Communication service impact analysis and detour path recommendation method based on Nx risk verification

By employing cross-layer topology stitching and tiered scheduling methods, damaged services in power communication networks are identified and optimized, resolving resource conflicts and failover issues in existing technologies and achieving more efficient communication service recovery and resource management.

CN123123232APending Publication Date: 2026-11-17STATE GRID HENAN INFORMATION & TELECOMM CO
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611305020.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-26
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing solutions are difficult to uniformly represent the cross-layer bearing relationship between the physical layer and the logical layer, do not fully consider the vulnerability of residual links after Nx failure, resulting in resource conflicts, path congestion and failover failures, and lack tiered scheduling and failover timing coordination, affecting the reliability and feasibility of communication services.

Method used

Damaged services are identified through cross-layer topology stitching, a spatiotemporal network topology map is constructed, and tiered scheduling is carried out by combining service importance and latency sensitivity. Multi-commodity flow collaborative optimization and resource conflict resolution are performed, implicit physical common source risks are verified, and a switching timing control sequence is generated.

Benefits of technology

It improves the completeness of damaged service identification, avoids redundant resource occupation, reduces the risk of secondary failures from detour paths, forms an executable switching sequence and maintenance push load, and enhances the reliability and executability of the communication network.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure FT_1
    Figure FT_1
  • Figure FT_2
    Figure FT_2
  • Figure FT_3
    Figure FT_3
Patent Text Reader

Abstract

This application discloses a communication service impact analysis and detour path recommendation method based on N-x risk verification. It forms a spatiotemporal network topology by stitching the physical and logical layers together, identifies damaged services through fault pruning, and implements tiered scheduling based on service importance and latency sensitivity. It quantifies residual link risks through secondary fault injection and combines switching windows to conduct collaborative optimization of concurrent service multi-product flows, resolve resource conflicts, and verify physical common-source risks. This improves the completeness of damaged service identification, avoids redundant occupation of bandwidth, ports, and other resources, reduces secondary failures and risks associated with shared cables and trenches in detour paths, and forms an executable switching sequence and maintenance push payload.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of intelligent recommendation, and more specifically, to a method for communication service impact analysis and detour path recommendation based on Nx risk verification. Background Technology

[0002] As the scale of power communication networks continues to expand, the reliance of protection, security control, and dispatch services on communication resources is constantly increasing. When multiple sites, optical cables, equipment, or trenches fail simultaneously due to maintenance, disasters, or other reasons, an Nx fault scenario is easily formed, triggering a concentrated switchover of multiple services. To accurately identify the scope of the fault's impact, ensure the priority restoration of critical services, and reduce the risk of secondary faults, it is necessary to construct a communication service impact analysis and detour path recommendation scheme oriented towards Nx risk verification.

[0003] Existing solutions typically establish topology relationships based on optical cables, equipment, optical paths, and service ledgers. They determine service interruption or transient status by judging whether the primary and backup service paths pass through failed resources, and employ algorithms such as Dijkstra's algorithm and K-shortest path algorithm to calculate alternative paths for affected services, avoiding failed resources, resources on the same cable, or overloaded resources. The analysis results are then pushed to maintenance tickets or maintenance plans. However, existing solutions mostly use static topology and single-service independent routing methods, making it difficult to uniformly represent the cross-layer bearing relationship between the physical and logical layers, and failing to fully consider the vulnerability of residual links to re-fail after an Nx failure. When multiple services recover concurrently, limited bandwidth, board ports, and cross-connection resources may be repeatedly occupied, leading to resource conflicts, path congestion, or failover failures. Furthermore, nominally separate paths may still have implicit physical co-source risks such as sharing the same optical cable or cable channel, and services with different importance and latency requirements lack tiered scheduling and failover timing coordination, resulting in insufficient reliability and executability of the recommended paths.

[0004] Therefore, an optimized method for communication service impact analysis and detour path recommendation based on Nx risk verification is needed. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a method for communication service impact analysis and detour path recommendation based on Nx risk verification.

[0006] According to one aspect of this application, a method for communication service impact analysis and detour path recommendation based on Nx risk verification is provided, which includes: Step 1: Based on the original data of the multi-source communication network, perform cross-layer topology stitching and fault pruning on the physical layer and logical layer to obtain a spatiotemporal network topology map and a list of candidate damaged services. Step 2: Based on the two-dimensional decision matrix composed of business importance and latency sensitivity, priority mapping and tiered sorting are performed on each business in the list of damaged business candidates to obtain the tiered scheduling queue and the switching window scheme. Step 3: Simulation calculations are performed to inject secondary faults into the residual links in the spatiotemporal network topology diagram one by one to obtain the spatiotemporal topology impedance matrix. Step 4: Using the spatiotemporal topology impedance matrix as the link cost weight and the switching time window scheme as the timing relaxation constraint, perform multi-commodity flow collaborative optimization and resource conflict resolution on the spatiotemporal network topology graph for concurrent services in the tiered scheduling queue to obtain a conflict-free concurrent detour path set and an updated spatiotemporal resource reservation graph. Step 5: Based on the updated spatiotemporal resource reservation map, perform implicit physical common source risk verification on the set of conflict-free concurrent detour paths, and generate a detour timing control sequence according to the detour window scheme to obtain the detour execution and maintenance push load.

[0007] Compared with existing technologies, this application provides a communication service impact analysis and detour path recommendation method based on Nx risk verification. It forms a spatiotemporal network topology by stitching the physical and logical layers together, identifies damaged services through fault pruning, and implements tiered scheduling based on service importance and latency sensitivity. It quantifies residual link risks through secondary fault injection and combines switching windows to conduct collaborative optimization of concurrent service multi-product flows, resolve resource conflicts, and verify physical common-source risks. This improves the completeness of damaged service identification, avoids redundant occupation of bandwidth, ports, and other resources, reduces secondary failures and risks associated with shared cables and trenches in detour paths, and forms an executable switching sequence and maintenance push payload. Attached Figure Description

[0008] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 This is a flowchart of a communication service impact analysis and detour path recommendation method based on Nx risk verification according to an embodiment of this application; Figure 2 This is a data flow diagram illustrating the communication service impact analysis and detour path recommendation method based on Nx risk verification according to an embodiment of this application; Figure 3 This is a flowchart of step one in the communication service impact analysis and detour path recommendation method based on Nx risk verification according to an embodiment of this application; Figure 4 This is a flowchart of step three in the communication service impact analysis and detour path recommendation method based on Nx risk verification according to an embodiment of this application. Detailed Implementation

[0010] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0011] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0012] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0013] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0014] In the technical solution of this application, a communication service impact analysis and detour path recommendation method based on Nx risk verification is proposed. Figure 1 This is a flowchart of a communication service impact analysis and detour path recommendation method based on Nx risk verification according to an embodiment of this application. Figure 2 This is a system architecture diagram of a communication service impact analysis and detour path recommendation method based on Nx risk verification, according to an embodiment of this application. Figure 1 and Figure 2As shown, the communication service impact analysis and detour path recommendation method based on Nx risk verification according to an embodiment of this application includes the following steps: S1, based on the original data of the multi-source communication network, cross-layer topology stitching and fault pruning are performed on the physical layer and logical layer to obtain a spatiotemporal network topology map and a list of candidate damaged services; S2, based on the two-dimensional decision matrix composed of service importance and latency sensitivity, priority mapping and hierarchical sorting are performed on each service in the list of candidate damaged services to obtain a hierarchical scheduling queue and a switching window scheme; S3, secondary faults are injected one by one into the residual links in the spatiotemporal network topology map. S4. Using the spatiotemporal topological impedance matrix as the link cost weight and the switching time window scheme as the timing relaxation constraint, multi-commodity flow collaborative optimization and resource conflict resolution are performed on the concurrent services in the tiered scheduling queue on the spatiotemporal network topology graph to obtain a conflict-free concurrent detour path set and an updated spatiotemporal resource reservation graph; S5. Based on the updated spatiotemporal resource reservation graph, implicit physical common source risk verification is performed on the conflict-free concurrent detour path set, and a switching timing control sequence is generated according to the switching time window scheme to obtain the detour execution and maintenance push load.

[0015] Specifically, S1, based on the original data of the multi-source communication network, performs cross-layer topology stitching and fault pruning on the physical and logical layers to obtain a spatiotemporal network topology map and a list of candidate damaged services. It should be understood that the logical channel of a power communication service is not directly equivalent to an independent physical link. A protection, security control, or dispatching service typically passes through multiple layers of resources, including the service channel, optical path, transmission equipment, boards, ports, fiber cores, optical cable segments, and trenches. If the failure status of a service is determined solely based on the service logical route or optical path name, it is impossible to accurately transmit the failure status of sites, equipment, boards, ports, optical cables, and trenches to the service layer, nor can it identify situations where multiple logical routes actually share the same optical cable segment, the same trench, or the same equipment at the physical layer. Therefore, in the technical solution of this application, cross-layer topology stitching is first performed on communication resource data from different sources and at different levels, so that any physical resource failure can be traced back to the corresponding optical path and service along the bearer relationship. Simultaneously, the bandwidth occupancy, remaining capacity, and transmission delay of the communication network at different points in time are not fixed. A topology formed solely from static ledgers can only indicate whether resources are connected, but it cannot indicate whether the corresponding links have sufficient capacity or transmission performance at the time of a failure. Therefore, state parameters from real-time service traffic snapshots are further injected into the topology, giving the resulting model both spatial connectivity and temporal state attributes.

[0016] It is worth noting that multi-source communication network raw data refers to a set of data obtained from different communication resource management systems, transmission network management systems, service management systems, and fault or maintenance management systems, which can collectively describe the physical structure, logical bearer relationships, real-time operating status, and fault scope of the communication network. Specifically, it includes optical cable physical segment ledger data, equipment board associated topology data, service logic routing mapping data, real-time service traffic snapshot data, and Nx fault target list data. Specifically, optical cable physical segment ledger data describes the unique identifier, name, starting and ending stations, primary line, trench, optical cable type, number of fiber cores, optical cable length, and adjacency relationships of optical cable segments. Equipment board associated topology data describes the inclusion and connection relationships between stations, equipment rooms, communication equipment, chassis, slots, boards, and ports, as well as the connection relationships between ports and optical paths and fiber cores. Service logic routing mapping data describes the bearer mapping between services and channels, channels and optical paths, optical paths and equipment ports, and optical paths and optical cable segments, and records the primary and backup service paths respectively. Real-time service traffic snapshot data is used to describe the used bandwidth, total capacity, bandwidth utilization, remaining available bandwidth, and transmission latency of each link, port, and channel at the time of failure or maintenance analysis. The Nx fault target list data describes x resources that are simultaneously assumed to be faulty during this verification, including resource identifier, resource type, failure time, failure duration, and failure scope.

[0017] Figure 3 This is a flowchart of step one in the communication service impact analysis and detour path recommendation method based on Nx risk verification according to an embodiment of this application. Figure 3 As shown, in the first embodiment of this application, S1 includes: S11, based on optical cable physical segment ledger data, equipment board associated topology data, and service logic routing mapping data, performing physical-logical topology cross-layer stitching from the physical pipeline layer to the service channel layer to obtain a multi-layer composite topology map; S12, injecting bandwidth utilization and transmission delay parameters from real-time service traffic snapshot data into the multi-layer composite topology map for dynamic spatiotemporal dimension superposition to obtain an initial spatiotemporal topology model; S13, based on Nx fault target list data, performing topology pruning and erasure and damaged service traversal extraction on the corresponding fault nodes and links in the initial spatiotemporal topology model to obtain a spatiotemporal network topology map and a damaged service candidate list.

[0018] Specifically, in step S11, based on the optical cable physical segment ledger data, equipment board associated topology data, and service logic routing mapping data, physical-logical topology cross-layer stitching is performed from the physical conduit layer to the service channel layer to obtain a multi-layer composite topology map. In this process, firstly, the optical cable physical segment ledger data is parsed to extract the unique identifier of the optical cable segment, optical cable name, starting station, ending station, primary line, trench, optical cable type, number of fiber cores, optical cable length, and connection relationships between optical cable segments. Based on this, the physical conduit layer and the optical cable fiber core layer are established. The physical conduit layer represents the sites, equipment rooms, trenches, and their spatial connection relationships, while the optical cable fiber core layer represents the optical cable segments, fiber cores, and their connection relationships between adjacent sites. Secondly, the equipment board associated topology data is parsed to extract the inclusion relationships between sites, equipment rooms, communication equipment, chassis, slots, boards, and ports, as well as the connection relationships between ports, between ports and fiber cores, and between ports and optical paths. Based on this, the equipment board port layer is established. For port connectivity relationships formed by cross-connections within the same communication device, logical connection edges are established between ports; for optical interfaces connected to fiber optic cores, physical connection edges are established between port nodes and fiber optic core nodes. Furthermore, the service logical routing mapping data is parsed to extract information such as the service's unique identifier, service type, primary service channel, backup channel, the optical path corresponding to the channel, the equipment and ports traversed by the optical path, and the optical cable segment and fiber optic core carried by the optical path. Based on this, the optical path transmission layer and service channel layer are established. For services with 1+1 protection, A / B port channels, 2M 1+1 protection, SNCP protection, or optical cut-off protection, primary and backup paths are established respectively, retaining primary / backup attributes, protection types, and path traversal order.

[0019] When different data sources use different resource codes, association is prioritized based on the unique resource identifier. If no unified identifier exists, matching is performed based on a combination of fields, including resource name, site, A-end and Z-end, device name, chassis slot, board name, port number, and transmission system. After matching is completed, a unified internal identifier is generated for each resource, while the original system identifier is retained to support subsequent write-back of analysis results.

[0020] Let the node sets of the physical conduit layer, optical fiber core layer, equipment board port layer, optical path transmission layer, and service channel layer be respectively... , , , and Then the node set of the multi-layer composite topology graph is:

[0021] in, It is a set of nodes in a multi-layered composite topology graph.

[0022] Let the set of internal connecting edges of each layer be . The set of cross-layer mapping edges between each layer is Then the multi-layer composite topology diagram is represented as:

[0023] in, It is a multi-layered composite topology graph. This is the set of internal connection edges for each layer. This is the set of cross-layer mapping edges between different layers. This includes static attributes such as resource type, resource name, site, capacity limit, physical length, protection type, and primary / backup attributes. Cross-layer mapping edges are established based on inclusion, connection, bearer, and routing relationships, and can be represented as:

[0024] in, This refers to inclusion relationships such as a board being included in a device, or a port being included in a board. This refers to the connection relationships between ports and resources such as fiber cores. For optical path carrying service channels and other bearer relationships, This represents the routing relationships of optical paths or service paths through equipment, optical cable segments, and trenches. After cross-layer stitching is completed, any service node can trace down its occupied optical paths, ports, equipment, fiber cores, optical cable segments, and trenches, or any physical resource node can query upwards all the optical paths and services it carries, outputting a multi-layer composite topology map.

[0025] Specifically, in step S12, the bandwidth utilization and transmission delay parameters from the real-time service traffic snapshot data are injected into a multi-layer composite topology graph for dynamic spatiotemporal overlay to obtain an initial spatiotemporal topology model. During this process, the real-time service traffic snapshot data corresponding to the fault analysis time is read, and each snapshot record is matched to a link, port, board, or device node in the multi-layer composite topology graph based on the unified resource identifier. The real-time service traffic snapshot includes at least the sampling time, resource identifier, total bandwidth capacity, used bandwidth, bandwidth utilization, and transmission delay; when the data source does not directly provide bandwidth utilization, it is calculated based on the total bandwidth capacity and used bandwidth.

[0026] Specifically, for the link At sampling time Total bandwidth capacity and used bandwidth Bandwidth utilization for:

[0027] in, For link At sampling time Bandwidth utilization.

[0028] Link at sampling time Remaining available bandwidth for:

[0029] in, For link At sampling time The remaining available bandwidth.

[0030] For a business path composed of multiple links in sequence Its end-to-end transmission delay is obtained by summing the transmission delay of each link segment in the path, the device processing delay, and the node queuing delay:

[0031] in, This refers to the transmission delay of each link segment in the path. Due to equipment processing delay, Queuing delay for nodes.

[0032] Bandwidth utilization, remaining available bandwidth, transmission latency, operating status, and sampling time are combined into resources. Dynamic attribute vectors:

[0033] in, For resources at all times The running status, For resources at all times Dynamic attribute vectors, Sampling time, For resources at all times bandwidth utilization For resources at all times The remaining available bandwidth, For resources at all times The transmission delay is calculated. For a single analysis moment, the dynamic attributes of that moment are directly attached to the corresponding nodes and links in the multi-layer composite topology graph; for multiple consecutive sampling moments, multiple time slices are established for the same resource, and time evolution edges are established between adjacent time slices to represent the changes in resource status over time.

[0034] By superimposing the multi-layered composite topology graph with dynamic state attributes, the initial spatiotemporal topology model is obtained:

[0035] in, This indicates that dynamic state attributes will be injected into the corresponding topology resources. For a moment The initial spatiotemporal topology model preserves all static bearer relationships from the physical pipeline layer to the service channel layer, while recording the bandwidth utilization, remaining capacity, transmission delay, and operating status of each link and node at the time of fault analysis. This provides a unified network state foundation for Nx fault pruning and subsequent detour path calculation.

[0036] Specifically, in step S13, based on the Nx fault target list data, topology pruning and erasure, as well as damaged service traversal extraction, are performed on the corresponding fault nodes and links in the initial spatiotemporal topology model to obtain a spatiotemporal network topology map and a candidate list of damaged services. During this process, resource identifiers, resource types, fault times, and fault ranges are read from the Nx fault target list, and corresponding sites, channels, optical cable segments, communication equipment, boards, or ports are matched in the initial spatiotemporal topology model. Let the set of fault targets included in this verification be:

[0037] in, This refers to the set of fault targets included in this verification.

[0038] After completing the direct fault target matching, the disability status is propagated along the cross-layer mapping relationship. When a site fails, the equipment, boards, ports, and intra-site connections that depend on the site are marked as disabled; when equipment fails, its boards, ports, and cross-connections are marked as disabled; when a board or port fails, the optical path connection established by the corresponding board or port is marked as disabled; when an optical cable segment or trench fails, the corresponding fiber core, optical path, and cross-layer bearing edge are marked as disabled.

[0039] Let the set of disabled nodes formed after fault matching and fault propagation be . The set of disabled links is Then the set of nodes after pruning is:

[0040] in, This is the set of disabled nodes.

[0041] The pruned link set is as follows:

[0042] in, This is the set of links after pruning.

[0043] Based on the pruned nodes, links, and their dynamic attributes, a spatiotemporal network topology diagram is formed:

[0044] in, To preserve the bandwidth utilization, remaining available bandwidth, transmission latency, and operational status of resources after a failure; This is a spatiotemporal network topology diagram. Then, iterates through all services in the business logic routing mapping data. For the first... Each service obtains its primary path. and alternative paths Check if the path contains disabled nodes or disabled links. Define the path disability indication function as follows:

[0045] The service status is determined based on the failure assessment results of the primary and backup paths:

[0046] This formula indicates that for services where the working path is not interrupted but the protection path is interrupted, the service status is recorded as normal, and a "protection path damaged" flag can also be recorded. For optical paths configured with optical cut-off protection, if only one side of the line is affected, it is considered normal; for cases without optical cut-off protection where the equipment is affected, it is considered interrupted.

[0047] Directly interrupted services and transiently recoverable services are added to the list of candidate damaged services. Each record in this list includes a unique service identifier, service name, service type, service level, scheduling level, maximum tolerable latency, bandwidth requirement, protection type, primary path status, backup path status, failed resources, affected optical paths, impact level, and measures taken. Finally, a spatiotemporal network topology diagram after fault pruning and a list of candidate damaged services are output. The former serves as the network foundation for secondary fault injection and detour path optimization, while the latter serves as the service input for service priority mapping, hierarchical sorting, and switchover window planning.

[0048] Specifically, in the first embodiment, after completing Nx fault target matching and topology pruning, sub-step S13 mainly identifies damaged services by determining whether the primary and backup paths pass through disabled nodes or disabled links. Essentially, it performs path connectivity determination on a single service basis: when both the primary and backup paths are interrupted, the service is determined to be directly interrupted; when the primary path is interrupted but the backup path remains connected, the service is determined to be momentarily interrupted and recoverable. This determination method can identify explicit path disconnections, but it assumes that backup path topology connectivity is equivalent to backup path protection switching being possible, without considering the competition between multiple services and shared protection resources.

[0049] In power communication network protection, security control, and dispatch service scenarios, even if backup paths maintain topological connectivity, they may still rely on the same protection ring segment, shared protection fiber optic group in the multiplexing segment, backup fiber optic pairs, switching board ports, or equipment cross-connection capacity. Nx faults typically cause multiple primary channels to fail simultaneously within the same switching cycle, leading to multiple services concurrently requesting the same protection resources. When the total bandwidth demand of these services exceeds the remaining capacity of the shared protection resources after the fault, the protection resources cannot simultaneously meet all switching requests. Some services, although nominally having backup paths, cannot actually recover due to insufficient link bandwidth, board ports, or cross-connection capacity, resulting in pseudo-protection interruption services.

[0050] While the initial spatiotemporal topology model recorded real-time bandwidth usage, remaining capacity, and link latency, the original process for identifying damaged services did not aggregate these state parameters according to the many-to-many relationship between services and shared protection resources, nor did it simulate concurrent protection requests within the same switchover window. Therefore, when a single service checks its backup path, it sees the same protection capacity that has not yet been deducted. Multiple services may be separately determined to be recoverable, ultimately resulting in duplicate commitments and over-allocation of the same resource. This deficiency is not due to a missing link capacity attribute, but rather the failure to model the competitive relationship between services formed through shared protection resources. Choosing the missing specific relationship as the entry point for improvement is more in line with the actual network operation mechanism.

[0051] The aforementioned misjudgment will cause pseudo-protection interruption services to be omitted from the list of candidate affected services. Subsequent tiered scheduling and detour path recommendations will not be able to reserve recovery resources for these services, and the impact scope in the maintenance ticket will be smaller than the actual impact scope. If the switchover is directly executed according to the misjudgment result, it may also cause overload of protection links, contention for board ports, cross-connection configuration failures, and repeated rollback of control signaling, further amplifying the congestion risk during fault recovery.

[0052] To address the aforementioned technical deficiencies, this application further proposes a second embodiment.

[0053] Specifically, firstly, based on the Nx fault target list data, a bipartite graph of fault nodes, fault pruning, and protection resource competition is constructed from the initial spatiotemporal topology model to obtain the spatiotemporal network topology graph and the protection resource competition bipartite graph. It should be understood that Nx faults simultaneously change physical link connectivity, logical channel reachability, and the available capacity of protection resources. If a competition relationship is established before fault pruning, already disabled links or ports may still be incorrectly included in the protection resource pool. Therefore, based on the resource identifiers in the fault list, corresponding physical sites, optical cable segments, duct segments, communication equipment, boards, and ports are matched in the initial spatiotemporal topology model. The disabled state is then propagated along the mapping relationship of the physical duct layer, optical cable core layer, optical path transmission layer, and service channel layer. The flow capacity and allocable capacity of the matched resources are set to zero, forming the spatiotemporal network topology graph after fault pruning.

[0054] Subsequently, the service logic channels in the spatiotemporal network topology are traversed. Services whose primary and backup paths both pass through disabled resources are marked as directly interrupted services, and this mark is written into the service layer attributes of the spatiotemporal network topology. For services whose primary path passes through disabled resources but whose backup path remains connected, they are identified as protection services to be switched over. The shared protection loop segment, backup fiber pair, switching board port, and cross-connection resources necessary to complete the switchover are extracted along the backup path. At the same time, the remaining available capacity of each resource at the time of the failure and the bandwidth requirements of the corresponding services are read.

[0055] Using the services to be switched as left-hand nodes and shared protection resources as right-hand nodes, an association edge is established between the service node and the resource node necessary for its switchover. The service bandwidth requirement is used as the weight of the association edge, and the remaining available capacity of the protection resource is used as the capacity attribute of the right-hand node, generating a bipartite graph of protection resource contention. This transforms the concurrent contention relationships originally implicit in multi-layered routing ledgers into a computable graph structure, allowing for unified identification of which services are handled by the same protection resource, the total bandwidth required for those services, and the remaining capacity available after a failure.

[0056] Next, protection capacity contention arbitration and pseudo-protection service identification are performed on the bipartite graph of protection resource contention to obtain the protection contention arbitration result. Here, it should be understood that simply identifying sharing relationships is insufficient to determine which services can complete protection switching. Therefore, all service bandwidth demands connected to the right-hand protection resource node are aggregated, and the aggregated bandwidth demands are compared with the remaining available capacity of the corresponding protection resource to calculate the protection contention overload ratio of the j-th shared protection resource node:

[0057] in, This represents the protection contention overload ratio of the j-th shared protection resource node; j represents the sequence number of the shared protection resource node. This represents the j-th shared protected resource; Indicates an application to use shared protected resources The set of services to be switched under protection; This represents the i-th protected service to be switched in the set; i represents the sequence number of the protected service to be switched. Indicates business The bandwidth required to complete the protection switch; Indicates shared protection resources The remaining available protection capacity after the Nx fault; the summation term represents the protection resources shared by all requests. The total bandwidth demand of the service. Specifically, when When the value is zero, no division operation is performed, and the corresponding resource is directly determined to be an unallocable resource.

[0058] When the protection contention overload ratio is no greater than 1, the capacity allocation result for the protection-related services awaiting handover associated with the corresponding resource node is registered. When the protection contention overload ratio is greater than 1, services are first sorted from highest to lowest service level, and then, if the service levels are the same, they are sorted from lowest to highest maximum tolerable latency. Finally, the remaining capacity of protection resources is checked item by item according to the sorting order. Considering that a backup path usually occupies multiple resources such as fiber optic cables, ports, and cross-connections, a cross-resource atomic allocation method is adopted. Allocation is confirmed and the capacity of each resource is deducted synchronously only when all protection resources on which the service depends meet the capacity requirements. If any necessary resource fails to meet the requirements, the temporary allocation of the service on other resources is revoked to avoid unusable fragmented reservations.

[0059] Services that acquire all necessary resources are marked as successfully switching over protection, while those that do not are marked as pseudo-protection interruptions. The competing resource identifier, capacity gap, arbitration order, and reason for failure are recorded to form the protection competition arbitration result. Overload ratio is used to locate protection resource bottlenecks, and atomized arbitration further converges the availability of backup paths to backup paths where the required resources can be actually allocated, eliminating the problem of multiple services repeatedly committing to the same protection capacity at the data level.

[0060] Furthermore, based on the spatiotemporal network topology map and the arbitration results of protection contention, pseudo-protection interruption services and direct path interruption services are merged, and the service level and bandwidth parameters of each service are extracted for damage classification fusion and encapsulation to obtain the spatiotemporal network topology map and the list of candidate damaged services. If the arbitration results are only used to generate alarms and are not written back to the topology, the subsequent detour path calculation will still be repeatedly allocated based on the undeducted protection resource status. Therefore, according to the protection contention arbitration results, the link bandwidth, board ports, and cross-connect capacity occupied by the successfully switched-out services are written back to the spatiotemporal network topology map, and the remaining capacity and service operation status of the corresponding resources are updated. For pseudo-protection interruption services, their canceled temporary capacity is not deducted, and the protection resources and capacity gaps that are in contention are recorded at the service layer.

[0061] Specifically, from the updated spatiotemporal network topology, services with direct interruptions where both the primary and backup paths are invalid are extracted. Services with pseudo-protection interruptions where backup paths are connected but resource allocation has failed are extracted from the protection contention arbitration results. Services with successful protection switching and successful resource allocation are also extracted. These three types of services are uniformly grouped and assigned the labels of direct interruption, pseudo-protection contention interruption, and transient recoverable interruption, respectively. For each service, a unique service identifier, service level, maximum tolerable latency, bandwidth requirement, protection type, invalid resources, contention resources, and arbitration status are simultaneously encapsulated to generate a candidate list of damaged services.

[0062] The resulting list of candidate damaged services covers both explicit interruptions caused by path loss and implicit interruptions caused by contention for shared protection resources. It also retains transient interruption states that can be restored through protection switching, and can be directly used as input for subsequent service prioritization, switching window planning, and detour path recommendation. The output is an updated spatiotemporal network topology and a list of candidate damaged services. The bipartite graph of protection resource contention is arbitrated to form a protection contention arbitration result, which is then used for state write-back and list generation. The entire data flow contains no unconsumed intermediate objects.

[0063] In summary, the second embodiment expands the original single-service path connectivity judgment into a processing chain that integrates fault topology pruning, shared protection resource contention modeling, protection capacity atomic arbitration, and service status write-back. Without changing the Nx fault input format, it enables service impact analysis to simultaneously consider path reachability and resource feasibility. Compared to simply checking whether the primary and backup paths are interrupted, the protection resource contention bipartite graph can centrally express the many-to-many dependencies of multiple services on the same protection ring segment, backup fiber group, board port, and cross-connection capacity. The protection contention overload ratio can quickly locate shared resources with insufficient capacity, avoiding redundant calculations combining unrelated services and resources across the entire network.

[0064] Atomized arbitration ensures that a service is considered recoverable only when all technical resources on which the entire backup path depends are available. This avoids configuration fragmentation where local resources are reserved but the overall service cannot be established, and reduces the processing load caused by repeated control plane deployments, configuration failures, and rollbacks. By writing the arbitration result back to the spatiotemporal network topology, subsequent detour paths are recommended using the actual remaining capacity after deducting protection occupancy. This reduces invalid candidate paths, excessive resource reservations, and redundant pathfinding calculations, improving the executability of fault recovery solutions.

[0065] The resulting candidate list of damaged services can identify pseudo-protection interruptions that nominally have backup paths but are actually unable to be switched over due to competition for protection resources. This makes the Nx risk verification results closer to the actual operating state of the communication network. This improves the completeness of damaged service identification and the success rate of protection switching, reduces the risk of missed reports of maintenance impact, shared link overload, and cascading congestion, and provides a consistent and verifiable network resource status for subsequent tiered scheduling, detour path recommendation, and maintenance ticket generation. This improvement is not a simple addition of judgments to existing capacity thresholds, but rather utilizes the multi-layered network topology after a fault, real-time remaining capacity, and service protection relationships to complete communication resource competition modeling and control decisions, directly impacting the allocation process of communication links, ports, and cross-connection resources.

[0066] Specifically, S2, based on a two-dimensional decision matrix composed of service importance and latency sensitivity, performs priority mapping and tiered sorting on each service in the candidate list of damaged services to obtain a tiered scheduling queue and a switchover window scheme. It should be understood that an Nx fault typically causes multiple protection, security control, scheduling data, and other communication services to be damaged simultaneously. Different services have different degrees of impact on the safe operation of the power grid, and their allowable interruption duration, recovery bandwidth requirements, and existing protection capabilities also differ. If switchover is performed only according to the order in which services are identified as damaged, services with lower importance or less stringent latency requirements may occupy remaining links, ports, and cross-connection resources first, resulting in important protection services or high latency-sensitive services not being able to recover within the maximum tolerable latency. Simultaneously, a large number of damaged services simultaneously initiating path calculation, port configuration, cross-connection establishment, and switchover confirmation requests to the control plane will cause a concentrated arrival of control signaling. The control plane needs to sequentially complete request parsing, path configuration, resource locking, equipment command issuance, and status confirmation. If switchover operations are not staggered according to control plane signaling processing latency, problems such as signaling queuing, equipment configuration contention, resource status not being written back before being reassigned, and switchover command timeouts may occur. Therefore, in the technical solution of this application, the importance of the service and its latency sensitivity are jointly determined, prioritizing important services that cannot be interrupted for extended periods in the recovery process. Furthermore, based on the signaling processing pace that the control plane can handle, the affected services are divided into multiple tiers and switchover windows. This process provides a definite service processing order for subsequent multi-product flow collaborative optimization and also provides time constraints for resolving resource conflicts.

[0067] In practice, firstly, normalized feature extraction and vectorized assembly are performed on the service level, maximum tolerable latency, bandwidth requirements, and protection type of each service in the candidate list of damaged services to obtain a service feature attribute matrix. During this process, for the first service in the candidate list of damaged services... Extract the service level of each service. Maximum tolerable delay Bandwidth requirements and protection type This forms the original service feature vector. The service level represents the importance of the service to the safe operation of the power grid and communication assurance. In practice, a service level mapping table is established based on service type, dispatch level, and service configuration assurance rules, converting the original levels of protection services, security control services, dispatch data services, and other services into ordered values. The larger the converted value, the higher the importance of the service. Range normalization is then applied to the service levels.

[0068] in, and These represent the minimum and maximum service levels in the candidate list of affected services, respectively. When they are equal, it indicates that all candidate services have the same service level. In this case, the normalized service level of all services is set to the same preset value to avoid the denominator being zero.

[0069] Maximum tolerable latency represents the maximum allowed time from the moment a service is affected by an Nx fault until the service completes protection switching or recovery. The smaller the maximum tolerable latency, the more sensitive the service is to recovery time. To ensure that services with stricter latency requirements obtain a larger characteristic value, we first take the reciprocal of the maximum tolerable latency, and then normalize the reciprocal:

[0070] in, and These are the minimum and maximum reciprocal values ​​among the candidate services, respectively. When the maximum tolerable latency of all services is the same, the normalized latency characteristics of each service are set to the same value.

[0071] Bandwidth requirements represent the continuous link bandwidth, board port capacity, and equipment cross-connection capacity that services need to obtain along the recovery path. Bandwidth requirements are normalized as follows:

[0072] in, and These represent the minimum and maximum bandwidth requirements for candidate services, respectively. When these two requirements are equal, the normalized bandwidth requirements for each service are set to the same value. Bandwidth requirements are primarily used for subsequent concurrent batching and multi-product flow capacity constraints, and do not directly represent the importance of a service based on their numerical value.

[0073] Protection type is a category attribute, including 1+1 protection, A / B port channel, 2M 1+1 protection, SNCP protection, optical switching protection, or no effective protection. Since the category codes for different protection types do not have a continuous order, a mapping table between protection type and protection urgency needs to be established based on service configuration assurance rules. Services with no effective protection, both primary and backup paths failing, or protection resources unavailable correspond to higher protection urgency; services with interrupted primary paths but still capable of failover via backup paths correspond to lower protection urgency. Normalizing the values ​​obtained from the protection type mapping yields the normalized protection urgency.

[0074] After completing the above feature transformation, the normalized service level, normalized latency characteristics, normalized bandwidth requirements, and normalized protection urgency of each service are vectorized and assembled according to the unique service identifier to obtain a service feature attribute matrix. Each row in this matrix corresponds to a damaged service, and each column corresponds to a normalized service feature. In addition to the matrix values, index information such as the unique service identifier, service name, service type, fault status, primary / backup path status, and failed resources is also retained to ensure that the subsequent scoring results can be accurately written back to the specific service.

[0075] Next, based on a two-dimensional decision matrix constructed from the reciprocal of business importance and latency sensitivity, a two-dimensional decision matrix mapping and scheduling priority score calculation are performed on each business in the business feature attribute matrix to obtain a business priority score list. In this process, firstly, business importance is calculated based on normalized service level and normalized protection urgency. The business importance of each item is:

[0076] in, For service level weighting, To ensure the protection urgency weight, both are non-negative numbers and their weights sum to one. The service level weight is usually no less than the protection urgency weight, so that the inherent importance of the business remains dominant; the protection urgency is used to prioritize businesses whose primary and backup paths are both interrupted or have no effective protection when the service levels are the same or close.

[0077] Delay sensitivity is calculated using the normalized result of the reciprocal of the maximum tolerable delay. Here, the smaller the maximum tolerable delay, the larger its reciprocal, and the higher the normalized delay sensitivity.

[0078] The importance and latency sensitivity of each service are assembled according to its unique identifier to form a two-dimensional decision matrix. The first column of this matrix represents the service importance, and the second column represents the latency sensitivity. Then, a weighted mapping is applied to each row of the two-dimensional decision matrix to calculate the scheduling priority score.

[0079] in, As a weight for determining the importance of business operations, The weights for latency sensitivity judgment are non-negative and summed to one. In scenarios where critical services such as protection and security control dominate, the weight for service importance judgment can be increased; in scenarios where service recovery periods are generally short, the weight for latency sensitivity judgment can be increased. These weights are pre-configured by the service assurance rules and remain unchanged during the same Nx risk verification process. After calculating the scheduling priority scores for all services, the unique service identifier, service name, service type, service level, service importance, maximum tolerable latency, latency sensitivity, bandwidth requirement, protection type, fault status, and scheduling priority score are associated and encapsulated to obtain a service priority score list.

[0080] Furthermore, based on the control plane signaling processing latency in the spatiotemporal network topology diagram, the service priority score list is sorted in descending order and divided into staggered time window intervals to obtain a tiered scheduling queue and a switching time window scheme. In this process, firstly, the service priority score list is sorted from highest to lowest according to the scheduling priority score. When scores are the same, the order is determined sequentially by maximum tolerable latency from smallest to largest, service level from highest to lowest, protection urgency from highest to lowest, and the unique service identifier. By setting deterministic tie-breaking rules, the same input will produce the same service order when executed repeatedly.

[0081] After sorting in descending order, services are divided into tiers based on the preset number of tiers or the scheduling priority score range. The first tier includes services within the highest score range, and subsequent tiers include services with lower priority in order of score range. When all services have the same score, all services are grouped into the same tier, and the internal order of the tier is determined by maximum tolerable latency, service level, and protection urgency.

[0082] To avoid excessive service requests within the same tier, leading to concentrated signaling requests, each tier is further divided into one or more service batches based on the maximum number of requests the control plane can process within a single time window. During batch division, both the number of services and bandwidth requirements are checked to ensure that the number of services within the same batch does not exceed the control plane's concurrent processing limit, and the total bandwidth requirements do not exceed the pre-scheduled capacity limit set based on the remaining resource status of the spatiotemporal network topology. Here, the pre-scheduled capacity is only used for initial batching; specific link capacity and port capacity constraints will be further verified in subsequent multi-product flow collaborative optimization.

[0083] Subsequently, the control node processing delay, control link propagation delay, device configuration delay, and status confirmation delay are read from the spatiotemporal network topology diagram. The control plane signaling processing delay for each service is obtained by sequentially accumulating the switchover request generation and parsing delay, the processing delay of each control node on the control signaling path, the propagation delay of each control link, the port configuration or cross-connection establishment delay, and the device status confirmation delay.

[0084] For the same service batch, the control plane signaling processing delay for each service within the batch is calculated separately, and the maximum value is used as the base processing time for the batch. A protection interval is then added to absorb signaling jitter, equipment response differences, and status write-back delays to obtain the switchover window length for the batch. Determining the window length based on the maximum processing delay within the batch ensures that all services in the same batch have sufficient signaling processing and status confirmation time.

[0085] Specifically, the first batch of the first tier begins at the moment the fault is confirmed and the initial resource state is frozen. The start time of the subsequent batch is determined by the start time of the previous batch, the time required for the previous batch to complete resource pre-locking, and the protection interval. That is, the subsequent batch starts after the previous batch completes resource pre-locking and passes the protection interval. When the control system allows parallel execution of the configuration distribution phase, adjacent switchover time windows can partially overlap in time, but the resource check and resource pre-locking phases should be staggered. When the control system does not support parallel configuration, adjacent switchover time windows are arranged end-to-end, with the subsequent batch starting after the previous batch completes all signaling processing and state confirmation. The end time of each service batch is obtained by adding the switchover time window length to the start time of that batch. After the initial division of the time windows, it is verified whether the end time of each service exceeds its latest recovery period. The latest recovery period of a service is obtained by adding the fault occurrence time to the maximum tolerable delay of that service. Only when the end time of the time window corresponding to a service is not later than the latest recovery period is it considered that the switchover time window meets the service recovery timeliness requirements. If a service cannot be switched over within the maximum tolerable time delay according to the initial tiered order, it will be moved to an earlier time window, but it must not overtake services with significantly higher importance that are also in a damaged state. If the maximum tolerable time delay still cannot be met after moving it forward, the service will be marked as a high-risk timeout service and given a mandatory priority processing flag during subsequent detour path optimization.

[0086] After the above-described descending sorting, tiered division, business batch division, and staggered time window calculation, a tiered scheduling queue and a switching time window scheme are output. The tiered scheduling queue includes at least the tier number, batch number, unique business identifier, business scheduling priority score, and internal tier order. The switching time window scheme includes at least the start time, end time, duration, control signaling offset, protection interval, and latest completion time for the corresponding business or business batch. This output serves as the business sequence constraint and timing relaxation constraint for subsequent multi-product flow collaborative optimization, ensuring that the concurrent detour path recommendation simultaneously meets the requirements of business importance, recovery time, and control plane processing capacity.

[0087] Specifically, S3 involves simulating the injection of secondary faults into each residual link in the spatiotemporal network topology to obtain the spatiotemporal topology impedance matrix. It should be understood that after the initial Nx fault pruning, although the residual links in the spatiotemporal network topology are in an available state, their reliability varies. Some residual links may be bridging links connecting two network areas; if they fail again, a large number of sites or service endpoints will lose connectivity. Some residual links may carry a large amount of traffic and have limited remaining bandwidth; even if the topology remains connected, they may become overloaded after concentrated service migration. Other links, although shorter in path, are adjacent to the fault area, and their endpoint devices or boards have a higher risk of failure again. If subsequent detour path recommendations only consider physical transmission delay or hop count as link costs, path calculations may focus on links with shorter apparent distances but higher vulnerability to secondary faults. When such a link fails again, it will cause the restored services to be interrupted again, and may lead to further service migration, shared link overload, and cascading congestion. Therefore, in the technical solution of this application, the re-failure of residual links and nodes is simulated before path optimization, and the network structure vulnerability, load pressure, current bandwidth utilization, and physical transmission delay are uniformly converted into link impedance. The larger the resulting spatiotemporal topological impedance, the higher the transmission cost, congestion risk, or secondary failure risk of the corresponding link; the smaller the spatiotemporal topological impedance, the lower the latency of the corresponding link, the more relaxed the bandwidth condition, and the smaller the impact of re-failure on the overall network.

[0088] Figure 4 This is a flowchart of step three in the communication service impact analysis and detour path recommendation method based on Nx risk verification according to an embodiment of this application. Figure 4 As shown, S3 includes: S31, performing Monte Carlo secondary fault injection and topology connectivity change simulation calculations on each residual link and node in the spatiotemporal network topology diagram to obtain secondary fault simulation results; S32, based on the total bandwidth demand of all services in the tiered scheduling queue, performing topology entropy increase and load collapse probability quantification assessment on the secondary fault simulation results to obtain vulnerability gradient vector; S33, based on physical delay and bandwidth utilization exponential penalty term, performing multi-factor nonlinear topology impedance synthesis on the vulnerability gradient vector and the state parameters of each link in the spatiotemporal network topology diagram to obtain spatiotemporal topology impedance matrix.

[0089] Specifically, in step S31, Monte Carlo secondary fault injection and topology connectivity change simulation calculations are performed on each residual link and node in the spatiotemporal network topology graph to obtain secondary fault simulation results. In this process, all residual nodes and residual links are first extracted from the spatiotemporal network topology graph, and the resource identifier, resource type, start and end nodes, current operating status, fault probability, dynamic remaining available bandwidth, and physical connection relationship of each residual resource are read. Links with an operating status of disabled or a dynamic remaining available bandwidth of zero are no longer considered as residual links for secondary fault injection. Before performing secondary fault injection, the baseline topology connectivity of the spatiotemporal network topology graph is calculated. Assume that the current spatiotemporal network topology graph contains... For any two distinct nodes, if there is at least one reachable path between them, the reachability indicator value is one; otherwise, it is zero. The baseline topological connectivity is the ratio of all reachable node pairs to the total number of node pairs:

[0090] in, This is a function indicating node reachability. With nodes The value is 1 when the network is connected and 0 when the network is disconnected.

[0091] Subsequently, each remaining link and each remaining node will be designated as the current forced disabling target. For the current target resources... The spatiotemporal network topology is copied to form a simulation topology copy, and the target resource is set to disabled in the simulation topology copy. When the target resource is a residual link, the link is deleted or its flow capacity and dynamic remaining available bandwidth are set to zero; when the target resource is a residual node, the node and all its connected incoming and outgoing links are synchronously set to disabled.

[0092] For each target resource, a predetermined number of Monte Carlo simulations are performed. In each simulation, the target resource is forcibly disabled, while other residual resources are randomly assigned a usable state or an additional disabled state based on their failure probability at the current moment. Resources in this simulation The status is:

[0093] in, A uniformly random number between zero and one. For resources At any moment The probability of secondary failure is given, with a state value of zero indicating resource failure and a state value of one indicating resource availability. If the current target resource is... Then, regardless of the value of the random number, it will... Set to zero.

[0094] After each fault state injection, a breadth-first search, depth-first search, or disjoint-set data structure is performed on the simulation topology to identify all connected components formed in that simulation. The connectivity of the simulation topology is then calculated using the same method as the baseline topology connectivity. Let the target resource... In the The topological connectivity in this simulation is: The decrease in connectivity in this simulation is:

[0095] in, For target resources In the The decrease in connectivity in the simulation.

[0096] Target Resources The average decrease in connectivity across all Monte Carlo simulations is:

[0097] in, This represents the total number of Monte Carlo simulations performed for a single target resource. A larger average connectivity decrease indicates a more severe loss of network reachability after the target resource becomes disabled again. After completing the simulation of all remaining links and nodes, the target resource identifier, target resource type, set of disabled resources in each simulation, number of connected components, number of nodes in each connected component, topological connectivity, connectivity decrease, source and destination node reachability status, and remaining carrying capacity after the fault are correlated and encapsulated to obtain the secondary fault simulation results. This result retains both the detailed data of each random simulation and the average connectivity decrease corresponding to each remaining link and node.

[0098] Specifically, in step S32, based on the total bandwidth demand of all services in the tiered scheduling queue, a topology entropy increase and load collapse probability quantification assessment are performed on the secondary fault simulation results to obtain a vulnerability gradient vector. In this process, firstly, the bandwidth demands of all services to be restored are read from the tiered scheduling queue and summed to obtain the total bandwidth demand of all services. Then, the topology entropy is calculated based on the connected component data. Let the target resource... In the Formed after the second simulation The connected component, the first The proportion of nodes contained in each connected component to the total number of available nodes is: Then the topological entropy of this simulation is:

[0099] When a network remains a single connected component after a failure, the connected component ratio is one, and the topology entropy is zero. When the network is divided into multiple connected components, the topology entropy is greater than zero. The more connected components there are and the closer their sizes are, the greater the topology entropy.

[0100] The baseline topology entropy without injecting secondary faults into the current target resource is calculated using the same method. Target resources In the The topological entropy increase in this simulation is:

[0101] The average topological entropy increase of the target resource across all simulations is:

[0102] Furthermore, the structural impact of a node's secondary failure is mapped to the residual links connected to that node. For endpoints that are nodes... and nodes Residual links This method weights and fuses the structural entropy increase caused by the failure of the link itself with the structural entropy increase caused by the failure of the two endpoint nodes. Through this processing, the risk of link failure and the risk of failure of equipment or sites at both ends of the link can be reflected in the structural vulnerability of the corresponding link.

[0103] Subsequently, the dynamic remaining available bandwidth of each residual link at the current moment is read from the spatiotemporal network topology diagram, and the ratio of the total bandwidth demand of all services to the dynamic remaining available bandwidth of the corresponding link is used as the load pressure term:

[0104] in, For residual links At any moment The dynamic remaining available bandwidth is determined. When the dynamic remaining available bandwidth is zero, division is not performed; instead, the link is marked as unallocable, and its load pressure is set to a preset maximum value. A load pressure greater than one indicates that the bandwidth required when all services concentrate on using this link exceeds its dynamic remaining available bandwidth. Here, to ensure that load pressures of different magnitudes have a uniform numerical range, a bounded transformation is performed on the load pressure.

[0105] Furthermore, a capacity feasibility check is performed on each Monte Carlo simulation following the failure of the target residual link. The capacity feasibility check reads the source node, destination node, and bandwidth requirements of each service in the tiered scheduling queue to determine whether the network after a secondary failure has a carrying scheme that satisfies source-destination reachability, flow conservation, and remaining link capacity constraints. When at least one mandatory guaranteed service's source and destination nodes are not connected, or the effective carrying capacity the network can provide after the failure is lower than the bandwidth requirements of the full range of services, the simulation is considered to have experienced load collapse. The bounded load pressure and load collapse probability are weighted and synthesized to obtain the comprehensive load risk of the residual link. Among these, the residual link... The load collapse probability is the proportion of the number of times the link is forcibly disabled in all simulations out of the total number of simulations.

[0106] The joint structural entropy increase of all residual links is normalized, and then the normalized joint structural entropy increase is weighted and synthesized with the overall load risk to obtain the secondary failure vulnerability gradient of each residual link. The larger the vulnerability gradient, the higher the risk of network fragmentation or service load collapse when the corresponding residual link fails again. Finally, the vulnerability gradients of all residual links are vectorized and assembled according to the residual link index order to obtain the vulnerability gradient vector.

[0107] Specifically, in step S33, based on the physical delay and bandwidth utilization exponential penalty term, a multi-factor nonlinear topological impedance synthesis is performed on the vulnerability gradient vector and the state parameters of each link in the spatiotemporal network topology diagram to obtain the spatiotemporal topological impedance matrix. In this process, firstly, the physical transmission delay, currently used bandwidth, total bandwidth capacity, current bandwidth utilization, and dynamically remaining available bandwidth of each residual link are extracted from the spatiotemporal network topology diagram. To eliminate the differences in the magnitude of physical transmission delay among different residual links, the ratio of the physical transmission delay of each residual link to the maximum physical transmission delay among all residual links is used as the normalized physical transmission delay. For the current bandwidth utilization, an exponential penalty method is used for nonlinear transformation to suppress subsequent path optimization from continuing to focus on selecting links close to saturation.

[0108] Subsequently, based on the residual link index, the secondary fault vulnerability gradient of the corresponding link is read from the vulnerability gradient vector. The normalized physical transmission delay, bandwidth utilization exponential penalty value, and secondary fault vulnerability gradient are then weighted and summed to obtain the spatiotemporal topological impedance of the residual link. For links that are disabled in operation or have zero dynamic remaining available bandwidth, their topological impedance is set to an unreachable value, preventing subsequent path optimization from selecting this link. After calculating the spatiotemporal topological impedance of each residual link, the remaining links are matrix-assembled according to their corresponding start and end nodes. For matrix positions corresponding to the same node, the impedance value is set to zero; for two different nodes with a direct residual link connection, the spatiotemporal topological impedance of the corresponding link is written into the corresponding node pair position; for node pairs without a direct connection, the impedance value is set to an unreachable value. It is worth noting that when multiple parallel residual links exist between two nodes, the index relationship between the node pair and multiple specific links can be retained, and the impedance of each link can be stored separately; when only one value is allowed to be stored in the node pair matrix, the minimum available impedance among the parallel links is written into the matrix, while the unique identifier of the selected link is stored separately.

[0109] Specifically, S4 uses the spatiotemporal topological impedance matrix as the link cost weight and the switching window scheme as the timing relaxation constraint to perform multi-commodity flow collaborative optimization and resource conflict resolution on the spatiotemporal network topology graph for concurrent services in the tiered scheduling queue, in order to obtain a conflict-free concurrent detour path set and an updated spatiotemporal resource reservation graph. It should be understood that an Nx fault may cause multiple protection, security control, and scheduling services to simultaneously initiate recovery requests within the same or adjacent switching windows. If the shortest path is calculated independently for each service, multiple services may simultaneously select the same residual link, board port, or device cross-connection resource with lower topological impedance. Each service may meet the capacity requirements when calculated individually, but after merging all services for execution, the sum of their bandwidth or port occupancy may exceed the remaining resource capacity, causing link overload, port contention, cross-connection establishment failure, or control command rollback. Therefore, in the technical solution of this application, multiple services within the same switchover window are jointly modeled as interconnected commodity flows, ensuring that path selection simultaneously satisfies constraints related to service source-destination connectivity, link capacity, node cross-connection capacity, port quantity, and switchover window. For links experiencing resource conflicts, the cost weight is increased, and candidate paths are reselected. This can guide some low-priority services to other available links, avoiding duplicate commitments to the same resource. After conflict resolution, the bandwidth and port occupancy of the selected path are written back to the spatiotemporal network topology map, enabling subsequent tiered services to continue routing based on the deducted actual remaining resources.

[0110] In practice, firstly, using the spatiotemporal topological impedance matrix as the cost weight and the switching window scheme as the timing constraint, a multi-product flow source-destination mapping and optimization model construction are performed on the spatiotemporal network topology graph for each service in the tiered scheduling queue to obtain a multi-product flow optimization model. During this process, according to the switching window scheme, the services in the tiered scheduling queue are divided into multiple service sets corresponding one-to-one with the switching window. Each service can only participate in path solving within its assigned switching window or the switching window allowed after timing relaxation.

[0111] For each service, the source node, destination node, bandwidth requirement, port usage requirement, service priority, and allowed switchover window are extracted from the service logic routing mapping data, and these attributes are combined into a single product flow. Each product flow has a defined start point, destination point, resource requirements, and time constraints. For the relationship between each service and each residual link, a path selection status is set. When a service selects the residual link within the corresponding switchover window, the path selection status is selected; otherwise, it is unselected. The path selection status of each service collectively constitutes the decision variables of the multi-product flow optimization model.

[0112] For each service, all remaining links selected within the corresponding switching time window are statistically analyzed. The impedance values ​​of these links in the spatiotemporal topology impedance matrix are read and accumulated. The accumulated path impedance of all services is then used as the overall target value. The solver selects the service path combination with the minimum overall target value while satisfying all constraints.

[0113] For each service, a flow conservation constraint is set. At the service source node, the number of selected outgoing links should be one more than the number of selected incoming links; at the service destination node, the number of selected incoming links should be one more than the number of selected outgoing links; at intermediate nodes other than the source and destination nodes, the number of selected incoming links should be equal to the number of selected outgoing links. This constraint ensures that each service forms a continuous and uninterrupted path between its source and destination nodes. For each residual link, a link remaining capacity constraint is set. Specifically, the bandwidth requirements of all services selecting the residual link within the same switchover window are summed to obtain the aggregated bandwidth occupancy of that link within the corresponding window. This aggregated bandwidth occupancy must not exceed the dynamic remaining available bandwidth of that link within that window. For each intermediate node, a cross-connection capacity constraint is set. The system counts all services passing through that node within the same switchover window and sums the bandwidth requirements of these services to obtain the cross-connection occupancy of the corresponding node within that window. This occupancy must not exceed the node's dynamic remaining cross-connection capacity. For services with discrete port occupancy requirements, a port number constraint is also required. The system adds up the service port demands passing through corresponding nodes within the same switchover window, and the total number of occupied ports must not exceed the number of remaining available ports on that node within that window. Regarding switchover window constraints, the actual start time of a service must not be earlier than the earliest start time allowed by the switchover window, and the sum of the actual start time of the service and the control plane processing and switchover execution time must not be later than the latest completion time allowed by the switchover window. This constraint allows for limited offsets of services within the switchover window, but cannot exceed the window's end point or the service's maximum tolerable latency. After completing the above service mapping, objective function construction, and setting constraints for flow conservation, link capacity, node cross-connection capacity, port quantity, and switchover window, a multi-product flow optimization model is obtained.

[0114] Next, the multi-commodity flow optimization model is solved iteratively in batches and concurrent collisions are resolved to obtain a set of conflict-free concurrent detour paths. In this process, firstly, based on the failover window scheme, the tiered scheduling queue is divided into multiple business sets corresponding one-to-one with the failover window, and each business set is solved iteratively in batches according to its priority from high to low. Specifically, the highest priority tier and its earliest failover window are solved first, and the resulting feasible paths and resource usage results are temporarily locked; then, the businesses in the next tier or the next failover window are solved. Confirmed resource reservations for high-priority businesses are retained first, and low-priority businesses must not overwrite their corresponding reservations.

[0115] Each service is mapped to a commodity flow with source nodes, destination nodes, and bandwidth requirements. The selection status of each service on residual links is used as the decision variable, and the optimization objective is to minimize the sum of the topological impedances of the selected residual links for each service. Flow conservation constraints, link remaining capacity constraints, node cross-connection capacity constraints, and switchover time window constraints are set. Specifically, for each round of solution results, the system calculates the aggregate bandwidth occupancy of each residual link within each switchover time window. Specifically, all services that select the corresponding link within the current time window are selected, and the bandwidth requirements of these services are summed. When the aggregate bandwidth occupancy exceeds the dynamic remaining available bandwidth of the link, a capacity conflict is determined for that link. The portion of the aggregate bandwidth occupancy exceeding the dynamic remaining available bandwidth is considered excess occupancy.

[0116] For links experiencing capacity conflicts, a non-negative capacity constraint multiplier is introduced. Specifically, based on the capacity constraint multiplier of the previous round for that link, the product of the excess occupancy and a preset update step size is added. The larger the excess occupancy, the more the capacity constraint multiplier increases. For links without capacity conflicts, the capacity constraint multiplier does not increase or gradually decreases according to preset rules, but it must not be less than zero. After updating the capacity constraint multiplier, the original spatiotemporal topological impedance of the link is added to the capacity conflict penalty to obtain the link cost for the next iteration. The capacity conflict penalty is determined jointly based on the link capacity constraint multiplier and the service bandwidth requirement. Therefore, high-bandwidth services incur higher costs when continuing to occupy overloaded links, and the path solving process will tend to transfer them to other residual links with available capacity.

[0117] When the sum of bandwidth usage on the same residual link by multiple services within the same handover window exceeds the dynamically remaining available bandwidth of that residual link, a capacity constraint multiplier is introduced to increase the cost weight of that residual link. Based on the updated cost weight, candidate paths are then reselected for the conflicting services. In other words, when a conflict occurs, the system prioritizes retaining the feasible paths already selected by high-priority services, and then performs candidate path reselection for low-priority services using conflicting resources. Candidate path reselection can employ the shortest path algorithm after updating the link cost. The shortest path algorithm or multi-commodity flow optimization solver completes the process. For conflicts in node cross-connection capacity or port quantity, the same handling logic as for link capacity conflicts is adopted. The system calculates the excess occupancy of node cross-connections or ports, updates the corresponding node capacity constraint multipliers or port constraint multipliers, and then distributes the node penalty to the relevant links entering or leaving that node, so that the probability of selecting conflicting nodes is reduced in the next round of candidate path reselection.

[0118] After candidate path reselection is completed, the aggregate bandwidth usage of each remaining link, node, and port within the corresponding switchover window is recalculated. If resource over-utilization still exists, the capacity constraint multiplier is updated and a new path is selected. This process is repeated until the aggregate bandwidth usage of all remaining links does not exceed their dynamic remaining available bandwidth, the cross-connection usage of all nodes does not exceed their dynamic remaining cross-connect capacity, and the port usage does not exceed their remaining available port count.

[0119] When a conflict-free path cannot be found for a service within the current failover window, the system adjusts its actual start time within the allowed timing relaxation range for that service, or moves it to a subsequent available window that can still meet the latest completion time. During adjustment, the remaining capacity of links, nodes, and ports within the target window is reread, and path calculation is performed again. If no feasible path exists for a service within all allowed failover windows, the service is marked as having an infeasible path, and no inoperable detour path is generated for it. Simultaneously, the conflicting links or nodes, capacity gaps, attempted candidate paths, and reasons for solution failures are recorded.

[0120] When all successfully solved services meet the constraints of link capacity, node cross-connection capacity, number of ports, and switching window, the unique identifier of each service, source node, destination node, selected residual link sequence, path topology impedance, bandwidth occupancy, port occupancy, tier, and switching window are associated and encapsulated to obtain a set of conflict-free concurrent detour paths.

[0121] Furthermore, based on the bandwidth and port occupancy of each path in the conflict-free concurrent detour path set, the spatiotemporal network topology map is dynamically updated with reservation deduction within the corresponding time window interval to obtain the updated spatiotemporal resource reservation map. After obtaining the conflict-free concurrent detour path set, the system updates the spatiotemporal network topology map according to the switching time window corresponding to each service. For each residual link and each switching time window, all confirmed services passing through the corresponding link within that time window are selected, and the bandwidth requirements of these services are added together to obtain the reserved bandwidth of that link within the corresponding time window. The dynamic remaining available bandwidth after the link has been reserved is obtained by deducting the reserved bandwidth within the corresponding time window from the dynamic remaining available bandwidth before reservation. The deduction result must not be less than zero; if the calculation result is less than zero, it indicates that there is an anomaly in the aforementioned conflict verification or data status, and the reservation should be revoked and resource conflict resolution should be re-executed.

[0122] For node cross-connect capacity, the system adds up all confirmed service bandwidth requirements passing through the corresponding node within the same time window to obtain the node cross-connect reserved capacity. This reserved capacity is then subtracted from the dynamic remaining cross-connect capacity before the node's reservation, resulting in the updated dynamic remaining cross-connect capacity. For port resources, the system adds up all confirmed service port requirements passing through the corresponding node within the same time window to obtain the port reserved quantity. This reserved quantity is then subtracted from the remaining available port quantity before the node's reservation, resulting in the updated remaining available port quantity.

[0123] After reserving bandwidth, cross-connect capacity, and number of ports, the link bandwidth utilization rate is recalculated based on the total link bandwidth capacity and the updated used bandwidth. The new bandwidth utilization rate, dynamic remaining available bandwidth, reserved service identifier, reservation start time, reservation end time, and reservation status are then written into the corresponding time slice of the spatiotemporal network topology map.

[0124] For adjacent but not completely overlapping switchover windows, the system divides the time axis into multiple continuous intervals according to the start and end times of each window, and calculates the reserved service capacity in the effective state within each interval. The same link can have different reserved bandwidth, remaining bandwidth, and bandwidth utilization in different time intervals.

[0125] When a service failover instruction has not yet been executed, the resource reservation status is set to pending activation. After a successful service failover, the reservation status changes from pending activation to occupied. If a service failover fails, the corresponding reservation is revoked and the resource capacity is restored. When the original path is restored, the service completes its back-switch, or the maintenance plan ends, the corresponding link bandwidth, node cross-connection capacity, and port resources are released. The resulting updated spatiotemporal resource reservation map retains the nodes, links, physical-logical mapping relationships, and fault states from the original spatiotemporal network topology map, while adding the resource reservation amount, remaining capacity, reserved service affiliation, reservation validity period, and reservation status within each failover window.

[0126] Specifically, S5, based on the updated spatiotemporal resource reservation map, performs implicit physical common-source risk verification on the set of conflict-free concurrent detour paths, and generates a switching timing control sequence according to the switching time window scheme to obtain the detour execution and maintenance push load. It should be understood that multi-commodity flow collaborative optimization can eliminate conflicts in link bandwidth, node cross-connect capacity, and port quantity, but no over-utilization of resources does not equate to the primary and backup paths being physically independent. Two logical routes may pass through different optical paths, different fiber cores, or even different optical cables, yet be laid together in the same trench, pipe, drop section, or station wiring interval; they may also pass through different logical ports but share the same equipment, board, or cross-connect unit. Such paths have no common link in logical topology, but may fail simultaneously due to the same physical event, forming an implicit physical common-source risk. Furthermore, recommended paths only have practical executability when converted into timed execution device instructions. Directly dismantling the old path before establishing the new one would cause service interruption; without considering the transmission delay difference between the old and new paths, message gaps, out-of-order messages, or duplicate receptions might occur during the switchover. Therefore, in the technical solution of this application, a lossless switchover method of establishing, switching, and releasing is adopted according to the switchover time window and the delay difference between the old and new paths, and each control command is arranged at the millisecond level. After completing path verification and timing arrangement, based on the actual site address, equipment, board, slot, and port information, the algorithm results are converted into a structured payload that can be recognized by the maintenance system and equipment control interface, enabling the maintenance ticket to automatically fill in the maintenance work content, affected services, detour paths, execution order, risk verification results, and rollback measures.

[0127] In specific implementation, firstly, based on the physical pipelines and co-channel ledgers in the updated spatiotemporal resource reservation map, implicit physical common-source isolation is verified for the primary and backup paths in the conflict-free concurrent detour path set, and local fine-tuning of illegal paths is performed to obtain the verified detour path set. In this process, firstly, the conflict-free concurrent detour path set is traversed, and according to the physical-logical mapping relationship in the updated spatiotemporal resource reservation map, the primary path, original backup path, and recommended detour path for each service are expanded layer by layer. The expansion process traces sequentially from the service channel layer to the optical path transmission layer, equipment board port layer, optical cable core layer, and physical pipeline layer, extracting the optical cable segments, trenches, pipelines, sites, equipment, boards, port groups, and cross-connection units traversed by each path. For the first... For each service, its primary path and backup or detour paths are specified in the first... The co-source ratio on physical resources is defined as the ratio of the number of resources shared by two paths to the total number of resources traversed by the two paths. Wherein, the... Physical resources can be optical cables, trenches, equipment, boards, or port groups. Next, the common-source ratios of different physical resource categories are weighted and synthesized according to risk weights to obtain the comprehensive physical common-source risk; the physical common-source isolation degree of the primary and backup service paths is one minus the comprehensive physical common-source risk. The physical common-source isolation degree is compared with the isolation threshold for the corresponding service type. When the isolation degree is not less than the threshold, the primary and backup paths are deemed to have passed the implicit physical common-source risk verification; when the isolation degree is less than the threshold, the corresponding detour path is deemed to be in violation. For critical services such as protection and security control, the same optical cable or the same cable channel can be set as a hard prohibition condition, that is, as long as the primary and backup paths share any restricted optical cable segment or cable channel, the verification is directly deemed to have failed.

[0128] For paths that fail verification, locate the physical resources where the primary path and the detour path share a common source and mark the logical links mapped to these physical resources as illegal links. Then, select a reachable node before the illegal link as the local replacement start point and select a reachable node after the illegal link as the local replacement end point, temporarily removing the illegal sub-path between the two nodes from the current detour path.

[0129] The search scope for local candidate paths is the spatiotemporal network topology subgraph between the local replacement start point and the local replacement end point. During the search, restricted optical cables, channels, devices, boards, or port resources shared with the primary path are prohibited from being selected. The remaining bandwidth, number of ports, cross-connect capacity, and switchover window status in the updated spatiotemporal resource reservation graph are read to exclude candidate links with insufficient resource capacity or that are unavailable within the corresponding window.

[0130] For each local candidate path, a local adjustment cost is calculated. This cost is weighted by the sum of the candidate path's spatiotemporal topological impedance, the increased transmission delay compared to the original non-compliant sub-path, the physical common-source risk of the candidate path, and the change in resource reservations. The system selects the candidate path that satisfies the constraints of physical isolation, link capacity, node cross-connection capacity, number of ports, and switchover window, while minimizing the local adjustment cost, to replace the original non-compliant sub-path. Before replacement, resource availability is checked again for the candidate path; if the check passes, the reserved resources of the original non-compliant sub-path are released atomically, and the resources required for the new local path are locked. If locking the new resources fails, the original reservation status remains unchanged, and another candidate path is selected.

[0131] After completing the partial replacement, the physical common-source isolation of the entire primary / backup path is recalculated. If the isolation threshold is still not met, the next common-source resource is located and local fine-tuning is performed until the verification is passed or no feasible alternative path exists. If the common-source risk cannot be eliminated, the corresponding service is marked as physically unavailable, and shared optical cables, shared trenches, shared equipment, capacity gaps, and manual handling suggestions are recorded. This path is not directly used as the automatic execution path.

[0132] The verified service paths and the partially adjusted service paths together form a set of verified detour paths. Each record in this set includes the service identifier, the original path, the detour path, the optical cables and trenches traversed, the physical common source isolation, the verification conclusion, the locally adjusted section, and the updated resource reservation information.

[0133] Next, combining the switching window scheme with the delay difference between the old and new paths in the updated spatiotemporal resource reservation map, millisecond-level lossless switching timing offset calculation and instruction arrangement are performed on the verified detour path set to obtain the switching timing control sequence. During this process, for each service in the verified detour path set, the transmission delay, node processing delay, and node queuing delay of the links traversed by the original path and the new detour path are read from the updated spatiotemporal resource reservation map. Following the original path's traversal order, the physical transmission delay of all links in the path is accumulated, along with the processing delay and queuing delay of all traversed nodes. The end-to-end transmission delay of the new detour path is calculated in the same way, i.e., the transmission delay of each link in the new path, as well as the processing delay and queuing delay of each node, are accumulated. Then, the delay difference between the new and old paths is obtained by calculating the difference between the end-to-end delay of the new detour path and the end-to-end delay of the original path. When the time delay difference is greater than zero, it means that the new path is slower than the old path, and the new path needs to be activated in advance or the dual-path transmission needs to be started in advance. When the time delay difference is less than zero, it means that the new path is faster than the old path, and the time difference needs to be absorbed by receiving buffer, sending delay or sequence number rearrangement. When the time delay difference is equal to zero, it means that the theoretical transmission delay of the new and old paths is the same, but a protection interval still needs to be reserved to absorb measurement errors and equipment response deviations.

[0134] The lossless switching timing offset is obtained by taking the absolute value of the delay difference between the old and new paths and then adding a preset protection interval. The protection interval is used to absorb link delay jitter, equipment response deviation, control signaling processing errors, and clock synchronization errors. Therefore, regardless of whether the new path is faster or slower than the old path, sufficient time is reserved for dual-path switching and message connection.

[0135] If the new path delay is greater than the old path delay, the new path activation time will be advanced by the corresponding delay difference and protection interval relative to the service switching time, so that the first batch of packets on the new path arrives at the receiving end near the switching time. If the new path delay is less than the old path delay, a buffer or delivery offset adapted to the delay difference will be set on the receiving side of the new path, so that packets from the new and old paths are delivered to the upper layer in the order of the service sequence, avoiding out-of-order delivery caused by packets from the new path arriving early.

[0136] For each service, the allowed start and end points of the time window for that service are read from the switchover time window scheme, and the service switchover time is determined by combining this with the time required for the control plane to establish and verify the new path. The service switchover time must be within the corresponding switchover time window, and sufficient time must be available before the switchover time for new path establishment, path verification, and lossless switchover offset.

[0137] Based on this, using the planned service switchover time as a benchmark, subtract the time required for establishing and verifying the new path, and then subtract the lossless switchover timing offset. The resulting time must not be earlier than the start of the switchover window. If the calculated result is earlier than the start of the window, the service switchover time needs to be adjusted backward; if the adjustment would exceed the end of the window, it indicates that the current window is insufficient to complete the lossless switchover, and it needs to be adjusted to another window within the allowed timing relaxation range, or the service should be marked as having an infeasible window.

[0138] The old path release time is determined jointly based on the larger end-to-end transmission delay between the old and new paths, the device status acknowledgment delay, and the protection interval. Specifically, based on the service handover time, the larger of the end-to-end delays of the old and new paths is added, followed by the status acknowledgment delay and the protection interval, to obtain the earliest time the old path is allowed to be released. This process ensures that all in-transit packets in the old path have been cleared, the new path's status has stabilized, and the service receiver has confirmed a successful handover.

[0139] The specific instruction sequence is as follows: First, lock the required links, board ports, and cross-connection resources for the new path; after successful resource locking, issue a new path cross-connection establishment instruction; after the new path is established, perform connectivity, available bandwidth, actual transmission latency, and error rate status verification; after successful verification, enter short-term dual-path transmission or establish-then-dismantle state; when the service switching time arrives, switch service traffic to the new path; after confirming that the new path is continuous and stable and that the packets in transit on the old path have been cleared, release the old path resources. Each instruction has pre-dependencies. For example, the new path cross-connection establishment instruction depends on successful resource locking, the path verification instruction depends on successful cross-connection establishment, the service switching instruction depends on successful path verification, and the old path release instruction depends on successful service switching and status confirmation. Subsequent instructions can only be executed if the pre-completion instructions are successful. When the new path establishment fails, connectivity verification fails, the measured latency exceeds the allowable value, the error rate exceeds the threshold, the resource status is inconsistent with the reserved status, or the status confirmation times out, the control sequence does not execute the old path release instruction, but triggers a rollback process. The rollback process cancels the cross-connection of the new path, releases the resources temporarily occupied by the new path, and maintains or restores the service carrying capacity of the old path; at the same time, it records the failure stage, target device, error status and rollback result.

[0140] For multiple services within the same switchover window, control commands are arranged according to the service priority in the tiered scheduling queue. If multiple service commands act on the same device, board, or port group, the execution times of these commands are staggered based on the signaling processing delay of the corresponding device to avoid multiple control commands arriving at the same resource at the same time.

[0141] Command conflict checking also requires comparing the resource locking interval, configuration issuance interval, and status confirmation interval of each service. When the commands of two services will not modify the same device resources simultaneously, partial parallel execution is allowed; when two services will modify the same board, port group, or cross-connection unit simultaneously, the relevant commands of the latter service should be executed after the former service has completed configuration and written back its status. After completing the command timing calculation, dependency check, and device-level conflict check for all services, all commands are arranged from earliest to latest according to their planned execution time; when the execution times are the same, they are further sorted according to service priority, command dependency level, and device locking order, thus obtaining the switchover timing control sequence.

[0142] Furthermore, based on the updated spatiotemporal resource reservation map of site addresses and board ports, the verified detour path set and the switchover timing control sequence are encapsulated into a structured maintenance ticket payload to obtain the detour execution and maintenance push payload. Specifically, the detour execution and maintenance push payload encapsulates at least the fault and maintenance identifier, Nx fault resources, affected services, service level, original primary / backup path, verified detour path, site addresses and physical resources along the route, physical common-source isolation, risk verification conclusion, bandwidth and port reservation, switchover time window, switchover timing control sequence, execution preconditions, success confirmation conditions, abnormal rollback instructions, maintenance work content, and service impact. In this process, for each control instruction, the payload records the instruction sequence number, planned execution time, target site address, target device, target board and port, operation type, operation parameters, preconditions, confirmation conditions, timeout time, and failure rollback instructions. Operation types include resource locking, cross-connection establishment, path verification, dual-path transmission, service switching, status confirmation, old path release, and abnormal rollback.

[0143] For verified detour paths, the payload records the sites, devices, boards, ports, optical paths, optical cable segments, and channels along the path in chronological order, and records the reserved bandwidth, reserved ports, reserved start time, reserved end time, and reserved status of each segment's resources. For paths that have undergone minor adjustments, it is also necessary to record the original non-compliant sub-paths, common-source risk resources, the replaced local paths, and changes in resource reservations.

[0144] After payload encapsulation, structural integrity and data consistency checks are performed. Structural integrity checks confirm that all required fields, such as service identifier, site address, device, board, port, path, time window, and control commands, are filled in, and that there are no missing execution objects or operation parameters. Data consistency checks confirm that the paths in the payload match the verified detour path set, and that the resource usage in the payload matches the updated spatiotemporal resource reservation. Figure 1 The control commands are executed within the corresponding failover window, and the order of commands—resource locking, path establishment, service switching, and resource release—is correct, with no circular dependencies. After successful verification, a unique execution identifier, version number, generation time, and integrity verification value are generated for the load, forming a detour execution load for the equipment control side and a maintenance push load for the maintenance management side. Specifically, the detour execution load drives the control system to complete resource locking, new path establishment, path verification, service switching, status confirmation, and old path release according to the failover sequence control sequence. The maintenance push load automatically fills the maintenance ticket and maintenance plan with faulty resources, maintenance work content, impact on services, recommended detour paths, physical common source risk verification results, execution window, and safeguards.

[0145] In summary, the communication service impact analysis and detour path recommendation method based on Nx risk verification according to the embodiments of this application is explained. It forms a spatiotemporal network topology by stitching the physical and logical layers together, identifies damaged services through fault pruning, and implements tiered scheduling based on service importance and latency sensitivity. It quantifies residual link risks through secondary fault injection and combines switching windows to conduct collaborative optimization of concurrent service multi-product flows, resolve resource conflicts, and verify physical common-source risks. This improves the completeness of damaged service identification, avoids redundant occupation of bandwidth, ports, and other resources, reduces secondary failures of detour paths and risks associated with shared cables and trenches, and forms an executable switching sequence and maintenance push load.

[0146] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for communication service impact analysis and detour path recommendation based on Nx risk verification, characterized in that, include: Step 1: Based on the original data of the multi-source communication network, perform cross-layer topology stitching and fault pruning on the physical layer and logical layer to obtain a spatiotemporal network topology map and a list of candidate damaged services. Step 2: Based on the two-dimensional decision matrix composed of business importance and latency sensitivity, priority mapping and tiered sorting are performed on each business in the list of damaged business candidates to obtain the tiered scheduling queue and the switching window scheme. Step 3: Simulation calculations are performed to inject secondary faults into the residual links in the spatiotemporal network topology diagram one by one to obtain the spatiotemporal topology impedance matrix. Step 4: Using the spatiotemporal topology impedance matrix as the link cost weight and the switching time window scheme as the timing relaxation constraint, perform multi-commodity flow collaborative optimization and resource conflict resolution on the spatiotemporal network topology graph for concurrent services in the tiered scheduling queue to obtain a conflict-free concurrent detour path set and an updated spatiotemporal resource reservation graph. Step 5: Based on the updated spatiotemporal resource reservation map, perform implicit physical common source risk verification on the set of conflict-free concurrent detour paths, and generate a detour timing control sequence according to the detour window scheme to obtain the detour execution and maintenance push load.

2. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 1, characterized in that, The raw data of the multi-source communication network includes optical cable physical segment ledger data, equipment board associated topology data, service logic routing mapping data, real-time service traffic snapshot data, and Nx fault target list data.

3. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 2, characterized in that, Step one includes: Based on the physical segment ledger data of optical cables, the associated topology data of equipment boards and the business logic routing mapping data, physical-logical topology cross-layer stitching is performed from the physical pipeline layer to the business channel layer to obtain a multi-layer composite topology map. The bandwidth utilization and transmission delay parameters in the real-time business traffic snapshot data are injected into a multi-layer composite topology graph and dynamically superimposed in spatiotemporal dimensions to obtain an initial spatiotemporal topology model. Based on the Nx fault target list data, topology pruning and erasure and damaged service traversal extraction are performed on the corresponding fault nodes and links in the initial spatiotemporal topology model to obtain the spatiotemporal network topology map and the candidate list of damaged services.

4. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 1, characterized in that, Step two includes: Normalized features are extracted and vectorized to assemble the service level, maximum tolerable latency, bandwidth requirements and protection type of each service in the candidate list of damaged services to obtain the service feature attribute matrix. Based on a two-dimensional decision matrix composed of the reciprocal of business importance and latency sensitivity, a two-dimensional decision matrix mapping and scheduling priority score calculation are performed on each business in the business feature attribute matrix to obtain a business priority score list. Based on the control plane signaling processing delay in the spatiotemporal network topology, the service priority score list is sorted in descending order and divided into staggered time window intervals to obtain the tiered scheduling queue and switching time window scheme.

5. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 1, characterized in that, Step three includes: Monte Carlo simulation calculations of secondary fault injection and topology connectivity changes were performed on the residual links and nodes in the spatiotemporal network topology diagram to obtain the simulation results of secondary faults. Based on the total bandwidth demand of all services in the tiered scheduling queue, the topology entropy increase and load collapse probability of the secondary fault simulation results are quantitatively evaluated to obtain the vulnerability gradient vector. Based on the physical delay and bandwidth utilization exponential penalty term, a multi-factor nonlinear topological impedance is synthesized from the vulnerability gradient vector and the state parameters of each link in the spatiotemporal network topology diagram to obtain the spatiotemporal topological impedance matrix.

6. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 1, characterized in that, Step four includes: Using the spatiotemporal topological impedance matrix as the cost weight and the switching time window scheme as the timing constraint, a multi-commodity flow source-destination mapping and optimization model is constructed on the spatiotemporal network topology graph for each service in the tiered scheduling queue to obtain a multi-commodity flow optimization model. The multi-commodity flow optimization model is solved by iterative step-by-step batching and concurrent collision resolution to obtain a set of conflict-free concurrent detour paths; Based on the bandwidth and port occupancy of each path in the conflict-free concurrent detour path set, the spatiotemporal network topology map is dynamically updated by reservation and deduction within the corresponding time window interval to obtain the updated spatiotemporal resource reservation map.

7. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 1, characterized in that, Step five includes: Based on the physical pipelines and co-channel ledgers in the updated spatiotemporal resource reservation map, implicit physical common-source isolation degree verification and local fine-tuning of illegal paths are performed on the primary and backup paths in the non-conflicting concurrent detour path set to obtain the verified detour path set. By combining the switching time window scheme with the delay difference between the old and new paths in the updated spatiotemporal resource reservation map, millisecond-level lossless switching timing offset calculation and instruction arrangement are performed on the verified detour path set to obtain the switching timing control sequence. Based on the site and board port mapping in the updated spatiotemporal resource reservation diagram, the verified detour path set and the switching timing control sequence are encapsulated into a structured maintenance ticket payload to obtain the detour execution and maintenance push payload.

8. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 3, characterized in that, Based on the Nx fault target list data, topology pruning and erasure, as well as damaged service traversal extraction, are performed on the corresponding fault nodes and links in the initial spatiotemporal topology model to obtain a spatiotemporal network topology map and a candidate list of damaged services, including: Based on the Nx fault target list data, the initial spatiotemporal topology model is constructed by fault nodes, fault pruning, and bipartite graphs of protection resource competition to obtain the spatiotemporal network topology graph and the bipartite graph of protection resource competition. To obtain the arbitration result of protection competition, a two-part diagram of resource competition is used to arbitrate protection capacity competition and identify pseudo-protection business. Based on the spatiotemporal network topology map and the arbitration results of protection competition, pseudo-protection interruption services and path direct interruption services are merged, and the service level and bandwidth parameters of each service are extracted for damage classification fusion and encapsulation to obtain the spatiotemporal network topology map and the candidate list of damaged services.

9. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 5, characterized in that, Based on the physical delay and bandwidth utilization exponential penalty term, a multi-factor nonlinear topological impedance synthesis is performed on the vulnerability gradient vector and the link state parameters in the spatiotemporal network topology diagram to obtain the spatiotemporal topological impedance matrix, including: Extract the physical transmission delay, current bandwidth utilization, and dynamic remaining available bandwidth of each residual link from the spatiotemporal network topology diagram; Based on the results of the secondary fault simulation, the entropy increase of the topology after the failure of each residual link is calculated, and the ratio of the total bandwidth demand of all services in the tiered scheduling queue to the dynamic remaining available bandwidth of the corresponding residual link is determined as the load pressure term. We weighted the entropy increase of the topology and the load pressure term to determine the secondary fault vulnerability gradient of each residual link; The physical transmission delay of each residual link, the exponential penalty value of the current bandwidth utilization, and the vulnerability gradient of the secondary fault are weighted and summed, and then the nodes corresponding to the residual links are matrix-assembled to obtain the spatiotemporal topological impedance matrix.

10. The communication service impact analysis and detour path recommendation method based on Nx risk verification according to claim 6, characterized in that, The multi-commodity flow optimization model is solved iteratively in batches and with concurrent collision resolution to obtain a set of conflict-free concurrent detour paths, including: Based on the switching time window scheme, the tiered scheduling queue is divided into multiple service sets that correspond one-to-one with the switching time window, and each service set is solved in batches iteratively according to the service priority from high to low. Each service is mapped to a commodity flow with source nodes, destination nodes and bandwidth requirements. The selection status of each service on the residual links is used as the decision variable. The optimization objective is to minimize the sum of the topological impedance of the residual links selected by each service. Flow conservation constraints, link remaining capacity constraints, node cross-connection capacity constraints and switching time window constraints are set. When the sum of the bandwidth usage of multiple services on the same residual link within the same handover window exceeds the dynamic remaining available bandwidth of the residual link, a capacity constraint multiplier is introduced to increase the cost weight of the residual link, and candidate paths are reselected for conflicting services based on the updated cost weight. Repeatedly perform cost weight updates and candidate path reselection until there is no excessive resource usage in any remaining links and nodes, in order to obtain a set of conflict-free concurrent detour paths.