A communication interface management method and electronic device

Abstract

The application discloses a communication interface management method and electronic equipment, relates to the technical field of communication interface management, and determines logical communication levels of multiple physical communication interfaces according to physical topological distribution of the multiple physical communication interfaces, selects logical communication interfaces of the multiple logical communication levels in turn from bottom to top by using the multiple physical communication interfaces, and generates a top-level logical communication interface until a top-level logical communication interface is generated; in response to the top-level logical communication interface receiving access data, determining an access source node and a physical topological relative distance between the access source node and multiple physical communication nodes, and selecting a physical communication interface closest to the access source node in the physical topological relative distance and in an active state as a target interface. The application can realize centralized management and dynamic selection according to topological relations of cross-node physical communication interfaces, improve communication continuity and path matching degree, and solve the problems of insufficient cross-node interface aggregation capability and difficulty in collaborative scheduling in a complex environment in the prior art.

Description

Technical Field

[0001] This application relates to the field of communication interface management technology, and in particular to a communication interface management method and electronic device. Background Technology

[0002] Network interface cards (NICs) are crucial input / output channels between servers and fundamental components used as physical communication nodes in distributed computing and data center networks. As applications shift from traditional business to massively parallel computing and artificial intelligence training, the demands on networks for high bandwidth, low latency, and high reliability have significantly increased, and some data transmission and reception processing is gradually being handled by the NICs.

[0003] To enhance network connectivity, related technologies typically aggregate multiple physical communication interfaces into a single external logical communication interface for redundancy and traffic distribution. However, such combination capabilities are often limited to a single physical device or a single network interface card (NIC), making it difficult to implement unified combination and scheduling of physical communication interfaces across different communication nodes. Furthermore, this limitation is amplified in cluster environments with multiple NICs, multiple nodes, and complex topologies. Summary of the Invention

[0004] This application provides a communication interface management method that enables centralized management and dynamic selection of physical communication interfaces across nodes based on topology, thereby improving communication continuity and path matching. This method effectively solves the problems of insufficient cross-node interface aggregation capability and difficulty in collaborative scheduling in complex environments in the prior art.

[0005] This application provides a communication interface management method, including: Receive multiple interface status information reported by multiple physical communication interfaces distributed across multiple physical communication nodes; Based on the physical topology distribution of multiple physical communication interfaces, determine the logical communication level of the multiple physical communication interfaces; Based on the logical communication hierarchy, multiple physical communication interfaces are selected from bottom to top to construct multiple logical communication interfaces for logical communication levels until the top-level logical communication interface is generated. In response to the top-level logical communication interface receiving access data, the access data is parsed to determine the access source node and the physical topological relative distance between the access source node and multiple physical communication nodes. Based on the physical topology relative distance and multiple interface status information, the physical communication interface that is closest to the access source node in terms of physical topology and is in an active state is selected as the target interface, so that the access data can be processed through the target interface.

[0006] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for implementing the steps of any of the above-described communication interface management methods when executing the computer program.

[0007] This application achieves centralized control over the operation of cross-node interfaces by acquiring the status information of physical communication interfaces distributed across multiple physical communication nodes, thus avoiding the problems of scattered interfaces and difficulty in unified scheduling in existing technologies. Furthermore, by constructing a bottom-up logical communication hierarchy based on the physical topology distribution, physical communication interfaces originally distributed across different nodes and levels can be logically unified in a hierarchical manner, which is beneficial for achieving cross-node link aggregation in complex topology environments. Moreover, since the top-level logical communication interface parses the access source node and determines its topological relative distance to each physical communication node after receiving access data, interface selection is not only based on interface status but also on the physical proximity of the access path. This allows for the selection of the most suitable physical communication interface as the target interface based on topological distance and interface activity status, resulting in a better actual forwarding path in multi-node, complex topology environments. In summary, this application achieves centralized management and dynamic selection of cross-node physical communication interfaces based on topological relationships, improving communication continuity and path matching while effectively solving the problems of insufficient cross-node interface aggregation capabilities and difficulty in collaborative scheduling in complex environments in existing technologies. Attached Figure Description

[0008] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A flowchart of a communication interface management method provided in an embodiment of this application; Figure 2 A schematic diagram of a communication interface management architecture provided for related technical embodiments; Figure 3 This application provides a schematic diagram of a communication interface management architecture. Figure 4 A schematic diagram of a communication interface management module provided in an embodiment of this application; Figure 5 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0010] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0011] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0012] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0013] This application provides a communication interface management method applicable to various computing devices and network systems that perform high-performance data communication under multiple physical communication interface conditions. It is particularly suitable for operating environments with multiple RDMA (Remote Direct Memory Access) physical communication interfaces and high requirements for communication stability, link reliability and real-time performance. For example, in systems such as data center servers, artificial intelligence training nodes, high-performance computing (HPC) clusters, distributed databases and distributed storage systems, cloud computing and virtualization platforms (such as virtualization environments based on SR-IOV or vRDMA), multiple RDMA network cards or multi-port RDMA devices are usually deployed to carry large-scale high-throughput, low-latency data exchange between nodes. The communication interface management method provided in this application can perform unified information management, operation status synchronization, and link switching control for multiple RDMA physical communication interfaces in the above-mentioned environment. It can be applied to scenarios such as RoCE (RDMA over Converged Ethernet), NVMe over Fabrics, distributed AI training frameworks (such as NCCL), object storage, and file system communication acceleration. When there are link jitter, port failures, or path congestion in the underlying network, the method of this application can realize the synchronization of communication records and seamless switching between member interfaces within the logical communication interface, thereby ensuring the continuity and stability of various high-performance communication tasks. Therefore, this application can be widely used in server communication systems with multiple RDMA channels, high-reliability network systems, and any application scenario that requires maintaining stable RDMA communication in a multi-link environment.

[0014] In one embodiment, such as Figure 1 As shown, this application provides a communication interface management method, including: Step 201: Receive multiple interface status information reported by multiple physical communication interfaces distributed across multiple physical communication nodes; Among them, physical communication nodes are used to represent devices that provide data transmission and reception capabilities, such as network interface cards (NICs). The system can contain multiple NICs, and each NIC participates in communication as an independent physical communication node; physical communication interfaces are used to represent NIC ports on physical communication nodes. Each NIC can contain multiple ports, such as Port0 and Port1, which serve as physical communication interfaces; interface status information is used to represent the operating status data reported by each NIC port, including but not limited to port link status (Up / Down), rate, error count, packet loss statistics, congestion information, and port enabled / disabled status; Step 202: Determine the logical communication level of the multiple physical communication interfaces based on the physical topology distribution of the multiple physical communication interfaces; Physical topology distribution refers to topological information that reflects the location and connection relationships of at least multiple physical communication nodes in the physical network, including but not limited to the cabinet / room / availability zone to which each physical communication node belongs, the connection level with the upper-level switching equipment, the physical link relationship between nodes, the number of path hops, the path bandwidth or path delay, etc., used to characterize the topological location and mutual distribution of multiple physical communication interfaces in the overall physical network. Step 203: Based on the logical communication hierarchy, select multiple physical communication interfaces from bottom to top to construct multiple logical communication interfaces for logical communication hierarchy until the top-level logical communication interface is generated. Among them, the logical communication interface refers to the virtualized and unified communication interface formed by aggregating multiple network interface card ports (physical communication interfaces). It can correspond to a link aggregation group (LAG) or form a distributed logical link across network interface cards, so that multiple ports can serve the upper layer as a unified entry point. The logical communication interface hides the heterogeneity of multiple ports at the bottom layer, and the upper layer system does not need to be aware of the specific physical ports. Step 204: In response to the top-level logical communication interface receiving access data, the access data is parsed to determine the access source node and the physical topology relative distance between the access source node and multiple physical communication nodes. Among them, the access source node refers to the node that sends access data, such as the client host, remote storage node or other network device. The access data carries information such as source address and session identifier, which can be used to identify the access source node. The physical topology relative distance refers to the distance or matching degree index obtained by comprehensively calculating the topology characteristic parameters such as the difference in physical connection level, path hops, path bandwidth index or path delay index between the access source node and each physical communication node based on the node topology mapping table or physical topology structure information. It is used to characterize the degree of proximity between the access source node and each physical communication node in the physical topology. Preferably, the smaller the value, the closer the topology distance. Step 205: Based on the physical topology relative distance and multiple interface status information, select the physical communication interface that is closest to the access source node in terms of physical topology relative distance and is in an active state as the target interface, so as to process the access data through the target interface. This embodiment provides a communication interface management method that, by acquiring the status information of physical communication interfaces distributed across multiple physical communication nodes, enables centralized control over the operation of cross-node interfaces, avoiding the problems of dispersed interfaces and difficulty in unified scheduling in existing technologies. Based on this, a bottom-up logical communication hierarchy is constructed according to the physical topology distribution, allowing physical communication interfaces originally distributed across different nodes and levels to be logically unified in a hierarchical manner, thus facilitating cross-node link aggregation in complex topology environments. Furthermore, since the top-level logical communication interface parses the access source node and determines its topological relative distance to each physical communication node after receiving access data, interface selection is not only based on interface status but also on the physical proximity of the access path. This allows for the selection of the most suitable physical communication interface as the target interface based on topological distance and interface activity status, achieving a better actual forwarding path in multi-node, complex topology environments. This realizes centralized management of cross-node physical communication interfaces and dynamic selection based on topological relationships, effectively solving the problems of insufficient cross-node interface aggregation capabilities and difficulty in collaborative scheduling in complex environments in existing technologies, while improving communication continuity and path matching.

[0015] In one embodiment, after receiving multiple interface status information reported by multiple physical communication interfaces distributed across multiple physical communication nodes, the method further includes: Parse multiple interface status information, identify multiple physical communication interfaces in the ready state, and register them in a pre-stored interface information table; Based on the multiple physical communication interfaces registered in the interface information table, multiple information storage areas are correspondingly divided in the preset storage space to store at least the interface status information and communication record information of the corresponding physical communication interfaces.

[0016] In an optional embodiment, after receiving interface status information reported by multiple physical communication interfaces, the status information is pre-parsed to identify the interface currently in the "ready state". The ready state may include, but is not limited to, the following: the link status of the interface is UP, the interface driver is loaded, the negotiation rate is normal, and there is no serious error count, thereby ensuring that the interface can perform data transmission and reception. All physical communication interfaces that meet the readiness conditions are registered in a pre-maintained interface information table. Then, based on the number of interfaces registered in the interface information table and their corresponding relationships, an independent information storage area is allocated for each physical communication interface in a pre-defined storage space that is either contiguous or logically divided. The information storage area is used to store at least the interface status information reported by the interface and subsequent communication record information, thereby ensuring that the interface-related data has a fixed storage location for easy retrieval later.

[0017] Specifically, in this embodiment, by parsing the interface status information and registering only interfaces in the ready state before partitioning the storage, unstable or unusable interfaces can be excluded at the beginning of management, reducing the storage resources occupied by invalid states. At the same time, by dividing the storage area based on the registered interface list, each physical communication interface has an independent information storage space, which enables the structured archiving of cross-node interface status information and communication records. This solves the problem of the difficulty in uniformly organizing multi-node interface information in the prior art, thereby providing a clear and stable data foundation for subsequent interface scheduling or status evaluation, and improving the manageability of physical communication interfaces in a multi-node environment.

[0018] In one embodiment, determining the logical communication layer of multiple physical communication interfaces based on their physical topology distribution includes: Obtain physical topology information containing at least multiple physical communication nodes, and determine the physical connection relationships between the multiple physical communication nodes; Based on physical connection relationships, determine the physical topology depth of physical communication nodes; Multiple physical communication interfaces connected to physical communication nodes at the same or adjacent physical topology depths are grouped into the same initial logical communication layer. Obtain the interface feature information of multiple physical communication interfaces, and adjust the initial logical communication layer of multiple physical communication nodes based on the interface feature information to generate the logical communication layer of multiple physical communication interfaces.

[0019] Specifically, in this embodiment, not only is the topology information of multiple physical communication nodes acquired, but a physical topology depth is also generated based on the physical connection relationships between nodes. This allows the system to accurately depict the positional relationships of each physical communication interface in the physical network, thus avoiding the problem in traditional solutions where the overall network structure cannot be reflected due to a single device perspective. Furthermore, physical communication interfaces with the same or adjacent physical topology depths are grouped into the same initial logical communication layer, enabling the logical layer to truly reflect physical proximity and path consistency, providing a clear and structured foundation for subsequent multi-level logical aggregation. Further, through... By adjusting the initial logical communication hierarchy based on interface feature information, the logical hierarchy can not only reflect the physical topology but also take into account the actual capabilities and business attributes of the interfaces, thereby constructing a more reasonable logical communication hierarchy system that is closer to the real network communication capabilities. Overall, this embodiment can realize hierarchical modeling of physical communication interfaces across nodes and paths, forming a structured logical communication hierarchy. This enables the system to have the prerequisites for subsequent execution of hierarchical aggregation, layer-by-layer routing, and path optimization, effectively improving the accuracy and flexibility of cross-node communication management and laying the foundation for dynamic interface scheduling and optimal path selection in complex topology environments.

[0020] In one exemplary embodiment, determining the logical communication layer of multiple physical communication interfaces based on the physical topology distribution of the multiple physical communication interfaces may include the following steps: Read topology information such as rack location, switching equipment level, node connection method, link bandwidth and physical hop count from the network controller, topology management module or preset topology description file; Based on this topology information, the physical connection relationships between multiple physical communication nodes are further determined, such as the ToR (Top of Rack) switches connected to the nodes, whether they belong to the same rack or the same data center, and the number of interconnection paths between the nodes. The physical topology depth of each physical communication node is determined based on this physical connection relationship. The physical topology depth can be defined according to the number of links or the difference in levels between the node and the core switching equipment or central node. For example, nodes connected to edge switches can be marked as having a lower depth, while nodes that span multiple switching levels have a higher depth. Multiple physical communication interfaces connected to physical communication nodes with the same or adjacent physical topology depths are grouped into the same initial logical communication level. For example, physical communication nodes in the same cabinet, connected to the same switch, or with a depth difference of 1 can have their corresponding multiple physical communication interfaces uniformly included in the same initial logical level to reflect physical adjacency and shared link characteristics. Obtain interface characteristic information for multiple physical communication interfaces. Interface characteristic information may include the interface's rate capability (e.g., 10G / 25G / 100G), actual available bandwidth, historical packet loss rate, congestion level, service priority, and the service type to which the interface belongs. Based on interface feature information, the initial logical communication layer of multiple physical communication nodes is adjusted. For example, if the bandwidth capacity of some interfaces is much higher than that of the interfaces of their neighboring nodes, or their service priorities are significantly different, they can be adjusted up or down from the current layer to generate a logical communication layer that better matches the actual communication capabilities and service attributes.

[0021] Through the above process, a logical communication hierarchy matching the physical topology distribution and interface capabilities of multiple physical communication interfaces is finally generated, providing an accurate hierarchical structure foundation for subsequent same-domain aggregation, mixed-domain aggregation, and layer-by-layer construction of logical communication interfaces.

[0022] In one embodiment, based on the logical communication hierarchy, multiple physical communication interfaces are selected sequentially from bottom to top to construct multiple logical communication layers of logical communication interfaces until the top-level logical communication interface is generated, including: Parse the status information of multiple interfaces to obtain the domain identification information of multiple physical communication interfaces; The communication domain to which multiple physical communication interfaces belong is determined based on the domain identification information; Select multiple physical communication interfaces that are at the same logical communication level and belong to the same communication domain as same-domain interfaces; Based on the preset first aggregation rule, at least two same-domain interfaces are selected as same-domain aggregation member interfaces. Based on the selected multiple same-domain aggregated member interfaces, a first logical communication interface is constructed, and the mapping relationship between the first logical communication interface and the multiple same-domain aggregated member interfaces is recorded. The constructed one or more first logical communication interfaces are merged with multiple physical communication interfaces other than the same-domain aggregation member interfaces to generate a candidate interface set for constructing the second logical communication interface.

[0023] Specifically, in this embodiment, by first dividing physical communication interfaces into the same communication domain based on domain identifiers, and then selecting interfaces within the same domain for combination based on the first aggregation rule, cross-device interfaces with similar physical locations and link conditions can be aggregated into a unified first logical communication interface without modifying the underlying topology. This not only breaks through the limitation of traditional interface aggregation being limited to a single network card or a single node, but also effectively reduces link imbalance forwarding caused by cross-domain differences. Furthermore, by recording the mapping relationship between logical communication interfaces and underlying physical interfaces and generating candidate interfaces that can continue to participate in aggregation, this embodiment enables logical communication interfaces to have hierarchical expansion capabilities, providing a structural foundation for subsequent cross-domain aggregation, cross-node aggregation, and the construction of higher-level logical links, thereby improving the system's interface organization capabilities, aggregation flexibility, and resource scheduling efficiency in complex topology environments.

[0024] In one embodiment, after generating a set of candidate interfaces for constructing the second logical communication interface, the method further includes: Get the candidate interface set; Based on the second aggregation rule, at least two candidate interfaces are selected from the candidate interface set as mixed-domain aggregation member interfaces, and the mixed-domain aggregation member interfaces include at least one first logical communication interface; Based on the selected mixed-domain aggregation member interface, a second logical communication interface corresponding to the upper logical communication layer of the first logical communication interface is constructed, and the mapping relationship between the second logical communication interface and the mixed-domain aggregation member interface is recorded. Remove the selected mixed-domain aggregation member interface from the candidate interface set, and add the constructed second logical communication interface to the candidate interface set; Based on the updated candidate interface set, the mixed-domain aggregation and logical communication interface construction are repeatedly executed until the top-level logical communication interface corresponding to the highest logical communication level is generated.

[0025] Specifically, through this embodiment, by performing mixed-domain aggregation on candidate interfaces after completing same-domain aggregation, the system can overcome the limitations of the communication domains (e.g., different racks, different switching partitions, or different availability zones) of physical communication interfaces. Interfaces distributed across multiple physical areas or cross-topology locations are combined layer by layer into higher-level logical communication interfaces, overcoming the limitations of aggregation based solely on the same-domain scope. This allows logical communication interfaces to expand across device boundaries and deployment areas, significantly improving the coverage and resource integration capabilities of logical communication links. Simultaneously, by requiring that each member interface in the mixed-domain aggregation contains at least one first logical communication interface, the aggregation process maintains a continuous hierarchy. This allows the logical communication structure to expand layer by layer from the bottom-level physical interfaces through same-domain aggregation to cross-domain levels, forming a complete multi-layered logical communication system. By recording the mapping relationship between the second logical communication interface and its mixed-domain aggregation member interfaces, not only is the aggregation structure maintained... The traceability of the structure provides a clear interface association basis for subsequent data forwarding, status monitoring, and fault switching. Furthermore, by updating the candidate interface after each mixed-domain aggregation and repeatedly executing the aggregation construction based on the updated candidate interface, the logical communication interface can be gradually expanded in multi-domain, multi-rack, or multi-availability zone environments until a top-level logical communication interface is formed. This provides a single, unified access point for upper-layer services, ensuring the dynamism and adaptability of the logical hierarchy construction. It enables the system to automatically adapt to topologies of different scales and depths, improving the overall scheduling capability and utilization efficiency of cross-node communication resources. In summary, this embodiment can realize the hierarchical construction of logical communication interfaces across domains and topologies, enabling the system to obtain a unified communication entry point with greater bandwidth aggregation capability, higher fault tolerance, and stronger scalability in complex topology environments. It solves the technical bottleneck in the prior art where logical communication interfaces cannot be aggregated across communication domains or multi-rack environments.

[0026] In one feasible embodiment, the process of constructing multiple logical communication interfaces from bottom to top based on logical communication hierarchy may include the following steps: The interface status information of multiple physical communication interfaces is parsed to extract the domain identification information carried by each interface. The domain identification information is used to identify the physical deployment domain to which the interface belongs, such as rack number, switching node area number, availability zone label, access switch number, local link domain ID, etc. Based on the domain identification information, multiple physical communication interfaces are classified according to their respective communication domains to distinguish physical interfaces that are in the same location, the same network node, or have similar topological relationships. Based on the above domain division, multiple physical communication interfaces belonging to the same communication domain are selected from the current lowest logical communication level as same-domain interfaces, and at least two interfaces are selected from the same-domain interfaces as same-domain aggregation member interfaces according to the first aggregation rule. Optionally, the first aggregation rule includes: the selected interfaces must simultaneously meet the following conditions: the link status is Up, the rate level is the same or the difference does not exceed a preset threshold, and the bandwidth remaining utilization rate is higher than 30%. For example, in the same rack environment, if Port0 and Port1 are both 25G links and are in a low-load state, they can be selected as same-domain aggregation member interfaces. Based on the selected same-domain aggregation member interface, a first logical communication interface is constructed, and the one-to-one mapping relationship between the logical communication interface and its underlying member interface is recorded for subsequent state backtracking and link selection. After construction is completed, the first logical communication interface is added to the candidate interface set for upper-layer aggregation, and the physical interfaces that have participated in aggregation are removed from the candidate interface set, so that the remaining interfaces and the newly generated logical interfaces can be used as the aggregation input for the next level. After the first logical communication interface is constructed, candidate interfaces are obtained, and mixed-domain aggregation is performed based on the second aggregation rule to construct a higher-level logical communication interface. The second aggregation rule can be: select at least one first logical communication interface from the candidate interfaces, and at least one physical or logical interface that does not belong to the same communication domain as the logical interface, and require that the average link bandwidth of the combined interfaces is not less than 70% of the bandwidth of the first logical communication interface. For example, the first logical communication interface (from rack A) is aggregated with another physical communication interface Port5 (from rack B, both with 25G bandwidth) to construct the second logical communication interface, thereby realizing cross-rack link integration. After constructing the second logical communication interface, add it to the candidate interface and delete the member interface of this round of aggregation to update the candidate interface set; Based on the updated candidate interface set, the mixed-domain aggregation and logical interface construction process is repeated until only one final logical aggregation interface remains among the candidate interfaces, that is, the top-level logical communication interface corresponding to the highest logical communication level is constructed as a unified external entry point. Ultimately, multiple physical communication interfaces distributed across different nodes, domains, and even racks will be aggregated layer by layer to form a unified logical entry point, enabling resource convergence and centralized scheduling capabilities.

[0027] In this embodiment, the system first completes intra-domain aggregation based on the first aggregation rule, and then completes cross-domain hybrid aggregation based on the second aggregation rule. This enables a layered construction approach for interfaces, from local integration to regional cross-regional integration, and finally to cluster unification. This removes the limitations of physical node location or network card isolation from logical communication interfaces. The first aggregation rule ensures that intra-domain interfaces are linked first, giving the aggregation starting point a natural advantage of low latency and high path consistency. The second aggregation rule enables logical interfaces to gradually expand across racks and regions, ultimately forming a three-dimensional aggregation topology, significantly improving the integration scale of multi-node communication resources. This embodiment not only realizes the hierarchical convergence of physical interfaces to a unified logical entry point, but also provides a data foundation and structural support for target link selection, load balancing, and link disaster recovery switching, which is an expansion capability that existing single-device aggregation mechanisms cannot cover.

[0028] In one embodiment, in response to the top-level logical communication interface receiving access data, the access data is parsed to determine the access source node and the physical topological relative distance between the access source node and multiple physical communication nodes, including: Parse the access data to obtain the access source address information, and determine the access source node based on the access source address information; Based on a node topology mapping table that includes at least multiple physical communication nodes and access source nodes, obtain the topology connection relationship between the access source node and multiple physical communication nodes from the node topology mapping table; Based on the topology connection relationship, determine the topology characteristic parameters of multiple physical communication nodes. The topology characteristic parameters include at least one of the following: the difference in physical connection level between nodes, the number of path hops, the path bandwidth index, or the path delay index. Based on topological feature parameters, topological distance indices are generated between the access source node and multiple physical communication nodes to characterize the physical topological relative distance between the access source node and multiple physical communication nodes.

[0029] Specifically, in this embodiment, after receiving access data at the top-level logical communication interface, the access source address information in the access data is first parsed to determine the access source node, thus clarifying the actual source device of the data request and laying the foundation for subsequent optimized scheduling based on the access path. Furthermore, the topological connection relationship between the access source node and multiple physical communication nodes is extracted based on the node topology mapping table, so that interface selection no longer depends solely on the interface's own state, but incorporates the physical path information between the access source and each physical node. Compared with the traditional scheduling method that only considers link load or interface state, this embodiment enables interface selection to comprehensively consider the physical topology factors of the access path, thus favoring the selection of interfaces that are closer to the access source, have shorter paths, or better bandwidth for data communication. This effectively reduces the additional latency costs caused by cross-rack and cross-switching layer access, and also reduces the bottleneck risk caused by path differences, improving the overall efficiency and stability of data forwarding. In summary, this embodiment, by introducing access source node identification and topology distance index calculation, enables the target interface selection process to have topology awareness capabilities, providing a more reasonable, efficient, and stable data access path in complex network environments, and significantly improving the system's communication performance in multi-node and multi-topology scenarios.

[0030] In one specific embodiment, based on topological feature parameters, topological distance indices are generated between the access source node and multiple physical communication nodes to characterize the physical topological relative distances between the access source node and the multiple physical communication nodes, including: The topology feature parameters corresponding to each physical communication node are organized and standardized to make different types of parameters (such as path hop count, connection level difference, path delay and path bandwidth) comparable under the same evaluation system. This process can be achieved by scaling, classifying or normalizing, so that different features can participate in distance evaluation in subsequent processes. Based on business needs or preset strategies, assign importance to each topology characteristic parameter. For example, for latency-sensitive services, path latency can be set as a key factor; for bandwidth-oriented services, the impact of path bandwidth can be increased. Based on these predefined levels of importance, different topological features are comprehensively considered, and a comprehensive topological evaluation value is generated for the access source node to each physical communication node. Specifically, the lower the evaluation value, the closer the access source node and the physical communication node are in physical topology, and the better the path conditions. The higher the evaluation value, the more network layers the path between the two crosses, the more hops, the higher the latency, or the weaker the bandwidth conditions. After obtaining the topology evaluation value, the topology distance index between the access source node and multiple physical communication nodes is further sorted to form a topology distance order from "closest" to "farthest". In one optional method, the topology distance is divided into several levels according to the comprehensive evaluation value, such as "close", "relatively close", "average", "relatively far", etc., to clarify the relative position differences of different nodes in the physical topology structure.

[0031] Specifically, in this embodiment, the above process can generate quantifiable topological distance indicators for the access source node and multiple physical communication nodes, so that the differences in network topology can be clearly reflected, thereby providing a reliable basis for the subsequent selection of target physical communication interfaces.

[0032] In one embodiment, based on the physical topology relative distance and multiple interface status information, the physical communication interface that is closest in physical topology distance to the access source node and is in an active state is selected as the target interface, including: If the logical communication level of the top-level logical communication interface is equal to the logical communication level of the first logical communication interface, then obtain multiple same-domain aggregated member interfaces corresponding to the first logical communication interface. Obtain and parse the interface status information corresponding to multiple same-domain aggregate member interfaces to obtain the interface status parameters of the same-domain aggregate member interfaces; The activity scores of multiple same-domain aggregated member interfaces are determined based on the interface status parameters, and the same-domain aggregated member interfaces with activity scores greater than the preset active status score threshold are identified as active interfaces. Obtain the physical topology relative distance between the physical communication node where multiple active interfaces are located and the access source node, and determine the active interface with the closest physical topology relative distance to the access source node as the target interface.

[0033] Specifically, in this embodiment, when the level of the top-level logical communication interface is the same as the logical level to which the first logical communication interface belongs, by obtaining the same-domain aggregated member interfaces corresponding to the first logical communication interface and parsing the status information of these interfaces, multiple physical interfaces within the same communication domain can be independently evaluated, thereby accurately identifying interfaces currently in an available or active state. Based on this, interfaces that are faulty, congested, or have abnormal loads can be excluded from the same-domain interface set, making the candidate objects more effective. Furthermore, by adjusting the relative physical topology distance between the physical communication nodes where these active interfaces are located and the access source node... By comparing data across nodes, the logical communication interface within the same domain can select the optimal interface from multiple candidate members for forwarding based on the actual network topology. This considers not only the interface's own operating status but also the topological proximity of the access path, allowing for a balance between state activity and topological distance within the same domain. Therefore, this embodiment can achieve more reasonable target interface selection among multiple network interface cards in the same rack, the same switching node, or the same location area, fully leveraging the advantages of same-domain aggregation. This improves the forwarding efficiency of data within the same domain, reduces additional path hops and latency, and significantly enhances communication stability and performance in same-domain access scenarios.

[0034] In one specific embodiment, the activity score of multiple same-domain aggregated member interfaces is determined based on interface status parameters, including: Get the interface status parameters corresponding to multiple same-domain aggregate member interfaces. The interface status parameters can be reported periodically or triggered by events by the physical communication interfaces of each physical communication node. The content includes, but is not limited to: link connection status, link speed, current bandwidth utilization, error packet count, packet loss rate, retransmission status, latency jitter, interface congestion level, and whether the interface is active or available. The interface status parameters of multiple aggregated member interfaces in the same domain are validated and standardized. Validation is used to determine whether the reported parameters are complete and whether there are outliers or missing data. Standardization is used to convert parameters with different dimensions and directions into a unified evaluation scale so that they can participate in the same scoring system. For example, negative indicators such as latency and error statistics are converted into indicators where the smaller the value, the higher the score, and positive indicators such as bandwidth utilization and link rate are converted into indicators where the larger the value, the higher the score. Based on the importance of multiple interface status parameters, preset weights are set for multiple interface status parameters, and the status element evaluation results of each same domain aggregate member interface are formed according to the correspondence between the weights and parameter values. The weights can be pre-configured by the system or dynamically adjusted according to the business type. For example, the influence weight of the latency parameter can be increased in business scenarios with high real-time requirements. The evaluation results of multiple factors of interface status parameters are comprehensively evaluated to obtain the status activity score of each intra-domain aggregate member interface. The comprehensive evaluation may include weighted summarization of multiple factors, threshold filtering, level conversion or segmented mapping, etc., so as to reflect the overall operation quality, stability and availability of each intra-domain aggregate member interface. The generated state activity score is used as the quality evaluation result of the corresponding same-domain aggregation member interface for subsequent interface selection, forwarding decision or master / slave switch logic. Optionally, same-domain aggregation member interfaces with higher state activity scores have higher priority in the target interface selection process and can be used as reliable candidate forwarding interfaces in mixed-domain aggregation or logical communication interfaces.

[0035] In one embodiment, the physical communication interface that is closest in physical topology relative distance to the access source node and is in an active state is selected as the target interface based on physical topology relative distance and multiple interface status information. The method further includes: If the logical communication level of the top-level logical communication interface is greater than that of the first logical communication interface, then obtain multiple mixed-domain aggregation member interfaces corresponding to the top-level logical communication interface. If the logical communication level of the mixed-domain aggregation member interface is greater than or equal to the logical communication level of the first logical communication interface, then the sub-interface of the mixed-domain aggregation member interface is obtained and it is determined whether the sub-interface is a physical communication interface. Therefore, based on the interface status information corresponding to the sub-interface, a corresponding sub-interface status activity score is generated. The status activity scores of multiple sub-interfaces are aggregated and statistically analyzed to generate the status activity score of the mixed-domain aggregated member interface. The scoring aggregation and statistical analysis is preferably obtained by weighted aggregation of multiple sub-interface status activity scores. The weight coefficients used can be obtained by linear mapping based on the historical failure frequency and importance of the sub-interface. The physical topology relative distances between the physical communication nodes where multiple sub-interfaces are located and the access source node are determined respectively, and the average value of the multiple physical topology relative distances is used to determine the physical topology relative distance between the mixed domain aggregation member interface and the access source node. If no response is received, the aggregated member interface of the next logical communication level of the sub-interface is obtained until the aggregated member interface is a physical communication interface. The state activity score of the aggregated member interface and the physical topology relative distance to the access source node are then traced back to the previous logical communication level to determine the state activity score of the logical communication interface of the previous logical communication level and the physical topology relative distance between the logical communication interface of the previous logical communication level and the access source node. Based on the physical topology relative distance and state activity score, the aggregate member interface that is closest to the access source node in physical topology relative distance and is in an active state is selected as the forwarding interface from top to bottom. If the forwarding interface is a physical communication interface, then the forwarding interface is determined as the target interface.

[0036] Specifically, in this embodiment, when selecting the target interface at the top-level logical communication interface, the mixed-domain aggregation member interfaces are parsed layer by layer and traced back to the bottom-level physical communication interface. This accurately obtains the true state and topological location of each physical interface in the cross-domain, multi-level aggregation structure. By weighted aggregation of the sub-interface states, the importance and stability of different sub-interfaces can be reflected when summarizing the logical interface states, making the state evaluation at the logical level more reliable. At the same time, by summarizing the topological distances corresponding to multiple sub-interfaces upwards, the logical communication interface has topology awareness. By selecting the optimal aggregation member interface from top to bottom based on state activity and topological distance, it can be ensured that the finally determined target physical communication interface is both active and topologically closest to the access source node. Therefore, this embodiment can achieve more accurate and efficient forwarding interface selection in complex multi-level aggregation scenarios, improving communication performance and stability under cross-domain access.

[0037] In one embodiment, after communicating the accessed data through the target interface, the method further includes: The information storage area corresponding to the target interface is defined as the first information storage area, and the interface status information stored in the first information storage area is defined as the first interface status information. The system periodically retrieves the first interface status information from the first information storage area and evaluates the operating status of the target interface based on the first interface status information. If one or more interface status parameters in the first interface status information do not meet the preset operating threshold, the target interface is determined to be operating abnormally. Obtain multiple physical communication interfaces that belong to the same logical communication interface as the target interface, select a backup interface from the multiple physical communication interfaces, and determine the information storage area corresponding to the backup interface as the second information storage area; The communication record information of the access data in the first information storage area is parsed to obtain the communication synchronization data; The communication synchronization data is synchronized to the second information storage area so that the backup interface can continue to process the access data based on the communication processing progress indicated by the communication synchronization data.

[0038] Specifically, in this embodiment, by periodically acquiring the interface status information of the target interface and evaluating its operational status, the system can promptly detect interface anomalies. When the target interface is determined to be operationally abnormal, a backup interface is selected from the members of the same logical communication interface, ensuring that subsequent communication continues through the same logical structure. Furthermore, communication record information is directly parsed from the first information storage area corresponding to the target interface and synchronized to the second information storage area corresponding to the backup interface. This allows the backup interface to continue subsequent data processing from the same communication processing progress as the target interface. This "direct synchronization within storage area" mechanism avoids the prior art method of retransmitting data over the network to restore the communication state, thereby reducing additional network burden and recovery latency. Simultaneously, this local or near-end record synchronization method ensures the continuity of communication state during cross-node interface switching, avoiding repetitive processing or state offsets caused by retransmissions, making interface switching smoother and more reliable. Therefore, it enables rapid fault-tolerant switching based on logical communication interfaces in a multi-node environment, resulting in higher continuity, lower recovery overhead, and better service availability for communication processing.

[0039] In a preferred embodiment, the information storage area corresponding to the target interface is periodically read to obtain the latest interface status information; based on the status information, it is determined whether the interface is operating abnormally, such as the status parameter being lower than a preset threshold, the error count surging, or the link jitter being severe. Once the target interface is determined to be abnormal, a backup interface is selected from the members of the same logical communication interface. To ensure the continuity of communication processing, the communication record information stored in the target interface's storage area is parsed, and the communication synchronization data (such as transmission sequence number, session status, incomplete transactions, etc.) is extracted and written to the information storage area corresponding to the backup interface. Since the synchronization process occurs directly between storage areas without re-forwarding historical data over the network, recovery latency can be significantly reduced, the additional overhead caused by repeatedly sending data can be avoided, and the backup interface can continue processing from the correct communication progress, thereby achieving faster and more stable interface switching.

[0040] In one embodiment, acquiring multiple physical communication interfaces that belong to the same logical communication interface as the target interface, and selecting a backup interface from the multiple physical communication interfaces, includes: Based on the maintained logical mapping relationship, determine the logical communication interface to which the target interface belongs, and obtain multiple physical communication interfaces associated with the logical communication interface. Among them, multiple physical communication interfaces are member interfaces in the same logical communication interface, and can all be used to carry the same type of access data. Remove the physical communication interfaces that are currently being used as target interfaces from a list of multiple physical communication interfaces, and combine the interface status information of each interface to determine the availability of the remaining physical communication interfaces. The availability determination includes, but is not limited to: whether the interface is in an enabled state, whether the link is normal, whether there are any faults or anomalies, and whether the preset performance threshold conditions are met. Priority evaluation is performed on multiple physical communication interfaces that meet availability conditions. The priority evaluation can be based on a comprehensive judgment of interface status information, historical communication record information and topology matching results. For example, physical communication interfaces with lower load, higher link quality or better topology relationship with the access source node are given priority. Among multiple candidate physical communication interfaces, the physical communication interface with the highest priority is determined as the backup interface, and the backup interface is established as a primary and backup interface with the target interface. This is recorded in the corresponding information storage area so that a fast switch can be performed when the target interface malfunctions or its performance degrades. When a target interface is detected to be faulty, the link is interrupted, or the communication quality requirements are not met, the access data is switched to the backup interface for subsequent communication processing by synchronizing the communication record information, thereby ensuring the continuity and reliability of the logical communication interface.

[0041] Specifically, this embodiment manages multiple physical communication interfaces within the same logical communication interface in a unified manner, and introduces a comprehensive evaluation mechanism based on interface activity and relative physical topology distance when selecting backup interfaces. This upgrades the selection of backup interfaces from the traditional "anything that works" to "preferential selection." This mechanism not only effectively avoids using interfaces with poor link quality, excessive load, or suboptimal topology locations as backup interfaces, but also ensures that backup interfaces have better carrying capacity in terms of path distance and communication performance during switching. By pre-establishing a primary / backup relationship between the target interface and the backup interface, a fast and smooth data switch can be achieved when the target interface malfunctions, without re-establishing the logical channel or interrupting upper-layer services, thereby significantly improving system stability and continuity. In summary, this embodiment enables more refined and intelligent management of physical communication resources within the logical communication interface, enhances the reliability of redundancy switching, and further improves the service quality and overall performance of the communication system in complex network environments.

[0042] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0043] It is worth noting that in a related technology, the communication interface management architecture is as follows: Figure 2As shown, in this architecture, multiple physical communication interfaces within each physical communication node can be aggregated into a logical communication interface through Link Aggregation (LAG) or Bond technology to improve link bandwidth or enhance link reliability. For example, in physical communication node 1, physical communication interface 1 and physical communication interface 2 can be combined into a logical communication interface through a local aggregation mechanism, forming a logical link in aggregation network device 1; in physical communication node 2, physical communication interface 3 and physical communication interface 4 can also be combined into another logical communication interface through a similar mechanism, forming a logical link in aggregation network device 2.

[0044] However, the link aggregation of the above technologies is limited to the same network card or the same physical communication node. Its aggregation boundary is limited by the management scope of the network card hardware or operating system driver. The physical communication interfaces between different physical nodes cannot form a unified logical communication interface across nodes and network cards. There is also no global unified scheduling and information synchronization mechanism between the logical communication interfaces. Therefore, when the cross-node communication link is abnormal, the existing technology cannot transparently switch based on the logical communication interface across network cards and nodes. Instead, it can only rely on the local LAG mechanism inside the node or re-establish the connection, making it difficult to achieve reliable global link management in multi-node collaborative or distributed systems.

[0045] In contrast, the embodiments of this application provide a communication interface management architecture such as Figure 3 As shown, unlike existing technologies that only support link aggregation within nodes, this application introduces a communication interface management module into the system to centrally manage multiple physical communication interfaces across physical communication nodes and network cards, and to build a unified logical communication interface (link group, LAG) based on these interfaces. The communication interface management module can uniformly obtain the physical communication interface status, configuration information and synchronization data of different nodes, and aggregate multiple physical communication interfaces from different nodes into a global logical communication interface according to the logical hierarchy, providing a unified communication entry point for the upper layer. When a physical communication interface in the logical communication interface fails, the management module can switch the communication to other available physical communication interfaces based on the membership relationship and real-time synchronization information within the logical communication interface, realizing transparent switching across nodes and network cards. Through the architecture of this application, the system has global logical communication interface management capabilities, which can significantly improve the stability and fault tolerance of multi-node, multi-link RDMA communication.

[0046] like Figure 4 As shown, embodiments of this application also provide a communication interface management module, including: an interface status read / write unit, a communication record read / write unit, a status synchronization parsing unit, an interface switching control unit, and a logical relationship management unit, wherein: This interface status read / write unit is used to read and write the operating status information of the physical communication interface. This unit obtains the interface status (e.g., link status, enable status, error flags, etc.) from the underlying interface driver or hardware device and writes the obtained status data to the interface status information storage area in the storage space. Simultaneously, this unit can also read status data from the interface status information storage area to support subsequent processing by this module or other modules. The communication record read / write unit is used to retrieve communication record information generated by the physical communication interface when receiving and processing access data from the storage space. The communication record information is the record data generated during the physical communication interface's RDMA access or other communication operations, including key communication records generated during the interface's processing of access requests, response messages, and receiving data packets. This unit reads the communication record information in the information storage area to provide basic data for subsequent parsing and synchronization operations. The state synchronization parsing unit is used to determine state synchronization data based on communication record information. This unit parses the communication records obtained by the communication record reading and writing unit, extracts information such as protocol fields, sequence numbers, session identifiers, and processing results from the records, and generates state synchronization data for multi-interface state synchronization. The state synchronization data represents the current processing progress of the accessed data and is used for subsequent state updates and interface switching. The interface switching control unit is used to write status synchronization data into the information storage area of ​​the selected backup interface in the storage space to realize the synchronization maintenance and interface switching control within the logical communication interface. When a physical communication interface malfunctions or its performance degrades, this unit can select other available physical communication interfaces within the logical communication interface and perform a switching operation based on communication records and communication status data. The logical relationship management unit is used to manage the logical hierarchy between multiple physical communication interfaces, build and maintain the membership relationship within the logical communication interface (or link group LAG). According to system configuration or logical topology requirements, this unit aggregates multiple physical communication interfaces into one logical communication interface, and maintains the mapping relationship, priority information and switching strategy between the logical communication interface and each physical communication interface, providing structured logical relationship support for the interface switching control unit.

[0047] In one specific embodiment, the communication interface management module interacts with a storage space, which contains multiple information storage areas for storing interface status data, communication record data, and intermediate data required during the synchronization phase. For example, the storage space includes: an interface status information storage area for storing the operating status information of each physical communication interface; a communication record information storage area for storing the communication record information generated by the interface during the reception and processing of access data; and other information storage areas for storing communication synchronization data generated by the status synchronization parsing unit or logical topology data maintained by the logical relationship management unit. The communication interface management module uses these information storage areas to achieve status collection, record acquisition, synchronization data storage, and information interaction related to interface switching.

[0048] like Figure 5 As shown, embodiments of this application also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above-described communication interface management method embodiments.

[0049] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0050] The communication interface management method and electronic device provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only intended to help understand the method and core ideas of this application. It should be noted that those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A communication interface management method, characterized in that, include: Receive multiple interface status information reported by multiple physical communication interfaces distributed across multiple physical communication nodes; Based on the physical topology distribution of the multiple physical communication interfaces, the logical communication layers of the multiple physical communication interfaces are determined; Parse the status information of multiple interfaces to obtain the domain identification information of multiple physical communication interfaces; The communication domain to which the multiple physical communication interfaces belong is determined based on the domain identification information; Multiple physical communication interfaces that are at the same logical communication level and belong to the same communication domain are selected as same-domain interfaces. At least two same-domain interfaces that are physically close and have similar link conditions are selected as same-domain aggregation member interfaces. Based on the selected multiple same-domain aggregation member interfaces, a first logical communication interface is constructed, and the constructed first logical communication interface is added to the candidate interface set. Select at least one first logical communication interface from the candidate interface set, and at least one physical or logical interface that does not belong to the same communication domain as the selected first logical communication interface as a mixed domain aggregation member interface. Based on the mixed-domain aggregation member interface, a second logical communication interface corresponding to the upper logical communication level of the first logical communication interface is constructed. The mixed-domain aggregation and logical communication interface construction are repeatedly executed until the top-level logical communication interface corresponding to the highest logical communication level is generated. In response to the top-level logical communication interface receiving access data, the access data is parsed to determine the access source node and the physical topological relative distance between the access source node and the multiple physical communication nodes; Based on the physical topology relative distance and the status information of multiple interfaces, the physical communication interface that is closest to the physical topology relative distance of the access source node and is in an active state is selected as the target interface, so as to perform communication processing on the access data through the target interface.

2. The communication interface management method according to claim 1, characterized in that, After receiving multiple interface status information reported by multiple physical communication interfaces distributed across multiple physical communication nodes, the method further includes: Parse the status information of multiple interfaces, determine the multiple physical communication interfaces that are in the ready state, and register them in the pre-stored interface information table; Based on the multiple physical communication interfaces registered in the interface information table, multiple information storage areas are correspondingly divided in the preset storage space to store at least the interface status information of the corresponding physical communication interfaces.

3. The communication interface management method according to claim 1, characterized in that, The step of determining the logical communication layer of the multiple physical communication interfaces based on the physical topology distribution of the multiple physical communication interfaces includes: Obtain physical topology information containing at least a plurality of the physical communication nodes, and determine the physical connection relationships between the plurality of physical communication nodes; Based on the physical connection relationship, the physical topology depth of the physical communication node is determined; Multiple physical communication interfaces connected to physical communication nodes at the same or adjacent physical topology depths are assigned to the same initial logical communication layer. The interface feature information of multiple physical communication interfaces is obtained, and the initial logical communication layer of multiple physical communication nodes is adjusted based on the interface feature information to generate the logical communication layer of multiple physical communication interfaces.

4. The communication interface management method according to claim 1, characterized in that, After constructing the first logical communication interface and before adding the constructed first logical communication interface to the candidate interface set, the process includes: Record the mapping relationship between the first logical communication interface and the multiple same-domain aggregate member interfaces; The constructed one or more first logical communication interfaces are merged with multiple physical communication interfaces other than the same-domain aggregated member interfaces to generate a candidate interface set for constructing the second logical communication interface.

5. The communication interface management method according to claim 1, characterized in that, Prior to the repeated execution of mixed-domain aggregation and logical communication interface construction, the method further includes: Record the mapping relationship between the second logical communication interface and the mixed-domain aggregation member interface; Remove the selected mixed-domain aggregation member interface from the candidate interface set, and add the constructed second logical communication interface to the candidate interface set; The construction of the repeated execution of mixed-domain aggregation and logical communication interface includes: Based on the updated candidate interface set, the mixed-domain aggregation and logical communication interface construction are repeatedly executed until the top-level logical communication interface corresponding to the highest logical communication level is generated.

6. The communication interface management method according to claim 1, characterized in that, In response to receiving access data at the top-level logical communication interface, the access data is parsed to determine the access source node and the physical topological relative distance between the access source node and the plurality of physical communication nodes, including: The access data is parsed to obtain the access source address information, and the access source node is determined based on the access source address information; Based on a node topology mapping table that includes at least a plurality of physical communication nodes and the access source node, the topological connection relationship between the access source node and the plurality of physical communication nodes is obtained from the node topology mapping table; Based on the topology connection relationship, the topology characteristic parameters of multiple physical communication nodes are determined. The topology characteristic parameters include at least one of the following: the difference in physical connection level between nodes, the number of path hops, the path bandwidth index, or the path delay index. Based on the topology feature parameters, topology distance indices are generated between the access source node and the multiple physical communication nodes to characterize the physical topology relative distance between the access source node and the multiple physical communication nodes.

7. The communication interface management method according to claim 1, characterized in that, Based on the physical topology relative distance and multiple interface status information, the physical communication interface that is closest in physical topology relative distance to the access source node and is in an active state is selected as the target interface, including: In response to the logical communication level of the top-level logical communication interface being equal to the logical communication level of the first logical communication interface, multiple of the same-domain aggregated member interfaces corresponding to the first logical communication interface are obtained. Obtain and parse the interface status information corresponding to multiple same-domain aggregation member interfaces to obtain the interface status parameters of the same-domain aggregation member interfaces; Based on the interface status parameters, the status activity scores of multiple same-domain aggregated member interfaces are determined, and the same-domain aggregated member interfaces whose status activity scores are greater than a preset active status score threshold are determined as active interfaces. The physical topology relative distances between the physical communication nodes where the multiple active interfaces are located and the access source node are obtained respectively, and the active interface with the closest physical topology relative distance to the access source node is determined as the target interface.

8. The communication interface management method according to claim 1, characterized in that, The step of selecting the physical communication interface that is closest in physical topology relative distance to the access source node and is in an active state as the target interface based on the physical topology relative distance and multiple interface status information further includes: In response to the fact that the logical communication level of the top-level logical communication interface is greater than the logical communication level of the first logical communication interface, multiple mixed-domain aggregation member interfaces corresponding to the top-level logical communication interface are obtained. In response to the logical communication level of the mixed-domain aggregation member interface being greater than or equal to the logical communication level of the first logical communication interface, the sub-interface of the mixed-domain aggregation member interface is obtained and it is determined whether the sub-interface is the physical communication interface. In response, a corresponding sub-interface status activity score is generated based on the interface status information corresponding to the sub-interface. The status activity scores of multiple sub-interfaces are aggregated and statistically analyzed to generate the status activity score of the mixed-domain aggregate member interface. The physical topology relative distances between the physical communication nodes where the multiple sub-interfaces are located and the access source node are determined respectively, and the average value of the multiple physical topology relative distances is determined as the physical topology relative distance between the mixed domain aggregation member interface and the access source node; If no response is received, the aggregated member interface of the next logical communication layer of the sub-interface is obtained until the aggregated member interface is the physical communication interface. The state activity score of the aggregated member interface and the physical topology relative distance with the access source node are backtracked to the previous logical communication layer to determine the state activity score of the logical communication interface of the previous logical communication layer and the physical topology relative distance between the logical communication interface of the previous logical communication layer and the access source node. Based on the physical topology relative distance and the state activity score, the aggregate member interface that is closest to the access source node in physical topology relative distance and is in an active state is selected from top to bottom as the forwarding interface. If the forwarding interface is the physical communication interface, then the forwarding interface is determined as the target interface.

9. The communication interface management method according to claim 2, characterized in that, After the communication processing of the access data through the target interface, the method further includes: The information storage area corresponding to the target interface is determined as the first information storage area, and the interface status information stored in the first information storage area is determined as the first interface status information. The first interface status information is periodically retrieved from the first information storage area, and the operating status of the target interface is evaluated based on the first interface status information. If one or more interface status parameters in the first interface status information do not meet the preset operating threshold, the target interface is determined to be operating abnormally. Obtain multiple physical communication interfaces that belong to the same logical communication interface as the target interface, select a backup interface from the multiple physical communication interfaces, and determine the information storage area corresponding to the backup interface as the second information storage area; The communication record information of the access data in the first information storage area is parsed to obtain communication synchronization data; The communication synchronization data is synchronized to the second information storage area so that the backup interface can continue to process the access data based on the communication processing progress indicated by the communication synchronization data.

10. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the communication interface management method as described in any one of claims 1 to 9 when executing the computer program.

Images