Shared link-aware redundant routing and CQF flow admission optimization methods and systems
By constructing a directed graph of time-sensitive networks and a resource-aware CQF decoder, the order of stream processing and path configuration are optimized, which solves the problems of inaccurate reliability assessment and resource conflict of shared links in sparse industrial topologies, and improves the reliability and resource utilization efficiency of time-sensitive networks.
Patent Information
- Application Number
- CN202611618159.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-20
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies suffer from inaccurate reliability assessments of shared links in time-sensitive networks with sparse industrial topologies or physical bottleneck links. Redundant replicas are prone to concentrated conflicts on link CQF periodic resources. Redundant path selection, flow admission, and periodic scheduling are disconnected, resulting in low reliability and resource utilization efficiency.
By constructing a directed graph of a time-sensitive network, a set of candidate redundant paths is generated. A resource-aware CQF decoder is used to evaluate the reliability and resource requirements of the path set, optimize the stream processing order and CQF start cycle, and use the HORRA algorithm to adjust the stream rearrangement and path configuration to ensure resource utilization efficiency and reliability.
It improves the accuracy of reliability determination of redundant path sets, reduces resource conflicts, optimizes reliability and resource scheduling in network transmission, and enhances network reliability and capacity utilization.
Smart Images

Figure CN123123248A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of time-sensitive networking and deterministic communication technology, specifically to a shared link-aware redundant routing and CQF flow admission optimization method and system. Background Technology
[0002] Time-Sensitive Networking (TSN) provides bounded-delay transmission for services such as industrial control, process automation, and unmanned systems through time synchronization, traffic shaping, resource reservation, and explicit path control. Cyclic Queuing and Forwarding (CQF) divides the switch egress queue into synchronization periods and alternates between receiving and transmitting in adjacent periods, allowing data frames to advance hop-by-hop according to the cycle when the single-hop forwarding condition is met.
[0003] To improve the fault tolerance of critical services, the Frame Replication and Elimination for Reliability (FRER) mechanism generates multiple frame replicas with the same sequence identifier at the replication end, transmits them via different paths, and removes duplicate replicas at the elimination end. When the replica paths are completely disjoint, the failure of a single link typically only affects one replica; however, in sparse industrial topologies or networks with physical bottleneck links, completely disjoint paths may not exist, and multiple replica paths may therefore share one or more directed links.
[0004] Existing methods face the following problems when dealing with the above scenarios: If the failure events of each path are regarded as independent, the failure probability of the shared link will be repeatedly counted in the failure probability of multiple paths, resulting in an overestimation of the reliability of the path set, and may misjudge the path set that has not actually reached the reliability threshold as feasible; When multiple replicas pass through the shared link, they occupy the resources of the same directed link in the same CQF cycle. If redundant routing, flow admission and cycle allocation are separated, concentrated conflicts are likely to form on a few combinations of directed links and CQF cycles, thereby reducing the amount of admissible services. Summary of the Invention
[0005] To address the shortcomings of existing technologies, the present invention aims to provide a shared link-aware redundant routing and CQF flow admission optimization method and system. This invention solves the technical problems of inaccurate reliability assessment under conditions of non-fully disjoint redundant paths, concentrated conflicts of redundant replicas on link CQF periodic resources, and the disconnect between redundant path selection, flow admission, and periodic scheduling.
[0006] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a shared link-aware redundant routing and CQF flow admission optimization method is provided, including: Construct a directed graph of a time-sensitive network to obtain the flow to be scheduled; Generate candidate paths for each flow to be scheduled and form multiple sets of candidate redundant paths; determine the reliability of the path set of the candidate redundant path set and form a fixed candidate path set library; Using the stream processing order, candidate redundant path set, and CQF start period as search decisions, a resource-aware CQF decoder is constructed. The byte requirements of each stream instance and each replica are accumulated according to the link period. Admission variables are generated based on path set reliability, CQF conservative latency, over-cycle boundary, and link period capacity constraints. Candidate schemes are evaluated according to the lexicographical order rule of weighted admission value priority and resource target sub-optimization. Within the preset decoding evaluation budget, using the fixed candidate path set library as the search space, the stream processing order, candidate redundant path set, and CQF start period are jointly adjusted, and the feasibility is determined and resources are evaluated through the resource-aware CQF decoder; the high-weight unadmitted streams blocked by capacity are adjusted by guiding the flow rearrangement through admission gain; when the weighted admission value stagnates continuously, the blocked capacity is released and the affected stream is re-inserted by guiding the destruction and reconstruction through the link cycle bottleneck, and the current optimal executable scheme is determined according to the lexicographical order rule; Based on the current optimal executable scheme, the redundant path configuration and CQF periodic configuration corresponding to the admitted flow are distributed to the corresponding end systems and switches.
[0007] Secondly, a shared link-aware redundant routing and CQF flow admission optimization system is provided, including: The data acquisition module is configured to: construct a directed graph of a time-sensitive network to acquire the flow to be scheduled; The fixed candidate library construction module is configured to: generate candidate paths for each flow to be scheduled and form multiple sets of candidate redundant paths; determine the reliability of the path set of the candidate redundant path set and form a fixed candidate path set library; The resource-aware CQF decoder construction module is configured to: construct a resource-aware CQF decoder based on the stream processing order, the set of candidate redundant paths, and the CQF start period as search decisions; accumulate the byte requirements of each stream instance and each replica according to the link period; generate admission variables based on the reliability of the path set, the conservative CQF latency, the over-period boundary, and the link period capacity constraints; and evaluate candidate schemes according to the lexicographical order rule of weighted admission value first and resource target second best. The joint optimization and optimal solution determination module is configured to: within a preset decoding evaluation budget, using the fixed candidate path set library as the search space, jointly adjust the stream processing order, the candidate redundant path set, and the CQF start period, and perform feasibility determination and resource evaluation through the resource-aware CQF decoder; guide the flow rearrangement through admission gain to adjust the high-weight unadmitted flows blocked by capacity; when the weighted admission value stagnates continuously, guide the destruction and reconstruction through link cycle bottleneck to release the blocking capacity and re-insert the affected flow; and determine the current optimal executable solution according to the lexicographical order rules. The configuration distribution module is configured to: distribute the redundant path configuration and CQF periodic configuration corresponding to the admitted flow to the corresponding end system and switch according to the current optimal executable scheme.
[0008] Thirdly, an electronic device is also provided, comprising: Memory is used to store computer-readable instructions in a non-transitory manner. Processor, for executing the computer-readable instructions, When the computer-readable instructions are executed by the processor, they perform the method described in the first aspect above.
[0009] Fourthly, a computer-readable storage medium is provided having a program stored thereon that, when executed by a processor, implements the method described in the first aspect above.
[0010] Fifthly, a time-sensitive network is provided, including: A time-sensitive network controller, at least one replication end system, multiple CQF-enabled switches, and at least one elimination end system; The time-sensitive network controller is configured to execute the method described in the first aspect above and distribute the currently optimal executable scheme to the replication end system, multiple switches and the elimination end system respectively.
[0011] The above technical solution has the following advantages or beneficial effects: (1) The present invention calculates the joint availability probability based on the union of directed links of path subsets, avoiding the repeated inclusion of shared links in multiple independent path failure events, thereby improving the accuracy of reliability determination of redundant path sets.
[0012] (2) In this invention, the transmission load of each replica is accumulated in a unified link CQF periodic capacity model, so that the reliability coupling and resource coupling caused by the shared link are jointly included in the configuration decision.
[0013] (3) The present invention ensures that candidates in different neighborhoods are evaluated using the same reliability, latency, capacity and overcycle constraints by using a fixed candidate library and a unified decoder, thereby reducing unexecutable schemes caused by local proxy targets.
[0014] (4) The present invention uses a single-stage lexicographical order acceptance rule to ensure that the obtained weighted admission value does not decrease due to resource optimization, and continuously reduces peak link CQF periodic congestion or duplicate transmission resource consumption when the admission value remains unchanged.
[0015] (5) The present invention can output the path and period configuration that can be directly mapped to the FRER replication / elimination function and the CQF alternation queue. The technical solution is applied to the network transmission process and improves the reliability judgment and internal communication resource scheduling. Attached Figure Description
[0016] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.
[0017] Figure 1 This is a flowchart of the shared link-aware redundant routing and CQF flow admission optimization method in a specific embodiment of the present invention; Figure 2 This is a flowchart of a shared link-aware redundant routing and CQF flow admission optimization system in a specific embodiment of the present invention. Detailed Implementation
[0018] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0019] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments of the invention. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0020] In this embodiment of the invention, "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the description of this invention, "multiple" refers to two or more.
[0021] Furthermore, to facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.
[0022] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.
[0023] All data acquisition in this embodiment is carried out in compliance with laws and regulations and with user consent, and the data is used legally.
[0024] Example 1 like Figure 1 As shown, this embodiment provides a shared link-aware redundant routing and CQF flow admission optimization method, including: S101: Construct a directed graph of a time-sensitive network to obtain the flow to be scheduled; S102: Generate candidate paths for each flow to be scheduled and form multiple sets of candidate redundant paths; determine the reliability of the path set of the candidate redundant path set and form a fixed candidate path set library; S103: Using the stream processing order, candidate redundant path set, and CQF start period as search decisions, construct a resource-aware CQF decoder, accumulate the byte requirements of each stream instance and each replica according to the link period, generate admission variables based on path set reliability, CQF conservative latency, over-cycle boundary, and link period capacity constraints, and evaluate candidate schemes according to the lexicographical order rule of weighted admission value priority and resource target sub-optimal; S104: Within the preset decoding evaluation budget, using the fixed candidate path set library as the search space, jointly adjust the stream processing order, candidate redundant path set, and CQF start period, and perform feasibility determination and resource evaluation through the resource-aware CQF decoder; adjust the high-weight unadmitted streams blocked by capacity by guiding flow rearrangement through admission gain; when the weighted admission value stagnates continuously, destroy and reconstruct the blocked capacity by guiding the link cycle bottleneck to release the blocked capacity and re-insert the affected stream, and determine the current optimal executable scheme according to the lexicographical order rule; S105: Based on the current optimal executable scheme, the redundant path configuration and CQF periodic configuration corresponding to the admitted flow are sent to the corresponding end systems and switches.
[0025] Step S101 is as follows: Representing Time-Sensitive Networks as Directed Graphs , where the set of nodes Includes end system collection and switch set , It is a set of directed links. Each directed link With transmission rate and link availability ,in, Link availability represents the probability that a link remains operational during the configuration period under consideration. This embodiment assumes that the availability events of different directed links are independent of each other and remain stable within the time interval under consideration, but there are related availability events between different paths that share a common directed link.
[0026] Switches in time-sensitive networks use a common CQF cycle. For directed links and CQF periodic index ,by This represents the available byte capacity allocated to time-triggered services, which is no greater than... Link rate CQF cycle in bits per second In seconds. This available capacity can be pre-calculated by deducting the capacity used by other service categories, protocol overhead, synchronization errors, and protection bands.
[0027] The set of flows to be scheduled is denoted as ,in This represents the number of flows to be scheduled. At least including the source end destination end Frame length Business cycle Deadline Redundancy level Minimum path set reliability and admission weight ,in In bytes, and In terms of time, The required number of replica paths. (Each flow's business cycle) for Integer multiples, supercycle Take the least common multiple of all scheduled flow service cycles, and the number of CQF cycles within the excess cycle is: Flow to be scheduled The number of instances within a supercycle is The cycle of periodic business is .
[0028] Step S102 is as follows: For each flow to be scheduled The time-sensitive network controller generates no more than the maximum number of candidate paths through a weighted shortest simple path process. The candidate paths are defined, and each candidate path is an acyclic directed path. The upper limit of the number of candidate paths is a pre-configured positive integer, which in principle satisfies... The budget setting is calculated based on the topology size, redundancy level, and candidate combinations; when the number of actually available acyclic paths is insufficient, all available paths are used. If the number of available paths is less than the redundancy level, a candidate set that satisfies that redundancy level cannot be formed.
[0029] Directed Link The reliability-aware link cost is set as follows: (1) in, Directed link The basic link cost; when using hop count-based path cost, the basic link cost for each directed link can be set to 1; For link availability, This is the reliability preference coefficient. This is a weighted link cost that combines basic link cost and reliability. To improve the spatial diversity of candidate paths, in the already selected... After selecting candidate paths, proceed with selecting subsequent candidate paths. Sort using the following formula: (2) in, Directed link Before the election The number of times it appears in the candidate paths. This is the penalty coefficient for directed link multiplexing. Representing a path The set of directed links included. For the first The path sorting cost during the second selection. Equation (2) is only used for candidate path generation and sorting, and does not replace the reliability and feasibility determination described in the subsequent equation (4).
[0030] After sorting A flow is formed by 10 distinct acyclic candidate paths. Path pool ,in According to redundancy level Each time, select exactly one path from the path pool. These distinct candidate paths form a set of candidate redundant paths. Specifically, for paths that satisfy... Each combination of indexes constitutes ,in This is a composite index; combinations containing the same path but arranged in different orders are considered the same set. If there is no limit to the number of combinations, a maximum of [number] can be generated. A set of different candidate redundant paths.
[0031] For any single candidate path Under the condition that the availability events of different directed links are independent of each other, the path availability is: (3) FRER successfully delivers a data frame if at least one replica path is available. This applies to any set of candidate redundant paths. Each non-empty path subset Determine its directed link union The reliability of the candidate redundant path set is calculated based on the inclusion-exclusion principle: (4) In equation (4), Indicates the number of paths contained in a non-empty path subset; for any federated path available event, the same shared directed link simultaneously located on multiple replica paths... Only one instance is counted. Therefore, under the assumption that the available events of each directed link are independent and stable, equation (4) gives the exact reliability of the path set considering the correlation of shared links. For lower redundancy levels, this reliability can be calculated and cached before the optimization search begins.
[0032] To limit computational complexity, the candidate redundant path set can first be deterministically pre-sorted based on path-independent approximate reliability, the number of shared directed link reuses, the total number of hops, and the combination index. When the number of combinations exceeds a preset upper limit for accurate evaluation... At that time, only those ranked first... The exact reliability calculation of equation (4) is performed on the candidate redundant path set; the remaining combinations are not included in the fixed candidate pool. The path-independent approximation is only used as a pre-sorting basis and is not used to determine whether the minimum path set reliability is met. .
[0033] During the candidate library construction phase, for each candidate redundant path set Calculate the reliability of the path set and check the CQF conservative latency for each path. Only when And all satisfy Only when this condition is met will the set be considered a feasible candidate, in which Indicates the first The number of directed links contained in each replica path; an additional CQF period is used to cover the worst-case period alignment wait time between the service frame release time and the first sendable CQF period. Perform dominant filtering on each feasible candidate: Let Represents a directed link In the set The number of times it appears in different replica paths defines the number of times the shared directed link is reused. Define the cost of supercycle replication transport estimation. For any two feasible sets and If both conditions are met , and And at least one of them is a strict inequality, then by Dominate Delete the latter. Sort the remaining unused candidate sets in descending order of path set reliability, ascending order of shared directed link reuse count, ascending order of estimated cost of ultra-cycle replication transmission, and ascending order of composite index, retaining at most the top [number missing]. A set, forming a fixed candidate path set library. ;in This sets the upper limit for the capacity of the fixed candidate path set library.
[0034] Step S103 is as follows: For all A stream to be scheduled, the complete scheduling configuration scheme is denoted as ,in Indicates the processing order of the streams to be scheduled; Select a vector for the set of candidate redundant paths. For flow In its fixed candidate path set library The index of the selected candidate set; The CQF initial period vector for the first instance of each flow; Let be the vector of admission variables, where The search decision includes the stream processing order, the set of candidate redundant paths, and the CQF start period; the admission variables are calculated by the resource-aware CQF decoder based on reliability, latency, overcycle boundaries, and capacity constraints, rather than being independently specified. For unadmitted streams, their path selection and start period values are not involved in the actual device configuration.
[0035] For the flow to be scheduled Candidate redundant path set and the first instance start period Its link cycle The periodic byte requirement is defined as follows: (5) Equation (5) sums up each flow instance, each replica, and hop-by-hop transmission within the supercycle, where For instance indexing, For the replica path index, The path skips to the next sequence number. For the candidate set The first copy path A directed link, This is an indicator function; it takes the value 1 if the condition is true, and 0 otherwise. When multiple replicas travel through the same directed link and are in the same CQF cycle, their byte requirements are accumulated separately.
[0036] The resource-aware CQF decoder first initializes the link-cycle load matrix. ,in Indicates link periodicity pairs The cumulative byte payload, initially set to zero. Processed in stream order. Process the stream sequentially ,enumerate and the allowed CQF start period And perform the following constraint checks: First, the candidate set satisfies the path set reliability constraint. Furthermore, each replica path satisfies the CQF conservative latency constraint. ; Second, define the maximum number of replica path hops in the candidate set. And check the supercycle boundary. This ensures that the last hop sending period of all instances is within the time limit under consideration; Third, for all link cycles within the supercycle period Check capacity constraints If at least one feasible insertion exists, then let... When multiple feasible insertions exist, select the CQF resource target that results in the inserted value. Find the minimum set of candidate redundant paths and their starting period, and add their requirements to... If no feasible insertion exists, then set The link cycle load matrix is not updated.
[0037] For the complete solution The weighted admission value is defined as follows: Link cycle utilization is defined as... Its peak value is Link cycles with zero available capacity are not allowed to schedule traffic and are not included in the utilization division calculation. The total replication transmission resource consumption within the supercycle period is... Normalized resource consumption is The denominator is positive and calculated based on all link cycles within the supercycle. The resource target is defined as: (6) For any two executable schemes and ,like If the former is superior to the latter; if and Therefore, the former is superior to the latter. In equation (6) This is a tradeoff between peak link cycle utilization and normalized replication transmission resource consumption. This lexicographical order rule ensures that resource optimization does not come at the expense of reducing the already obtained weighted admission value.
[0038] Step S104 is as follows: This embodiment employs the Hierarchical Optimization for Reliable Routing and Admission (HORRA) framework. HORRA includes Admission-Gain-Guided Flow Reordering (AG-FR) and Link-CycleBottleneck-Guided Destroy-and-Repair (LCB-DR). AG-FR is the Admission-Gain-Guided Flow Ordering Neighborhood operator, used to prioritize the reordering of high-weight flows that have not been admitted due to capacity congestion. LCB-DR is a repair operator triggered when admission values remain stagnant, used to release capacity and reconstruct affected schemes around the congested link cycle. The candidate schemes generated by both are verified by the same resource-aware CQF decoder and their acceptance is determined according to the same lexicographical order rules.
[0039] HORRA uses a fixed candidate path set library formed by S102. For the search space, the resource-aware CQF decoder of S103 is used as a unified feasibility assessment and resource evaluation tool. Within the same decoding evaluation budget, the stream processing order, the set of candidate redundant paths, and the CQF start period are jointly adjusted. HORRA outputs the optimal executable solution obtained by evaluation in lexicographical order. .
[0040] Time-Sensitive Network Controllers based on resource efficiency metrics Generate the initial flow order; where Estimate the cost of the hypercycle replication transport defined in S102. To prevent the use of preset constants where the denominator is zero, short deadlines and short business cycles can be used as deterministic criteria for ranking in parallel.
[0041] In total decoding assessment budget Within each round, a maximum number of candidate batches are generated from the following three neighborhoods corresponding to the decision dimension, and these batches are mutually exclusive. The maximum number of evaluations allowed for calling the unified decoder: Flow order neighborhood: includes swapping the positions of two flows, moving the unadmitted flow forward, moving the low-value bottleneck contribution flow backward, and AG-FR; The redundant path set neighborhood includes candidate set replacement, set replacement that avoids the current congested link, and set replacement that reduces the number of times the congested link replicas are reused; CQF periodic allocation neighborhoods include single-flow period offsets, offsets that avoid peak link period pairs, and joint offsets performed on two or three conflicting flows.
[0042] Candidate solutions retain the decoded prefix prior to the earliest modified flow, and only re-decode the affected flow and its suffix. The path set or starting period specified by the neighborhood is enforced at the target flow; if this enforcement decision is not feasible, the candidate is discarded. All candidates are evaluated using the same resource-aware CQF decoder and lexicographical order rules in S103.
[0043] AG-FR is the admission-oriented flow ordering neighborhood operator in HORRA. Its inputs include the currently decoded scheme, the fixed candidate path set of the unadmitted target flow, the candidate start period, and the current link period load matrix. Its output is a new flow processing order that keeps the relative order of other flows unchanged, and the affected suffixes are re-decoded by the resource-aware CQF decoder.
[0044] For unapproved target flows For its candidate path set and CQF start period Calculate the link cycle capacity gap: (7) in This indicates taking the non-negative part. This represents the total capacity gap. Based on the total capacity gap, the maximum single link cycle capacity gap, the number of congested contributing flows, the number of congested link cycle pairs, the replication transmission resource requirements, and the path set and cycle index, the path with the fewest conflicts is selected for placement. .
[0045] The link cycle pairs with positive capacity gaps constitute the set of blocked link cycle pairs. For in Admitted flows with positive load contribution Define its load contribution ,by This represents the available payload on the blocked set that can alleviate the target flow capacity gap, and its release capacity per unit admission weight is calculated. Among them, the admission weight AG-FR prioritizes target flows with high admission weights and small total capacity gaps, moving them to one or more... Before the major blocking contribution stream, the relative order of other streams remains unchanged, and the affected suffixes are re-decoded. Stream rearrangement is accepted only if the re-decoded complete scheme is strictly better lexicographically than the current scheme.
[0046] LCB-DR is the stagnation-triggered repair operator in HORRA. Its inputs include the currently decoded scheme, the admission stagnation count, the minimum collision placement determined by AG-FR, and the corresponding set of blocked link period pairs. Its output is the complete repair scheme after first inserting the unadmitted target flow and then re-inserting the destroyed flow. The current scheme is only replaced if the repair scheme passes the same decoder verification and satisfies strict lexicographical improvement.
[0047] When continuous LCB-DR is triggered when the weighted admission value is not increased in the round, where The preset threshold for the number of rounds of stagnation for admission is used. Target flows that are not admitted due to capacity congestion are selected from a set of feasible candidate flows with reliable performance and conservative CQF latency. Minimum collision placement calculated using AG-FR and blocking link periodicity set For each Define the capacity to be released. .
[0048] Candidate contribution stream set This includes all admitted flows that have a positive load contribution on at least one blocking link cycle pair. For the destroyed flow set... Define temporary admission weight loss The enumeration size is 1 to... The candidate destruction set, in which This sets the preset maximum number of destroyed flows. Prioritizes all flows. All meet The complete coverage destruction set is selected; among such destruction sets, the selection is based on the following order: smaller temporary admission weight loss, larger effective capacity release, fewer destroyed flows, larger bottleneck contribution, and deterministic flow identifier. If no complete coverage set exists, the destruction set with higher capacity gap coverage is prioritized, and then the above sorting order is applied.
[0049] During reconstruction, first attempt to insert the target stream. Then, flows from the destruction set are re-inserted according to a deterministic order of decreasing admission weights, increasing estimated resource requirements, and flow identifiers. The reconstruction process uses only a fixed set of candidate paths and allowed CQF start cycles. A complete remediation plan is generated and checked by the same resource-aware CQF decoder, and is only considered if its weighted admission value increases, or if the weighted admission value remains unchanged and the resource objective is met. The current solution will only be replaced when the threshold is lowered.
[0050] Step S105 is as follows: After optimization, the Time-Sensitive Network Controller (TSN) determines the optimal executable solution based on the current lexicographical order. , to satisfy The flow to be scheduled Once a flow is identified as admitted, its admitted flow, redundant path configuration, and CQF cycle configuration are written to the admission table. For each admitted flow, a FRER flow identifier or flow handle, replication port or path identifier, hop-by-hop exit port for each replica path, and the first instance start cycle are generated. Period offset of subsequent instances and the Jump transmission cycle .
[0051] When CQF has queues for alternating receive and transmit, the corresponding egress queue can be determined based on the parity of the transmit cycle, eliminating the need to set separate queue selection variables for each transmission. The Time-Sensitive Network Controller (TSC) writes flow classification, egress selection, cycle queues, and capacity reservation entries to switches along the path, writes replica generation rules to the replication endpoint, and writes sequence recovery or duplicate frame elimination rules to the elimination endpoint. After this distribution is complete, network devices forward service frames according to this configuration.
[0052] In a sample configuration, the CQF period can be set to 125 microseconds, the time-triggered frame length can be selected from 128, 256, 512, and 1024 bytes, and the redundancy level can be set to 3. The maximum number of candidate paths can be configured to 16, the maximum number of candidate path combinations for precise evaluation can be configured to 80, and the maximum capacity of the fixed candidate path set library can be configured to 10. The above values are for illustrative purposes only; actual values should be set based on network topology, link speed, synchronization error, guard band, service cycle time, reliability requirements, and available computing budget.
[0053] To verify the necessity of shared link reliability assessment, a deterministic audit can be performed on the candidate three-path set. In an example containing a set of 108,000 tested paths, 46.67% of the set deemed feasible using path-independent approximation did not meet the precise reliability threshold given by equation (4), and the misjudgment rate increased with the average path overlap. This result indicates that counting shared links only once in the joint availability event can avoid significant overestimation of reliability.
[0054] Example 2 like Figure 2 As shown, this embodiment provides a shared link-aware redundant routing and CQF flow admission optimization system, including: The data acquisition module is configured to: construct a directed graph of a time-sensitive network to acquire the flow to be scheduled; The fixed candidate library construction module is configured to: generate candidate paths for each flow to be scheduled and form multiple sets of candidate redundant paths; determine the reliability of the path set of the candidate redundant path set and form a fixed candidate path set library; The resource-aware CQF decoder construction module is configured to: construct a resource-aware CQF decoder based on the stream processing order, the set of candidate redundant paths, and the CQF start period as search decisions; accumulate the byte requirements of each stream instance and each replica according to the link period; generate admission variables based on the reliability of the path set, the conservative CQF latency, the over-period boundary, and the link period capacity constraints; and evaluate candidate schemes according to the lexicographical order rule of weighted admission value first and resource target second best. The joint optimization and optimal solution determination module is configured to: within a preset decoding evaluation budget, using the fixed candidate path set library as the search space, jointly adjust the stream processing order, the candidate redundant path set, and the CQF start period, and perform feasibility determination and resource evaluation through the resource-aware CQF decoder; guide the flow rearrangement through admission gain to adjust the high-weight unadmitted flows blocked by capacity; when the weighted admission value stagnates continuously, guide the destruction and reconstruction through link cycle bottleneck to release the blocking capacity and re-insert the affected flow; and determine the current optimal executable solution according to the lexicographical order rules. The configuration distribution module is configured to: distribute the redundant path configuration and CQF periodic configuration corresponding to the admitted flow to the corresponding end system and switch according to the current optimal executable scheme.
[0055] The descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.
[0056] The proposed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and the division of modules described above is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.
[0057] Example 3 This embodiment also provides an electronic device, including: one or more processors, one or more memories, and one or more computer programs; wherein, the processor is connected to the memory, and the one or more computer programs are stored in the memory. When the electronic device is running, the processor executes the one or more computer programs stored in the memory to cause the electronic device to perform the method described in Embodiment 1.
[0058] It should be understood that in this embodiment, the processor can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.
[0059] Memory may include read-only memory and random access memory, and provides instructions and data to the processor. A portion of memory may also include non-volatile random access memory. For example, memory may also store information about the device type.
[0060] In the implementation process, each step of the above method can be completed by the integrated logic circuits in the processor hardware or by software instructions.
[0061] In this embodiment, the electronic device can be deployed as a centralized time-sensitive network controller, a software-defined network controller, or a network management server.
[0062] The method in Embodiment 1 can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor. The software modules can reside in readily available storage media in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, a detailed description is not provided here.
[0063] Those skilled in the art will recognize that the units and algorithm steps described in connection with the various examples of this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention.
[0064] Example 4 Embodiment 4 of the present invention provides a computer-readable storage medium.
[0065] A computer-readable storage medium having a program stored thereon that, when executed by a processor, implements the steps of the method as described in Embodiment 1 of the present invention.
[0066] The detailed steps are the same as those provided in Example 1, and will not be repeated here.
[0067] Example 5 Embodiment 5 of the present invention provides a time-sensitive network.
[0068] A time-sensitive network includes: a time-sensitive network controller, at least one replication end system, multiple switches supporting cyclic queuing and forwarding, and at least one elimination end system; The time-sensitive network controller is configured to execute the method described in Embodiment 1 above, obtain the optimal complete executable scheme in the current lexicographical order, and issue corresponding path and period configurations to the replication end system, the plurality of switches supporting cyclic queuing and forwarding, and the elimination end system according to the complete executable scheme, so that the replication end system generates data frame replicas according to the complete executable scheme, the plurality of switches supporting cyclic queuing and forwarding perform flow classification, egress selection, periodic queuing and capacity reservation according to the complete executable scheme, and the elimination end system performs sequence recovery of replica frames or duplicate frame elimination according to the complete executable scheme.
[0069] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A shared link-aware redundant routing and CQF flow admission optimization method, characterized in that, include: Construct a directed graph of a time-sensitive network to obtain the flow to be scheduled; Generate candidate paths for each flow to be scheduled and form a set of multiple candidate redundant paths; Determine the reliability of the path set for the candidate redundant path set and form a fixed candidate path set library; Using the stream processing order, candidate redundant path set, and CQF start period as search decisions, a resource-aware CQF decoder is constructed. The byte requirements of each stream instance and each replica are accumulated according to the link period. Admission variables are generated based on path set reliability, CQF conservative latency, over-cycle boundary, and link period capacity constraints. Candidate schemes are evaluated according to the lexicographical order rule of weighted admission value priority and resource target sub-optimization. Within the preset decoding evaluation budget, using the fixed candidate path set library as the search space, the stream processing order, candidate redundant path set, and CQF start period are jointly adjusted, and the feasibility is determined and resources are evaluated through the resource-aware CQF decoder; the high-weight unadmitted streams blocked by capacity are adjusted by guiding the flow rearrangement through admission gain; when the weighted admission value stagnates continuously, the blocked capacity is released and the affected stream is re-inserted by guiding the destruction and reconstruction through the link cycle bottleneck, and the current optimal executable scheme is determined according to the lexicographical order rule; Based on the current optimal executable scheme, the redundant path configuration and CQF periodic configuration corresponding to the admitted flow are distributed to the corresponding end systems and switches.
2. The shared link-aware redundant routing and CQF flow admission optimization method as described in claim 1, characterized in that, The process of generating candidate paths for each flow to be scheduled and forming multiple candidate redundant path sets specifically involves: for each flow to be scheduled, generating acyclic candidate paths of no more than a preset upper limit for the number of candidate paths through a weighted shortest simple path process, wherein the upper limit for the number of candidate paths is not less than the redundancy level of the flow to be scheduled; sorting subsequent candidate paths according to the reuse status of directed links in the selected candidate paths to form a path pool containing distinct candidate paths; selecting candidate paths from the path pool that are the same number as the redundancy level and distinct from each other according to the redundancy level of the flow to be scheduled to form a candidate redundant path set, and forming multiple candidate redundant path sets through different combinations of candidate paths.
3. The shared link-aware redundant routing and CQF flow admission optimization method as described in claim 1, characterized in that, The process of determining the reliability of the path set of candidate redundant paths and forming a fixed candidate path set library specifically involves: for each non-empty path subset of the candidate redundant path set, determining the union of directed links of each candidate path in the path subset, and calculating the joint availability probability based on the availability of each directed link in the union of directed links. Based on the joint availability probability corresponding to each non-empty path subset, the path set reliability of the candidate redundant path set is determined through inclusion-exclusion relationships; candidate redundant path sets that meet the minimum path set reliability requirement and whose replica paths satisfy the CQF conservative delay requirement are selected; the selected candidate redundant path sets are subjected to domination filtering based on path set reliability, shared directed link reuse count, and estimated cost of supercycle replication transmission, and the undominated portion of the candidate redundant path sets are retained according to a preset sorting rule to form a fixed candidate path set library.
4. The shared link-aware redundant routing and CQF flow admission optimization method as described in claim 1, characterized in that, The process involves accumulating the byte requirements of each flow instance and replica based on the link cycle, generating admission variables according to path set reliability, CQF conservative latency, over-cycle boundary, and link cycle capacity constraints, and evaluating candidate schemes according to a lexicographical order rule prioritizing weighted admission values and secondarily optimizing resource targets. Specifically, this includes: initializing the link cycle load matrix; processing each flow to be scheduled sequentially according to the flow processing order; enumerating the set of candidate redundant paths and the allowed CQF start cycles in the fixed candidate path set library for the current flow to be scheduled, and accumulating the byte requirements of each flow instance and replica on the link cycle pair; checking the path set reliability, CQF conservative latency, over-cycle boundary, and link cycle capacity constraints; generating an admission variable indicating that the flow has been admitted when at least one feasible insertion satisfies the constraints; and selecting the candidate redundant path set and CQF start cycle that minimizes the CQF resource target after insertion when multiple feasible insertions exist, while updating the link cycle load matrix; and generating an admission variable indicating that the flow has not been admitted when no feasible insertion exists. After all the scheduled flows have been processed, a weighted admission value is determined based on the admission weight and admission result of each flow. Candidate solutions are evaluated according to the lexicographical order rule of prioritizing weighted admission values and, in the case of the same weighted admission value, prioritizing the one with the smaller resource target.
5. The shared link-aware redundant routing and CQF flow admission optimization method as described in claim 1, characterized in that, The process of adjusting high-weight excluded flows blocked by capacity through admission gain-guided flow rearrangement specifically involves: for target flows excluded due to insufficient capacity, calculating the capacity gap on each link-period pair based on its candidate redundant path set and CQF start period, and forming a blocked link-period pair set with positive capacity gaps; identifying admitted flows that have positive load contributions on the blocked link-period pair set, and determining the degree of blocking contribution based on the ratio of the effective load that each admitted flow can release on the set to its admission weight; prioritizing the selection of excluded target flows with higher admission weights and smaller total capacity gaps, moving them before one or more major blocking contribution flows, and re-decoding the affected flow processing suffixes while maintaining the relative order of other flows; accepting the flow rearrangement only if the re-decoded complete scheme is strictly superior to the current scheme according to the lexicographical order rule.
6. The shared link-aware redundant routing and CQF flow admission optimization method as described in claim 1, characterized in that, The process of releasing blocked capacity and re-inserting affected flows by guiding destruction and reconstruction through link cycle bottlenecks specifically involves: when the weighted admission value does not increase for a preset number of consecutive rounds, selecting a set of candidate redundant paths that meet the reliability requirements of the path set and the conservative delay requirements of CQF, but which are not admitted due to insufficient link cycle capacity; based on the set of blocked link cycle pairs corresponding to the target flow, identifying admitted flows that have positive load contributions on at least one of the blocked link cycle pairs as candidate contributing flows; selecting one or more flows to be destroyed from the candidate contributing flows to form a destruction set to release the capacity on the blocked link cycle pairs; during reconstruction, first attempting to insert the target flow, and then re-inserting the flows in the destruction set in a deterministic order of decreasing admission weight, increasing estimated resource requirements, and flow identifier, and using only a fixed candidate path set library and allowed CQF start cycles during the reconstruction process; verifying the complete scheme after reconstruction through the resource-aware CQF decoder, and replacing the current scheme only when its weighted admission value increases, or when the weighted admission value remains unchanged and the resource target decreases.
7. A shared link-aware redundant routing and CQF flow admission optimization system, characterized in that, include: The data acquisition module is configured to: construct a directed graph of a time-sensitive network to acquire the flow to be scheduled; The fixed candidate library construction module is configured to: generate candidate paths for each flow to be scheduled and form multiple sets of candidate redundant paths; determine the reliability of the path set of the candidate redundant path set and form a fixed candidate path set library; The resource-aware CQF decoder construction module is configured to: construct a resource-aware CQF decoder based on the stream processing order, the set of candidate redundant paths, and the CQF start period as search decisions; accumulate the byte requirements of each stream instance and each replica according to the link period; generate admission variables based on the reliability of the path set, the conservative CQF latency, the over-period boundary, and the link period capacity constraints; and evaluate candidate schemes according to the lexicographical order rule of weighted admission value first and resource target second best. The joint optimization and optimal solution determination module is configured to: within a preset decoding evaluation budget, using the fixed candidate path set library as the search space, jointly adjust the stream processing order, the candidate redundant path set, and the CQF start period, and perform feasibility determination and resource evaluation through the resource-aware CQF decoder; guide the flow rearrangement through admission gain to adjust the high-weight unadmitted flows blocked by capacity; when the weighted admission value stagnates continuously, guide the destruction and reconstruction through link cycle bottleneck to release the blocking capacity and re-insert the affected flow; and determine the current optimal executable solution according to the lexicographical order rules. The configuration distribution module is configured to: distribute the redundant path configuration and CQF periodic configuration corresponding to the admitted flow to the corresponding end system and switch according to the current optimal executable scheme.
8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the shared link-aware redundant routing and CQF flow admission optimization method as described in any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the steps of the shared link-aware redundant routing and CQF flow admission optimization method as described in any one of claims 1 to 6.
10. A time-sensitive network, characterized in that, include: A time-sensitive network controller, at least one replication end system, multiple CQF-enabled switches, and at least one elimination end system; The time-sensitive network controller is configured to: execute the steps of the shared link-aware redundant routing and CQF flow admission optimization method as described in any one of claims 1 to 6, and distribute the current optimal executable scheme to the replication end system, multiple switches and the elimination end system respectively.
