency.

[0168] The time-sensitive network switch device involved in the embodiments of the present invention will be further described below.

[0169] This invention provides a time-sensitive network switch device, such as... Figure 3 As shown, the system includes a control plane 100 and a forwarding plane 200. The control plane 100 receives control policies from the central controller 300, the control policies including a set of constraints, which may include one or more of performance constraints, resource constraints, and behavioral constraints. Based on the constraints in the control policies, it determines flow configuration parameters according to service requirements and the hardware and software capabilities of the switch. It then sends the flow configuration parameters to the forwarding plane 200 and reports summary information to the central controller 300. The summary information includes the processing results of TSN flow forwarding performed by the forwarding plane 200 and / or aggregated preset performance indicators obtained based on the processing results. The forwarding plane 200 generates and loads hardware entries related to Time-Sensitive Network (TSN) flows according to the flow configuration parameters sent by the control plane 100, performs TSN flow forwarding, and feeds back the TSN flow forwarding processing results to the control plane 100.

[0170] The beneficial effects of the switching device in this embodiment are consistent with those of the aforementioned control method for the switch control plane, and will not be repeated here. The specific working method of the device can be found in the descriptions of each step in the method. For example, the control plane 100 can be equipped with a configuration management module, used to determine flow configuration parameters based on constraints in the control policy, according to service requirements and the switch's hardware and software capabilities, to issue the flow configuration parameters to the forwarding plane 200, and to report summary information to the central controller 300. A reliability assurance module can also be further set up to detect link interruptions, port failures, or device anomalies, and to send a notification to the configuration management module in the event of an anomaly, so that the configuration management module can adjust the flow configuration parameters based on the constraints in the control policy. A device management module can also be further set up for network topology discovery and updating, initiating link status monitoring, synchronizing basic link information, and sending a notification to the configuration management module after discovering a new neighboring switch joining the network and completing link negotiation, so that the configuration management module can synchronize the control policy to the new neighboring switch. A distributed coordination module can also be further set up to negotiate with neighboring switches based on constraints in the control policy when cross-node coordination is required.

[0171] Compared to a centralized control model, the solution in this embodiment shifts control from a single point at the network center to every edge node of the network, providing TSN switch devices that integrate control and forwarding. Based on the TSN switch devices provided in this embodiment, the central controller 300 no longer needs to tell the TSN switch devices "to open the door to queue B at time A," but rather tells them "to ensure that this type of data flow arrives at its destination within time C, and takes priority over other data flows." Each TSN switch device gains autonomy, achieving a qualitative change in network resilience. When the network scales up, the new nodes are those with autonomous capabilities, rather than adding burden to the central controller.

[0172] When the control plane is moved from the central controller 300 to the TSN switch device, the configuration management, device management, protocol control, reliability assurance and other functions of the control plane focus on localization and real-time response, and can be deeply integrated with the forwarding plane.

[0173] The internal module architecture of the control plane 100 and the forwarding plane 200 is illustrated below.

[0174] In some possible implementations, such as Figure 4 As shown, the control plane 100 may include a configuration management module 110, a device management module 120, a protocol control module 130, a reliability assurance module 140, and a distributed collaboration module 150; the forwarding plane 200 may include an ingress port processing module 210, a cache and queue management module 220, a core scheduling mechanism module 230, and an egress port processing module 240.

[0175] The following is combined with Figure 5 The detailed settings and interaction flow of the functional modules of the control plane 100 are illustrated by example.

[0176] The configuration management module 110 may include a basic configuration submodule 111, a TSN protocol configuration submodule 112, an intelligent scheduling submodule 113, and a network configuration submodule 114.

[0177] The basic configuration submodule 111 is used to configure the switch's port parameters, such as port speed, working mode (full-duplex / half-duplex), and basic system parameters, such as global VLAN and static MAC.

[0178] TSN protocol configuration submodule 112 is used to configure basic TSN protocol parameters, including detailed configuration of time synchronization protocols (such as IEEE 1588) and parameter configuration of traffic scheduling protocols (such as Qbv, Qci, etc.).

[0179] The intelligent scheduling submodule 113 is used to determine flow configuration parameters based on constraints in the control policy, service requirements, and the hardware and software capabilities of the switch; and to automatically adjust the data transmission path and priority based on real-time network traffic monitoring data and preset scheduling policies to optimize network resource utilization and improve overall network performance. The preset scheduling policies include time-aware scheduling, traffic shaping scheduling, frame preemption (optimizing latency upper bounds), basic priority, and fairness.

[0180] The network configuration submodule 114 is used to configure and manage the network topology, including adding, deleting, and modifying network nodes and links, as well as setting network routing rules.

[0181] The device management module 120 may include a topology management submodule 121, a link management submodule 122, and a device management submodule 123.

[0182] The topology management submodule 121 is used to monitor the device connectivity in the network in real time, automatically discover newly added and disconnected devices, and update the network topology. Preferably, it can display the network topology graphically, allowing administrators to intuitively understand the network's operating status.

[0183] The link management submodule 122 is used to monitor and manage links in the network, including link status monitoring (such as link connectivity, signal strength, etc.), and link bandwidth allocation and management. When a link fails, it can promptly detect the failure and take corresponding recovery measures, such as automatically switching to a backup link. After detecting a new neighboring switch joining the network and completing link negotiation, it sends a neighboring switch joining notification to the configuration management module 110 so that the configuration management module 110 can synchronize control policies with the new neighboring switch.

[0184] The device management submodule 123 is used to manage the switch itself and other devices connected to the switch, including device registration, deregistration, status monitoring, firmware upgrades, etc.

[0185] Protocol control module 130 is used to manage TSN-compatible industrial bus / industrial Ethernet protocols, including Modbus-RTU, Modbus-TCP, Profinet, Ethercat, Ethernet / IP, RS232, RS485, CAN, etc.

[0186] The reliability assurance module 140 integrates protocols such as High-availability Seamless Redundancy (HSR), Parallel Redundancy Protocol (PRP), and IEEE 802.1 Qci. For example, PRP transmits the same data simultaneously through two independent links, allowing the receiving end to select either link to receive the correct data, thus improving data transmission reliability. IEEE 802.1 Qci employs filtering and control strategies for each data stream to ensure that input traffic conforms to specifications, thereby avoiding abnormal traffic issues caused by faults or malicious attacks and guaranteeing the bandwidth and latency requirements of critical services. When the reliability assurance module 140 detects a link interruption, port failure, or device malfunction, it sends a fault notification to the configuration management module 110, which then adjusts the flow configuration parameters.

[0187] The distributed coordination module 150 is used to negotiate with the control plane of adjacent TSN switches based on a distributed negotiation algorithm when cross-node coordination is required. This includes cross-node bandwidth allocation, path coordination, or priority mapping. The final negotiation result is then reported to the central controller 300 as part of the summary information.

[0188] Based on the above control plane functional module settings, such as Figure 5 Here is an example of an interactive flow:

[0189] A1: Topology management submodule 121 completes automatic network topology discovery and updates the topology structure; link management submodule 122 starts link status monitoring and synchronizes basic link information. Device management module 120 summarizes device, topology, and link status and sends it to configuration management module 110.

[0190] A2: The configuration management module 110 receives the control strategy issued by the central controller and hands it over to the intelligent scheduling submodule 113 to determine the flow configuration parameters.

[0191] A3: Configuration management module 110 sends the flow configuration parameters to forwarding plane 200.

[0192] A4: During normal operation, the topology management submodule 121 updates the topology in real time, the link management submodule 122 monitors the link status and bandwidth, and the device management submodule 123 continuously monitors the device status and synchronizes it to the configuration management module 110.

[0193] A5: Protocol control 130 continuously monitors the operating status of the industrial Ethernet protocol and promptly reports any abnormalities to the configuration management module 110.

[0194] A6: Reliability Assurance 140 continuously monitors the link and port status, performs traffic filtering control, and sends a fault notification to the configuration management module 110 when an anomaly occurs.

[0195] A7: The intelligent scheduling submodule 113 adaptively adjusts the transmission path, priority and scheduling parameters based on real-time traffic and load data, and sends them to the forwarding plane 200.

[0196] A8: When cross-node coordination is required, the distributed coordination module 150 negotiates with the control plane of the adjacent switch to complete operations such as bandwidth allocation and path coordination, and feeds back the negotiation results to the configuration management module 110.

[0197] A9: Based on the fault notification or negotiation results, the configuration management module 110, in conjunction with the intelligent scheduling submodule 113, adjusts the flow configuration parameters to ensure compliance with the control strategy constraints, and then reissues them to the forwarding plane 200.

[0198] A10: Configuration management module 110 summarizes various statuses, scheduling adjustments, fault handling, and negotiation results, and reports them as summary information to the central controller.

[0199] This embodiment further refines the internal functional module settings and interaction relationships of the control plane. The modules work together to realize the local autonomous closed loop of the TSN switch after receiving the control policy. This ensures both the real-time nature of local decision-making and that all autonomous behaviors always operate within the constraints of the global policy.

[0200] The following is combined with Figure 6 The detailed settings and interaction flow of the functional modules of the forwarding plane 200 are illustrated by example.

[0201] In the forwarding plane 200, the ingress port processing module 210 may include an ingress port physical layer interface submodule 211, a timestamp unit submodule 212, and a traffic classifier submodule 213.

[0202] The ingress physical layer interface submodule 211 is used for physical connection with external network devices to achieve data reception and transmission. It typically supports multiple transmission media such as Ethernet cable and fiber optic cable to meet the needs of different application scenarios.

[0203] The timestamp unit submodule 212 is used to add a precise timestamp to each data frame entering the switch, serving as the basis for time synchronization and traffic scheduling. The accuracy of the timestamp should meet the requirements of the TSN network, typically in the nanosecond range.

[0204] The traffic classifier submodule 213 is used to classify incoming traffic to the switch based on information such as the source address, destination address, port number, and protocol type of the data frame, and to allocate different types of traffic to different queues for subsequent processing and scheduling.

[0205] The cache and queue management module 220 may include a supervisor submodule 221, a filter submodule 222, and a queue mapping submodule 223.

[0206] The monitor submodule 221 is used to monitor traffic entering the switch. Based on preset traffic limit rules, it processes traffic exceeding the limit, such as dropping or marking it. This prevents illegal or excessive traffic from affecting the network and ensures its normal operation.

[0207] The filter submodule 222 is used to filter data frames according to preset filtering rules, allowing only data frames that meet the rules to pass through the switch. This can be used to prevent malicious attacks, virus propagation, etc., thereby improving network security.

[0208] The queue mapping submodule 223 maps the categorized traffic to the corresponding queues. Each queue can be configured differently according to its priority and business requirements.

[0209] The core scheduling mechanism module 230 may include a gating list submodule 231 and a time-aware shaper submodule 232.

[0210] The gating list submodule 231 is used to generate a gating list based on a time scheduling strategy to control the timing of data traffic transmission. The gating list can be precise down to each time period, determining which queues can send data and which queues need to wait within a certain time period.

[0211] The time-aware shaper submodule 232 shapes the data traffic based on the gating list and time synchronization information to ensure that the data is sent at predetermined time intervals and rates, thereby guaranteeing the determinism and real-time performance of the network.

[0212] The output port processing module 240 may include an output port physical layer interface submodule 241, a scheduler submodule 242, and a shaper submodule 243.

[0213] The output port physical layer interface submodule 241, which functions similarly to the input port physical layer interface submodule 211, is responsible for sending the processed data frames to external network devices through the physical link.

[0214] The scheduler submodule 242 is used to select appropriate data frames for transmission based on queue priorities and gating list instructions. It possesses efficient scheduling algorithms that ensure high-priority data frames are transmitted first while fully utilizing the bandwidth resources of the output link. These scheduling algorithms can include: Strict Priority (SP); Enhanced Transmission Selection (ETS), including weighted round-robin, deficit round-robin, and weighted fair queuing; Time-Aware Shaper (TAS); Traffic Shaping, including Credit-Based Shaper (CBS), Asynchronous Traffic Shaping (ATS), and Cyclic Queuing and Forwarding (CQF); and Enhanced Scheduling, including Frame Preemption and hybrid TAS+CBS / CQF scheduling.

[0215] The shaper submodule 243 is used to shape the data traffic to be transmitted to meet the bandwidth and rate requirements of the output link, thus avoiding data conflicts and congestion. Precise traffic control can be achieved by adjusting the transmission interval and sequence of data frames.

[0216] Based on the above functional module settings of the forwarding plane, such as Figure 6 As shown. An exemplary interaction flow is as follows:

[0217] B1: The control plane configuration management module 110 (intelligent scheduling submodule 113) issues configuration instructions to each module of the forwarding plane 200: issues timestamp accuracy and traffic classification rules to the ingress port processing module 210; issues traffic monitoring, filtering and queue mapping rules to the cache and queue management module 220; issues gating scheduling and time synchronization parameters to the core scheduling mechanism module 230; and issues scheduling algorithms and shaping parameters to the egress port processing module 240.

[0218] B2: The ingress port physical layer interface submodule 211 of the ingress port processing module 210 receives external data frames, the timestamp unit submodule 212 adds nanosecond-level timestamps to the data frames, and the traffic classifier submodule 213 completes traffic classification and marks priority according to the control plane rules, and then transmits the classified traffic (with priority) + nanosecond-level timestamps to the buffer and queue management module 220.

[0219] B3: The supervisor submodule 221 of the cache and queue management module 220 monitors traffic according to control plane rules and discards / marks excessive traffic; the filter submodule 222 filters illegal frames and only allows frames that meet the rules to pass; the queue mapping submodule 223 maps the classified traffic to the corresponding queues and transmits the status of each queue to the core scheduling mechanism module 230. If traffic exceeds the threshold or illegal frames are detected, the abnormality is reported to the control plane reliability assurance module 140 simultaneously.

[0220] B4: The gating list submodule 231 of the core scheduling mechanism module 230 generates a gating list based on the time scheduling strategy issued by the control plane and the queue status; the time-aware shaper submodule 232 performs preliminary shaping of the queue traffic based on the time synchronization information, and then transmits the gating permission instruction + sending timing / rate limit to the output port processing module 240.

[0221] B5: The scheduler submodule 242 of the outgoing port processing module 240 selects high-priority queue data frames for priority scheduling based on the scheduling algorithm specified by the control plane and in conjunction with the gating enable command; the shaper submodule 243 shapes the traffic to be sent to meet the requirements of the output link; the outgoing port physical layer interface submodule 241 sends the processed data frames to the external device.

[0222] B6: The scheduler submodule 242 of the outgoing port processing module 240 collects the outgoing port bandwidth utilization and link status in real time, feeds back the outgoing port bandwidth status to the core scheduling mechanism module 230, and simultaneously reports it to the control plane link management submodule 122.

[0223] B7: The core scheduling mechanism module 230 dynamically adjusts the gating list and transmission timing based on the bandwidth status fed back from the output port. If a timing deviation or bandwidth abnormality occurs, it reports to the intelligent scheduling submodule 113 of the control plane and requests adjustment of the scheduling parameters.

[0224] B8: The cache and queue management module 220 continuously feeds back queue occupancy and traffic monitoring results to the control plane. The intelligent scheduling submodule 113 of the control plane dynamically updates traffic classification, monitoring, queue mapping, and other rules based on the feedback and sends them to the corresponding modules in the forwarding plane to achieve adaptive optimization.

[0225] This embodiment further refines the internal functional module settings and interaction relationships of the forwarding plane. The modules form a closed loop through state feedback, so that the forwarding plane is not only a hardware path for passively executing instructions, but also has dynamic perception and adjustment capabilities. It can cooperate with the control plane to ultimately transform the control strategy into microsecond-level forwarding behavior.

[0226] The following section, using industrial field data as an example, illustrates the data transmission method of a time-sensitive network switch, based on the design of the control plane and forwarding plane functional modules of the aforementioned TSN switch. The data transmission method of the time-sensitive network switch specifically includes the following steps:

[0227] S31. Control Policy Distribution: Based on the overall network topology, service transmission requirements, and resource distribution status, the central controller generates control policies through centralized calculation and then distributes them to the relevant TSN switches.

[0228] S32, Local Policy Resolution and Autonomous Execution: (e.g.) Figure 7 As shown, the interaction process between the control plane 100 and the forwarding plane 200 of the TSN switch is as follows:

[0229] C1: The topology management submodule 121 of the control plane device management module 120 completes automatic network topology discovery and update, the link management submodule 122 starts link status monitoring and synchronizes basic link information, and the device management submodule 123 monitors device status and summarizes all status data to send to the configuration management module 110.

[0230] C2: The configuration management module 110 receives the control policy issued by the central controller 300 and hands it over to the intelligent scheduling submodule 113. Based on the constraints in the control policy, the module determines the flow configuration parameters according to the business requirements and the hardware and software capabilities of the switch.

[0231] C3: Configuration management module 110 sends configuration instructions to each module of the forwarding plane, such as: sending timestamp accuracy and traffic classification rules to ingress port processing module 210; sending traffic monitoring, filtering and queue mapping rules to cache and queue management module 220; sending gating scheduling and time synchronization parameters to core scheduling mechanism module 230; and sending scheduling algorithm and shaping parameters to egress port processing module 240.

[0232] C4: The ingress physical layer interface submodule 211 of the forwarding plane ingress port processing module 210 receives external data frames, the timestamp unit submodule 212 adds nanosecond-level timestamps to the data frames, and the traffic classifier submodule 213 completes traffic classification and marks priorities according to the control plane rules, and passes it to the buffer and queue management module 220.

[0233] C5: The supervisor submodule 221 of the forwarding plane buffer and queue management module 220 monitors traffic and handles excessive traffic, the filter submodule 222 filters illegal frames, the queue mapping submodule 223 maps traffic to the corresponding queue, feeds back the queue status to the core scheduling mechanism module 230, and reports to the control plane reliability assurance module 140 when there is an anomaly.

[0234] C6: The gating list submodule 231 of the forwarding plane core scheduling mechanism module 230 generates a gating list, the time-aware shaper submodule 232 shapes the traffic, and transmits the gating permission command and the sending timing and rate limit to the outgoing port processing module 240.

[0235] C7: The scheduler submodule 242 of the forwarding plane outgoing port processing module 240 schedules data frames according to the algorithm specified by the control plane, the shaper submodule 243 shapes the traffic, the outgoing port physical layer interface submodule 241 sends data to external devices, and at the same time feeds back the outgoing port bandwidth status to the core scheduling mechanism module 230, and synchronously reports to the control plane link management submodule 122.

[0236] C8: Normal operation of the control plane: The protocol control module 130 monitors the status of the industrial Ethernet protocol and reports to the configuration management module 110 when there is an abnormality; the intelligent scheduling submodule 113 adaptively adjusts the transmission path, priority and scheduling parameters based on the traffic and load data fed back by the forwarding plane and resends them to the forwarding plane.

[0237] C9: Special scenario handling: When cross-node coordination is required, the distributed collaboration module 150 negotiates with adjacent switches and feeds back the results to the configuration management module 110; when a fault occurs, the reliability assurance module 140 sends a fault notification to the configuration management module 110, and the configuration management module 110, in conjunction with the intelligent scheduling submodule 113, adjusts the flow configuration parameters and resends them to the forwarding plane.

[0238] C10: The configuration management module 110 summarizes the topology, links, device status, as well as scheduling adjustments, fault handling, and cross-node negotiation results, and reports them as summary information to the central controller 300 to support global optimization.

[0239] In real-world scenarios, each module in the control plane can summarize the operational results and indicator data of its respective domain as needed. After data aggregation, deduplication, classification, and statistical analysis, the data is reported as summary information.

[0240] For example, in the configuration management module 110, the indicators of the basic configuration submodule 111 may include the total number of global basic configurations, the number of effective configurations / the number of configurations to be issued; the configuration issuance success rate, the number of configuration issuance failures, and the number of configuration rollback triggers; the compliance verification pass rate of device basic parameters, such as the compliance of maximum transmission unit (MTU), VLAN, and port rate duplex; the configuration change frequency (number of changes per day / hour), the audit count of high-risk configuration operations; the number of configuration backups, the most recent backup time, and the number of backup file verification anomalies. The TSN protocol configuration submodule 112 indicators may include gPTP configuration: number of PTP domain configurations, number of clock port configuration compliances, and number of time synchronization parameter configuration anomalies; 802.1Qbv (TAS gating): total number of GCL flow table configurations, number of gating queue bindings, and number of timeslot configuration conflicts; 802.1Qav (CBS): number of credit shaping configuration flows and number of bandwidth quota overruns; 802.1Qch (CQF): number of period offset configuration anomalies and number of dual queue binding errors; 802.1Qbu / 802.3br preemption: number of preemption-enabled port configurations and preemption granularity configuration error counts; number of TSN full protocol configuration consistency verification failures and number of cross-device configuration synchronization differences. The intelligent scheduling submodule 113 indicators may include the number of intelligent scheduling policy templates and the number of service flows bound to policies; the number of times automatic routing scheduling takes effect and the number of times scheduling replanning is triggered; the number of times automatic timeslot allocation fails and the number of times bandwidth intelligent reservation is insufficient alarms; and the percentage of scheduling algorithms enabled (statistics on the number of online TAS / CBS / CQF policies). The network configuration submodule 114 metrics may include the total number of Virtual Local Area Network (VLAN) configurations, the number of Layer 3 interface configurations, the number of Access Control List (ACL) access control rules, the number of normal / abnormal QoS priority mapping configurations, the number of DSCP-PCP binding errors, the number of static route / multicast configuration entries, and the count of residual redundant configuration entries.

[0241] In the device management module 120, the device management sub-module 123 metrics may include the total number of TSN switches in the entire network, the number of online devices / offline devices / alarm devices; device firmware version compliance rate, the number of devices in firmware upgrade / upgrade failure; average / peak device CPU utilization, memory utilization, Flash storage utilization; device runtime, number of device restarts, and number of heartbeat keep-alive disconnections. The topology management sub-module 121 metrics may include the total number of topology nodes discovered in the entire network, topology convergence completion status, number of topology updates; number of topology loop detections, number of loop anomaly alarms, topology change trigger time; number of offline dumb nodes, topology edge node / core node health scores; and the number of service path topology consistency verification anomalies. The 122 metrics of the link management submodule can include the total number of physical links, the number of up links / down links, the number of link oscillation switching times; link bandwidth utilization, the number of link packet errors / packet loss / Cyclic Redundancy Check (CRC) errors; the number of TSN dedicated service links bound, the link latency baseline & real-time jitter value; the number of link negotiation anomalies, and the number of optical module power threshold over-limit alarms.

[0242] In the reliability assurance module 140, HSR / PRP redundancy protection indicators may include the number of online HSR ring network nodes and the ring network closed-loop integrity status; PRP dual-link working status (primary link / backup link occupancy count); HSR / PRP switching trigger count and redundancy failover time (switching latency); redundant path failure alarm count and duplicate frame drop statistics count; and the number of redundancy group configuration matching anomalies. 802.1QCI flow filtering and policing indicators may include the total number of QCI flow policing instances and the number of single-flow rate over-limit triggers; illegal frame interception and drop count and malicious attack flow cleaning count; frame length violation drop count and burst traffic suppression effectiveness count; whitelist flow passage count and blacklist interception hit count. Overall network reliability general indicators include the total number of network-wide fault alarms / cleared alarms / unprocessed emergency alarms; critical service flow protection failover success rate and uninterrupted service runtime; fault self-healing trigger count and self-healing recovery time statistics; and network-wide Service Level Agreement (SLA) availability indicators.

[0243] The metrics in the distributed collaboration module 150 can include the number of online cluster controller nodes, the number of master / slave role switching times; distributed data synchronization latency, the number of packet loss times in cross-node configuration synchronization; collaborative decision-making interaction message throughput, the number of heartbeat timeouts between nodes; the success rate of sharded task collaborative processing, and the number of distributed lock contention conflicts.

[0244] The reported information can adopt a standardized data format, containing only core results and aggregated indicator values, without transmitting the original detailed data, thus reducing the bandwidth consumption of the reporting. The reporting frequency can be dynamically adjusted according to business needs. During normal operation, reports are made periodically, and reports are made immediately after a major event occurs, ensuring that the controller obtains key information in a timely manner.

[0245] S33 Global Optimization and Policy Adjustment: After receiving summary information such as fault handling results, aggregated key performance indicators, and final results of distributed collaboration reported by all TSN switches across the network, the central controller summarizes and integrates the network data. Combined with preset global optimization goals, it uses data analysis and intelligent algorithms to deeply uncover global problems hidden in local autonomous operation. Based on real-time network topology, resource occupancy status, service priority distribution, and the latest service demands, the central controller readjusts global network performance goals and constraints, generating a new round of global policies adapted to the current network state. Subsequently, the updated global policies are distributed to relevant TSN switches, calibrating and scientifically guiding distributed autonomous behavior, ensuring the local operation and autonomy of each switch, and collaboratively serving the overall network optimization goals, forming a closed-loop control mechanism of distribution, execution, feedback, and optimization.

[0246] In practical engineering, regarding hardware deployment, the control plane of the TSN switch can be implemented using a high-performance FPGA chip, integrating hardware logic circuits for configuration management, device management, protocol control, and reliability assurance to ensure rapid processing of control commands. The forwarding plane can use an ASIC chip to implement ingress port processing, buffer queue management, core scheduling, and egress port processing. The switching matrix can adopt a Crossbar architecture, supporting simultaneous line-speed forwarding of 16 ports at a port rate of 10Gbps. The control plane and forwarding plane can interact via the AXI4-Stream high-speed interface, with an interface bandwidth of no less than 40Gbps, ensuring efficient transmission of control commands and data frames.

[0247] In terms of software and protocol configuration, the control plane of the TSN switch loads an embedded operating system based on FreeRTOS and runs the TSN protocol stack (supporting IEEE 802.1Qbv time-sensitive scheduling, IEEE 802.1AS clock synchronization, and IEEE 802.1Qcc flow configuration protocol). The configuration management module of the control plane supports receiving remote configuration commands or local manual configuration via the NETCONF protocol, and the flow configuration parameters can be dynamically adjusted. The reliability assurance module integrates the Bidirectional Forwarding Detection (BFD) protocol, and the fault detection time is less than 1ms.

[0248] like Figure 8As shown, this embodiment of the invention also provides an electronic device, including a memory 400, a processor 500, and a computer program / instructions stored in the memory 400 and executable on the processor 500. When executed by the processor 500, the computer program / instructions implement the steps of the control method for the time-sensitive network switch in any of the above embodiments. This electronic device has the same beneficial effects as the control method described above, and will not be repeated here.

[0249] like Figure 9 As shown in the illustration, this embodiment of the invention also provides a control system for a Time-Sensitive Network Switch (TSN), including a central controller 600 and a TSN switch control plane 700. The TSN switch control plane 700 is used to implement the control methods for the switch control plane in any of the aforementioned embodiments, and the central controller 600 is used to implement the control methods for the central controller in any of the aforementioned embodiments. This embodiment distributes complex, real-time-critical scheduling tasks across various TSN switch control planes, allowing the central controller to focus on non-real-time, global optimization tasks, thus improving the overall performance ceiling and engineering practicality of the TSN network. The provided control system possesses both a centralized, global perspective and distributed, autonomous, real-time response capabilities, making it better suited for applications such as industrial control, autonomous driving, and aerospace, which have extremely high requirements for network latency, jitter, and determinism.

[0250] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems or apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and relevant parts can be referred to the method section.

[0251] It should also be noted that relational terms such as "first" and "second" in this specification are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The term "and / or" includes any and all combinations of one or more of the associated listed items. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0252] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A control method for a time-sensitive network (TSN) switch, applied to the control plane of a TSN switch, characterized in that, include: The system receives control strategies issued by the central controller. The control strategies include a set of constraints, which includes one or more of performance constraints, resource constraints, and behavioral constraints. Based on the constraints in the control strategy, the flow configuration parameters are determined according to business requirements and the hardware and software capabilities of the switch. The flow configuration parameters are sent to the forwarding plane of the TSN switch to instruct the forwarding plane to generate and load TSN flow-related hardware entries and perform TSN flow forwarding. The summary information is reported to the central controller. The summary information includes the processing result of the forwarding plane performing TSN flow forwarding and / or the aggregated preset performance index obtained based on the processing result.

2. The control method for a time-sensitive network switch according to claim 1, characterized in that, The determination of flow configuration parameters based on constraints in the control strategy, according to service requirements and the hardware and software capabilities of the switch, includes: Extract the performance constraints, resource constraints, and / or behavioral constraints contained in the control strategy; Based on the extracted constraints, with business requirements as the target and the switch's hardware and software capabilities as the physical feasible domain, the executable flow configuration parameters of the switch's forwarding plane are determined. The flow configuration parameters include one or more of the following: flow identification rules, scheduling gating list, bandwidth threshold, queue mapping, priority, flow supervision parameters, and redundant path table.

3. The control method for a time-sensitive network switch according to claim 1, characterized in that, Also includes: In the event of a detected link interruption, port failure, or device malfunction, the flow configuration parameters are adjusted based on the constraints in the control policy. The adjustment of the flow configuration parameters enables path switching, bandwidth reallocation, and / or gating scheduling adjustments. The adjustment results are reported to the central controller as part of the summary information.

4. The control method for a time-sensitive network switch according to claim 1, characterized in that, Also includes: After detecting that a new neighboring switch has joined the network and completed link negotiation, the control policy is synchronized to the new neighboring switch; the summary information includes the switch's network entry result.

5. The control method for a time-sensitive network switch according to claim 1, characterized in that, Also includes: When a switch needs to coordinate across nodes, it negotiates with neighboring switches using a distributed negotiation algorithm based on constraints in the control policy. The negotiation includes bandwidth allocation, path coordination, or priority mapping. The negotiation results are then reported to the central controller as part of the summary information.

6. A control method for a time-sensitive network switch, applied to a central controller, characterized in that, include: A control policy is issued to the control plane of a Time-Sensitive Network (TSN) switch. The control policy includes a set of constraints, which includes one or more of performance constraints, resource constraints, and behavioral constraints. The control policy is used to constrain the switch control plane to determine flow configuration parameters. The flow configuration parameters are determined based on the constraints in the control policy, according to service requirements and the switch's hardware and software capabilities. They are used to instruct the switch forwarding plane to generate and load TSN flow-related hardware entries and execute TSN flow forwarding. The receiver receives summary information reported by the switch control plane. The summary information includes the processing results of the forwarding plane performing TSN flow forwarding and / or the aggregated preset performance indicators obtained based on the processing results.

7. The control method for a time-sensitive network switch according to claim 6, characterized in that, Also includes: The control strategy is updated by performing global optimization based on the summary information.

8. The control method for a time-sensitive network switch according to claim 7, characterized in that, The global optimization includes: The central controller summarizes and analyzes the summary information reported by all switches in the network. The data analysis includes one or more of the following: multi-node data normalization and alignment, statistical analysis of network performance indicators, anomaly and bottleneck identification, load balancing analysis, failure mode and reliability analysis, and resource utilization efficiency analysis. Based on the data analysis results and combined with the preset global optimization objectives, a new round of control strategies adapted to the current state of the entire network are generated through intelligent algorithms. The global optimization objectives consider one or more of the following: network resource utilization, end-to-end latency, network load balancing, and redundancy cost. The intelligent algorithms include one or more of the following: constraint satisfaction problem solving algorithms, Bayesian network probabilistic reasoning algorithms, fault propagation graph deduction algorithms, genetic algorithms, and particle swarm optimization algorithms.

9. A time-sensitive network switch device, characterized in that, Includes the control plane and the forwarding plane; The control plane is used to receive control policies issued by the central controller. The control policies include a set of constraints, which includes one or more of performance constraints, resource constraints, and behavioral constraints. Based on the constraints in the control policies, the control plane determines flow configuration parameters according to service requirements and the hardware and software capabilities of the switch. The control plane issues the flow configuration parameters to the forwarding plane. The control plane also reports summary information to the central controller. The summary information includes the processing results of the forwarding plane performing TSN flow forwarding and / or the aggregated preset performance indicators obtained based on the processing results. The forwarding plane is used to generate and load hardware entries related to Time Sensitive Network (TSN) flows based on the flow configuration parameters issued by the control plane, and to perform TSN flow forwarding. And the processing results of TSN stream forwarding are fed back to the control plane.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the control method for a time-sensitive network switch according to any one of claims 1-8.

11. A control system for a time-sensitive network switch, characterized in that, The system includes a central controller and a Time-Sensitive Network (TSN) switch control plane, wherein the switch control plane is used to implement the control method according to any one of claims 1 to 5, and the central controller is used to implement the control method according to any one of claims 6 to 8.

Images