A method and system for coordinated communication of a distributed switch cluster

Abstract

The application provides a kind of collaborative communication method and system of distributed switch cluster, it is related to communication technical field, the application is by first collecting the active data stream information of switching unit in distributed switch cluster to generate global flow table, the utilization rate data of multistage interconnection link is monitored to construct the network state information of the distributed switch cluster, then based on this view, identify congestion area to generate key service flow path optimization scheme and redundant path scheduling strategy, subsequently respectively through centralized control process to generate flow table instruction adapted to redundant strategy, through distributed control process to generate local path decision and are divided into the target switching unit in the distributed switch cluster, finally coordinate execution instruction set and decision result and dynamically adjust service flow path allocation, can realize the collaborative operation of centralized control and distributed execution of distributed switch cluster, improve the efficiency and stability of cluster communication.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a collaborative communication method and system for a distributed switch cluster. Background Technology

[0002] In scenarios such as data centers and high-performance computing, distributed switch clusters are the core infrastructure for carrying massive data exchange. The efficiency of their traffic scheduling and collaborative communication directly determines the service quality and operational performance of the entire network.

[0003] In existing technologies, a mainstream traffic scheduling method employs a fully centralized control model, such as the typical Software-Defined Networking (SDN) architecture. This method uses a central controller to collect state information from the entire network and calculates and distributes forwarding table entries for all data flows in the network. This approach can achieve globally optimal path planning, but it relies on the controller possessing sufficiently powerful processing capabilities and real-time response capabilities.

[0004] However, in environments with complex business types and highly dynamic network traffic, the aforementioned fully centralized control method exposes its inherent shortcomings. The central controller needs to indiscriminately process path calculation requests for all business flows (including high-priority critical services and low-priority non-critical services). When the network scales up or traffic surges, the controller easily becomes a performance bottleneck, leading to delayed responses to dynamic changes such as network congestion, and failing to simultaneously ensure path protection for critical services and flexibly manage ordinary services. Therefore, existing technologies face the technical challenge of improving overall network resource utilization and dynamic adaptability while ensuring the quality of service for critical services. Summary of the Invention

[0005] The purpose of this application is to provide a collaborative communication method and system for a distributed switch cluster, in order to solve the problem in the prior art that while ensuring the quality of critical business services, it is impossible to improve the overall network resource utilization and dynamic adaptability.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a collaborative communication method for a distributed switch cluster, comprising:

[0007] Collect active data flow information of all switching units in the distributed switch cluster to generate a global flow table, and monitor the utilization data of multi-level interconnection links in the distributed switch cluster to construct the network status information of the distributed switch cluster.

[0008] Based on the network status information, congested areas are identified, and path optimization schemes for critical service flows are generated according to the identification results. A corresponding redundant path scheduling strategy is then established based on the path optimization schemes.

[0009] The global flow table and the path optimization scheme are parsed through a centralized control process. Based on the parsing results, the optimal forwarding path of the critical service flow is calculated, flow table instructions adapted to the redundant path scheduling strategy are generated, and the flow table instructions are distributed to the target switching unit in the distributed switch cluster.

[0010] The local link status information of the switching unit is obtained through a distributed control process. The local link status information and the utilization data are combined to perform dynamic path selection for non-critical service flows, generate local path decisions, and distribute the local path decisions to the relevant switching units in the distributed switch cluster.

[0011] The system coordinates the execution of the flow table instructions and the local path decisions, and dynamically adjusts the allocation of service flow paths among multi-level switching units according to network status, thereby achieving coordinated operation of centralized control and distributed execution.

[0012] Optionally, the step of identifying congested areas based on the network state information, generating path optimization schemes for critical service flows based on the identification results, and establishing corresponding redundant path scheduling strategies based on the path optimization schemes includes:

[0013] Analyze the utilization data of multi-level interconnected links in the network status information, identify link segments whose utilization rate continuously exceeds a predetermined threshold, and mark the identified high-utilization link segments as congested areas;

[0014] Identify all critical business flows passing through the congested area, analyze the service quality requirements and timeliness constraints of the critical business flows, and generate path optimization schemes based on the analysis results;

[0015] Based on the backup path information determined in the path optimization scheme, a corresponding redundant path scheduling strategy is established. The redundant path scheduling strategy includes path switching conditions based on network state changes, execution timing requirements to ensure service continuity, and fault handling mechanisms to deal with emergencies.

[0016] Optionally, the step of parsing the global flow table and the path optimization scheme through a centralized control process, calculating the optimal forwarding path for critical service flows based on the parsing results, generating flow table instructions adapted to the redundant path scheduling strategy, and distributing the flow table instructions to the target switching units in the distributed switch cluster includes:

[0017] The global flow table is parsed to obtain the feature information of the key business flow. The service quality target and path selection constraints of the key business flow are determined by combining the feature information of the key business flow and the optimization requirements defined in the path optimization scheme.

[0018] Based on the service quality objectives and path selection constraints, and combined with the current network topology and utilization data, the optimal forwarding path for the critical service flow in the distributed switch cluster is calculated.

[0019] Based on the calculated optimal forwarding path and the path switching rules defined in the redundant path scheduling strategy, flow table instructions are generated. The flow table instructions contain detailed forwarding rules for the primary path and backup path, and the flow table instructions are distributed to the target switching units in the distributed switch cluster.

[0020] Optionally, the step of obtaining local link status information of the switching unit through a distributed control process, combining the local link status information with the utilization data to perform dynamic path selection for non-critical service flows, generating local path decisions, and distributing the local path decisions to relevant switching units in the distributed switch cluster includes:

[0021] The local link status information of each switching unit is obtained through a distributed control process. The local link status information includes link utilization, transmission delay and available bandwidth data.

[0022] The local link status information is correlated with the utilization data to generate a comprehensive link quality assessment result. Based on the comprehensive link quality assessment result and the transmission characteristic requirements of non-critical service flows, dynamic path selection is implemented for non-critical service flows.

[0023] Based on the results of dynamic path selection, a local path decision containing path selection strategy and forwarding rules is generated to ensure that the local path decision is coordinated with the flow table instructions, and the local path decision is distributed to the relevant switching units in the distributed switch cluster.

[0024] Optionally, the step of obtaining local link status information of each switching unit through a distributed control process includes link utilization, transmission delay, and available bandwidth data, including:

[0025] A data acquisition agent is deployed on the switching unit through a distributed control process to establish a communication connection with the controller;

[0026] The data acquisition agent periodically collects local link operation status data, including the number of communication data packets transmitted, latency measurement information, and bandwidth usage. The collected operation status data is preprocessed to extract link utilization indicators, transmission delay indicators, and available bandwidth indicators.

[0027] The extracted indicator data are standardized in format and combined to generate structured local link status information, which is then transmitted to the controller through the established communication connection.

[0028] The received local link status information is stored in a distributed database for use in subsequent path decision-making processes.

[0029] Optionally, the coordinated execution of the flow table instructions and the local path decision, dynamically adjusting the allocation of service flow paths among multi-level switching units according to network status, to achieve coordinated operation of centralized control and distributed execution, includes:

[0030] Receive the flow table instruction and the local path decision, and coordinate the flow table instruction and the local path decision;

[0031] Based on network conditions, the allocation of service flow paths among multi-level switching units is dynamically adjusted to prioritize the quality of service requirements of critical service flows while optimizing the transmission efficiency of non-critical service flows. This ensures that critical service flows are forwarded strictly according to the paths specified by the flow table instructions, while non-critical service flows flexibly select paths based on the local path decision to avoid path conflicts.

[0032] Monitor the effectiveness of path adjustments, optimize the decision-making process based on feedback, and achieve coordinated operation of centralized control and distributed execution.

[0033] Optionally, the step of collecting active data flow information from all switching units in the distributed switch cluster to generate a global flow table, and monitoring the utilization data of multi-level interconnection links within the distributed switch cluster to construct network status information for the distributed switch cluster, includes:

[0034] Periodically collect active data flow information of local forwarding from each switching unit in the distributed switch cluster, including the source address, destination address and service type information of the data packets, and integrate and process the collected active data flow information;

[0035] Extract the characteristic parameters of each active data stream, including traffic size, duration and quality of service requirements, and generate a unified global flow table based on the extracted characteristic parameters;

[0036] The utilization data of each link on the multi-level interconnection link is continuously monitored, the bandwidth usage and load status of each link are recorded, and the utilization data is associated with and stored in the global flow table.

[0037] Based on the correlated data, a network status information system is constructed that includes information on the distribution of data flows across the entire network and the link load status. This network status information reflects the network's operational status.

[0038] Secondly, this application provides a cooperative communication system for a distributed switch cluster, comprising:

[0039] The construction module is used to collect active data flow information of all switching units in the distributed switch cluster to generate a global flow table, and to monitor the utilization data of multi-level interconnection links in the distributed switch cluster to construct the network status information of the distributed switch cluster.

[0040] A module is established to identify congested areas based on the network status information, generate path optimization schemes for critical service flows based on the identification results, and establish corresponding redundant path scheduling strategies based on the path optimization schemes.

[0041] The distribution module is used to parse the global flow table and the path optimization scheme through a centralized control process, calculate the optimal forwarding path of the critical service flow based on the parsing result, generate flow table instructions that are compatible with the redundant path scheduling strategy, and distribute the flow table instructions to the target switching unit in the distributed switch cluster.

[0042] The second distribution module is used to obtain the local link status information of the switching unit through a distributed control process, combine the local link status information with the utilization data to perform dynamic path selection for non-critical service flows, generate local path decisions, and distribute the local path decisions to the relevant switching units in the distributed switch cluster.

[0043] The coordination module is used to coordinate the execution of the flow table instructions and the local path decisions, and dynamically adjust the allocation of service flow paths among multi-level switching units according to the network status, so as to realize the coordinated operation of centralized control and distributed execution.

[0044] Thirdly, this application provides an electronic device, comprising:

[0045] Memory, used to store computer programs;

[0046] A processor, used to execute the computer program to implement the steps of the cooperative communication method for a distributed switch cluster as described in the first aspect above.

[0047] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the cooperative communication method for a distributed switch cluster as described in the first aspect above.

[0048] The collaborative communication method for distributed switch clusters provided in this application generates a global flow table by collecting active data flow information from all switching units in the distributed switch cluster, and constructs network status information of the distributed switch cluster by monitoring the utilization data of multi-level interconnection links within the distributed switch cluster. This provides comprehensive and accurate basic data support for subsequent path optimization and decision-making, enabling a clear understanding of the overall network status. By identifying congested areas based on network status information, generating path optimization schemes for critical service flows, and establishing redundant path scheduling strategies, congestion risks can be avoided in advance, ensuring the stability of critical service flow transmission and improving the network's ability to cope with congestion. The method also calculates the optimal path for critical service flows by centrally controlling the parsing of the global flow table and path optimization schemes. Forwarding paths and generating appropriate flow table instructions for distribution to target switching units in the distributed switch cluster ensures that critical service flows are forwarded along the optimal path, meeting their high-priority requirements and achieving precise management of critical services. By acquiring local link status information of switching units through a distributed control process and combining it with utilization data, dynamic path selection is performed on non-critical service flows, generating local path decisions and distributing them. This flexibly adapts to the needs of non-critical service flows, improves local resource utilization, and enhances the network's responsiveness to local changes. By coordinating the execution of flow table instructions and local path decisions, and dynamically adjusting service flow path allocation based on network status, the coordinated operation of centralized control and distributed execution is achieved, optimizing overall service flow path allocation and improving cluster communication efficiency and stability.

[0049] Furthermore, the generation process for congestion hotspot identification, critical service flow path optimization schemes, and redundant path scheduling strategies has been refined, as well as the process of calculating the optimal path for critical service flows and generating and distributing flow table instructions during centralized control. These refined steps make congestion hotspot identification more accurate, critical service flow path optimization more aligned with actual needs, and redundancy strategies more operable. At the same time, they enable more refined management of critical service flows by centralized control, improving the accuracy and effectiveness of critical service flow path planning and execution, and ensuring the stable transmission of critical service flows in complex network environments. Attached Figure Description

[0050] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0051] Figure 1 A flowchart illustrating a collaborative communication method for a distributed switch cluster provided in an embodiment of this application;

[0052] Figure 2 This is a schematic diagram of the structure of a collaborative communication system for a distributed switch cluster provided in an embodiment of this application. Detailed Implementation

[0053] In view of this, the main objective of the present invention is to provide a collaborative communication method for a distributed switch cluster, so as to solve the performance bottleneck caused by completely centralized control in the prior art, and the technical problem that it is difficult to balance ensuring the quality of critical business services and improving the overall network resource utilization.

[0054] To achieve the above objectives, this invention provides an innovative collaborative communication method. The core idea of ​​this method is to decouple control tasks, employing a hybrid mode combining centralized control and distributed execution. Specifically, through a centralized control process, a macroscopic analysis of the overall network state is performed, pre-calculating and planning optimal forwarding paths and redundant backup paths specifically for high-priority critical service flows, generating clearly guiding flow table instructions. This ensures that the service quality of critical services receives the highest priority. Simultaneously, for a large number of low-priority non-critical service flows, a distributed control process authorizes each switching unit to perform real-time, dynamic path selection based on its own local link status.

[0055] In this way, the present invention significantly reduces the burden on the central controller, allowing it to focus on ensuring the stability and performance of critical services, thereby avoiding performance bottlenecks. Simultaneously, decentralizing path selection for non-critical services grants network edge nodes greater autonomy and flexibility, enabling them to quickly adapt to changes in local network load, thus significantly improving the overall utilization of network resources and the ability to adapt to dynamic traffic. Ultimately, the present invention achieves a synergistic effect that simultaneously ensures high reliability for critical services and efficiently and flexibly schedules network resources.

[0056] 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. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0057] The core of this application is to provide a collaborative communication method for a distributed switch cluster, and a flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes:

[0058] S101. Collect active data flow information of all switching units in the distributed switch cluster to generate a global flow table, monitor the utilization data of multi-level interconnection links in the distributed switch cluster, and construct network status information of the distributed switch cluster.

[0059] Active data stream information refers to descriptive information about data streams currently being transmitted in the network, typically including the source address, destination address, and service type of data packets.

[0060] A global flow table is a summary table that integrates key characteristics of all active data flows in the entire cluster network, used to provide a macroscopic understanding of the network's traffic status. Multi-level interconnection links refer to the physical or logical network links within a distributed switch cluster used to connect different levels or regional switching units. Utilization data is a key indicator of link busyness, reflecting the ratio of the actual amount of data carried on the link to its maximum capacity. Network status information is a dynamic dataset that integrates the distribution of data flows across the entire network with the link load status.

[0061] Optionally, S101 may specifically include the following steps:

[0062] S1011 Periodically collects active data stream information that is forwarded locally from each switching unit in the distributed switch cluster, and integrates and processes the collected active data stream information.

[0063] The active data stream information includes the source address, destination address, and service type information of the data packets.

[0064] S1012. Extract the feature parameters of each active data stream, and generate a unified global flow table based on the extracted feature parameters.

[0065] The characteristic parameters include traffic volume, duration, and service quality requirements.

[0066] S1013. Continuously monitor the utilization data of each link on the multi-level interconnection link, record the bandwidth usage and load status of each link, and associate the utilization data with the global flow table for storage.

[0067] S1014. Construct network status information containing the distribution of data flow across the entire network and link load status information based on the associated data.

[0068] Among them, network status information reflects the network's operational status.

[0069] In this embodiment of the application, the execution process of S101 firstly forms preliminary raw data through periodic data collection and integration; then, features are extracted from the raw data to generate a structured global flow table; next, the load status of physical links is monitored in parallel, and the link status is associated with the data flow in the global flow table; finally, the associated data is presented in a visual manner to form intuitive network status information.

[0070] As an example, this process can be implemented as follows:

[0071] First, data acquisition and integration are performed via S1011. In this embodiment, within an industrial control network scenario, an 8-second acquisition cycle is set, and an sFlow protocol acquisition tool is used to acquire data from switching unit A (responsible for the East zone), switching unit B (responsible for the West zone), and switching unit C (responsible for the North zone) in the cluster. Within one cycle, the following information is acquired:

[0072] Switching Unit A: Transmits a video surveillance data stream with a source address of 192.168.3.1 and a destination address of 192.168.4.1.

[0073] Switching Unit B: Transmits a file backup data stream with a source address of 192.168.3.2 and a destination address of 192.168.4.2.

[0074] Switching unit C: Transmits two data streams, one of which is a video surveillance data stream that is the same as that in A, and the other is a voice scheduling data stream with a source address of 192.168.3.3 and a destination address of 192.168.4.3.

[0075] The collected information is sent to a centralized data processing node. This node identifies and removes duplicate video surveillance data streams by comparing key information such as the five-tuple (source / destination IP, source / destination port, protocol number) of the data streams, and then performs unified formatting on the cleaned three valid data streams.

[0076] Secondly, feature extraction and flow table generation are performed via S1012. For the three cleaned data streams, a statistical feature extraction algorithm is run. For example, the traffic volume of a data stream is calculated by multiplying the total number of data packets and the average packet length within a statistical unit of time (e.g., 1 second); the duration is calculated based on the difference between the timestamp of the first data packet and the current timestamp; and the quality of service requirements are obtained from a pre-defined policy library based on the service type. After calculation and matching, the feature parameters of each data stream are obtained, and a global flow table as shown in Table 1 is generated.

[0077]

[0078] Next, link monitoring and data association were performed via S1013. Using an SNMP-based monitoring tool, continuous monitoring was conducted on the three multi-level interconnect links (link 1, link 2, and link 3) connecting switching units A, B, and C to the core switching unit, assuming a total bandwidth of 100Mbps for each link. Monitoring revealed that link 1's current bandwidth usage was 60Mbps, link 2 was 25Mbps, and link 3 was 80Mbps. The link load level was calculated using the following formula:

[0079]

[0080] in, Indicates the load level. Indicates the bandwidth already used. This indicates the total bandwidth.

[0081] The preset load level threshold is: Load State .

[0082] Link 1 calculation process: (60Mbps / 100Mbps) 100% = 60%, which is considered a medium load.

[0083] Link 2 calculation process: (25Mbps / 100Mbps) 100% = 25%, which is considered a light load.

[0084] Link 3 calculation process: (80Mbps / 100Mbps) 100% = 80%, which is considered a heavy load.

[0085] Subsequently, using database association technology, the utilization data of each link is associated with the data flow entries passing through that link in the global flow table, using the "link identifier" as the key, and then stored in the network status database.

[0086] Finally, a visual view of network status information is constructed using S1014. Data visualization tools such as ECharts are used to read relevant data from the network status database. For example, in the generated visualization interface, green lines represent lightly loaded link 2, yellow lines represent medium-loaded link 1, and red lines represent heavily loaded link 3. directional arrows indicate the transmission paths of each data flow, with their service types labeled next to them. One side of the view supplements the detailed bandwidth usage of each link and key information such as traffic and duration of the corresponding data flow in a dynamic table format, thus forming a comprehensive view that intuitively reflects the overall network operation status.

[0087] The above example is only one specific implementation of this application. In practical applications, the collection period, protocol type, load threshold, etc. can be flexibly set according to the network environment and management needs. This application does not limit these aspects.

[0088] Through the above-mentioned scheme, this application can accurately and comprehensively collect and integrate dynamic data from the entire network, forming a unified and interconnected global network status view that includes data flow characteristics and link load status, providing reliable data support for subsequent efficient collaborative communication and intelligent traffic scheduling.

[0089] S102. Identify congested areas based on network status information, generate path optimization schemes for critical business flows based on the identification results, and establish corresponding redundant path scheduling strategies based on the path optimization schemes.

[0090] A congested area refers to a link segment in the network whose utilization rate consistently exceeds a preset safety threshold, representing a potential bottleneck in network performance. Critical service flows refer to high-priority data flows within the business system; their transmission quality directly impacts the normal operation of core services, such as real-time control commands or voice communication data.

[0091] The path optimization scheme is a new set of path planning measures designed to ensure that critical business flows bypass congested areas. The redundant path scheduling strategy is the specific execution manual for implementing this path optimization scheme, which defines in detail the conditions and steps for path switching, as well as how to deal with unexpected situations during the switching process.

[0092] Optionally, S102 may specifically include the following steps:

[0093] S1021. Analyze the utilization data of multi-level interconnected links in the network status information, identify link segments whose utilization rate continuously exceeds a predetermined threshold, and mark the identified high-utilization link segments as congested areas.

[0094] S1022. Identify all critical business flows passing through the congested area, analyze the service quality requirements and timeliness constraints of the critical business flows, and generate path optimization schemes based on the analysis results.

[0095] S1023. Based on the backup path information determined in the path optimization scheme, establish a corresponding redundant path scheduling strategy. The redundant path scheduling strategy includes path switching conditions based on network status changes, execution timing requirements to ensure service continuity, and fault handling mechanisms to deal with emergencies.

[0096] In this embodiment of the application, the execution process of S102 firstly involves performing trend analysis on network status information to accurately locate congested areas in the network; then, identifying critical service flows affected by congestion and tailoring one or more backup paths for them to form a path optimization scheme; finally, establishing an automated scheduling strategy that includes triggering conditions, execution steps, and contingency plans around the scheme.

[0097] As an example, this process can be implemented as follows:

[0098] First, congestion area identification is performed via S1021. This embodiment employs a sliding window-based data trend analysis algorithm to analyze the network status information constructed in S101. The algorithm sets a sliding window with a width of N monitoring periods and continuously calculates the average utilization rate of each link within the window. The predetermined threshold is 80%.

[0099] Taking the aforementioned industrial control network scenario as an example, the system continuously monitors the utilization data of link 3. At a certain point in time, the utilization rates collected continuously within the sliding window (set N=3, monitoring period is 5 seconds) are 82%, 83%, and 81%, respectively. Since all data points within the window exceed the predetermined threshold of 80%, the system determines that link 3 has entered a congested state and marks it as a congested area.

[0100] Secondly, a path optimization scheme is generated through S1022. The system first retrieves the global flow table, filters out all data flows passing through link 3, and identifies the key business flow, namely Flow-01 (voice scheduling flow), based on the preset business importance level. Subsequently, the system parses the service quality requirements of this data flow (latency ≤100ms, packet loss rate ≤0.1%).

[0101] Next, the system initiates an improved Dijkstra path planning algorithm. This algorithm uses "meeting service quality requirements" as a constraint and "minimizing the overall path load" as the optimization objective to find an alternative path within the current network topology. Algorithm analysis reveals that the link 2 traversed by the path "Switching Unit C → Switching Unit B → Core Unit" is under "light load," fully meeting the service quality requirements of the voice scheduling flow. Therefore, the system generates a path optimization scheme, designating this path as the primary and backup path for Flow-01.

[0102] Finally, a redundant path scheduling strategy is established through S1023. Based on the path optimization scheme generated in the previous step, the system formulates a detailed scheduling strategy for Flow-01. This strategy is defined in a structured form, as shown in Table 2 below:

[0103]

[0104] Here, T represents the moment when the triggering condition is met. This execution sequence follows the "Make-Before-Break" principle to ensure uninterrupted service during the switchover process. The fault handling mechanism provides further protection in case the backup path itself fails.

[0105] The above example is only one specific implementation of this application. In practical applications, the congestion determination algorithm, the path planning objective, and the details of the scheduling strategy can all be adjusted according to business needs, and this application does not limit them.

[0106] Through the above-mentioned scheme, this application can accurately predict and locate congestion risks from the macro network status, and automatically generate a complete scheduling plan including backup paths and fine-grained execution rules for the affected critical service flows, thereby transforming the network from passively responding to faults to actively avoiding risks, and significantly improving the transmission stability of critical services.

[0107] S103. The global flow table and path optimization scheme are parsed through a centralized control process, and the optimal forwarding path of the critical business flow is calculated based on the parsing results. Flow table instructions that are compatible with the redundant path scheduling strategy are generated and distributed to the target switching unit in the distributed switch cluster.

[0108] Centralized control refers to a process in which a central control node uniformly executes decisions and calculations. Its advantage lies in its ability to make optimal judgments based on global information. Quality of Service (QoS) objectives translate abstract business requirements (such as "guaranteed transmission") into measurable and actionable network performance metrics, for example, specifying latency requirements down to the millisecond level.

[0109] Optionally, S103 may specifically include the following steps:

[0110] S1031. Parse the global flow table to obtain the characteristic information of the critical business flow. Combine the characteristic information of the critical business flow with the optimization requirements defined in the path optimization scheme to determine the service quality target and path selection constraints of the critical business flow.

[0111] Path selection constraints are a series of "hard rules" that must be followed when calculating paths, such as avoiding certain links or passing through certain nodes. The optimal forwarding path is the best transmission path calculated based on a specific optimization objective (such as minimizing overall cost) while satisfying all constraints.

[0112] S1032. Based on the quality of service objectives and path selection constraints, and combined with the current network topology status and utilization data, calculate the optimal forwarding path for critical service flows in the distributed switch cluster.

[0113] S1033. Based on the calculated optimal forwarding path and the path switching rules defined in the redundant path scheduling strategy, generate flow table instructions. The flow table instructions contain detailed forwarding rules for the primary path and backup path, and distribute the flow table instructions to the target switching units in the distributed switch cluster.

[0114] Flow table commands are specific execution commands issued by the central controller to the switching unit. They precisely define the forwarding behavior of data packets within the switch in a machine-readable format, such as "data packets coming in from port A, if they meet a certain condition, will be sent out from port B".

[0115] In this embodiment of the application, the execution process of S103 firstly parses global data to quantify the service objectives and path constraints of critical business traffic; then, under the premise of satisfying these constraints, a path calculation algorithm is used to find an optimal path and one or more backup paths in the current network; finally, these path information are compiled into standard flow table instructions and sent to all relevant switching units along the path.

[0116] As an example, this process can be implemented as follows:

[0117] First, the calculation objective is determined through S1031. The central controller parses the feature information of the critical business flow Flow-01 (voice scheduling flow) from the global flow table, and combines it with the path optimization scheme generated in S102. Through the demand mapping algorithm, the optimization requirement "avoiding link 3 congestion and ensuring transmission" is transformed into more stringent service quality objectives (such as latency ≤80ms, packet loss rate ≤0.08%) and explicit path selection constraints (such as not passing through link 3, and the utilization rate of the links passing through ≤60%).

[0118] Secondly, the optimal forwarding path is calculated using S1032. The controller runs a path calculation algorithm based on the current network topology and real-time utilization data. This algorithm first eliminates all links that do not meet the path selection constraints, and then searches for candidate paths that meet the quality of service (QoS) objectives in the remaining topology. For each candidate path, the algorithm calculates its comprehensive weight according to the following formula:

[0119] (1)

[0120] in, The overall weight representing the path; This represents the importance coefficient of latency in the weight calculation (e.g., 0.6). The predicted delay value for this path (unit: ms); This represents the importance coefficient of packet loss rate in the weight calculation (e.g., 0.4). This represents the predicted packet loss rate (percentage) for this path.

[0121] Taking the aforementioned industrial network scenario as an example, the algorithm finds two candidate paths:

[0122] 1. Candidate path 1: "C→B→Core unit", prediction latency 60ms, packet loss rate 0.05%.

[0123] 2. Candidate path 2: "C→A→B→Core unit", prediction latency 75ms, packet loss rate 0.06%.

[0124] The calculation process is as follows:

[0125] The overall weight of path 1 is 0.6 × 60 + 0.4 × 0.05 = 36 + 0.02 = 36.02.

[0126] The overall weight of path 2 is 0.6 × 75 + 0.4 × 0.06 = 45 + 0.024 = 45.024.

[0127] Since candidate path 1 has a smaller overall weight, the system determines it as the optimal forwarding path (i.e., the primary path) and uses candidate path 2 as the backup path.

[0128] Finally, flow table instructions are generated and distributed via S1033. Based on the calculated primary and backup paths, and in conjunction with the redundant path scheduling strategy defined in S102, the controller generates a set of flow table instructions. These instructions are distributed to the switching units C, B, and A involved in the path via the standard OpenFlow protocol. An example of the instruction content is as follows:

[0129] Instructions issued to switching unit C:

[0130] Primary rule: Matching condition (source IP=192.168.3.3, destination IP=192.168.4.3), action (forward from port 2 to switching unit B).

[0131] Alternative rule: Matching condition (source IP=192.168.3.3, destination IP=192.168.4.3), action (forward from port 3 to switching unit A).

[0132] Instructions issued to switching unit B:

[0133] Primary rule: Matching condition (entering from port 3, source IP=192.168.3.3), action (forwarding from port 1 to the core unit).

[0134] Alternative rule: Matching condition (entering from port 4, source IP=192.168.3.3), action (forwarding from port 1 to the core unit).

[0135] Instructions issued to switching unit A:

[0136] Alternative rule: Match condition (entering from port 2, source IP=192.168.3.3), action (forwarding from port 4 to switching unit B).

[0137] After receiving the instruction, each switching unit preloads these rules. Under normal circumstances, the primary rule is enabled, and when the switching conditions defined in the redundant path scheduling policy are met, it automatically switches to the backup rule.

[0138] The above example is only one specific implementation of this application. In practical applications, the weight coefficient of path calculation, the number of backup paths, etc. can be configured according to actual needs, and this application does not limit them.

[0139] This application, through the aforementioned scheme, systematically transforms macro-level business assurance requirements into quantifiable computational goals and constraints, and calculates the optimal path based on real-time network conditions. This scheme not only plans efficient and reliable primary paths for critical business flows but also proactively pre-sets backup paths and corresponding forwarding rules, directly deploying these contingency plans into network devices. This ensures rapid and seamless path switching during network fluctuations, significantly improving the continuity and stability of critical services.

[0140] S104. Obtain the local link status information of the switching unit through the distributed control process, combine the local link status information with the utilization data to perform dynamic path selection for non-critical business flows, generate local path decisions, and distribute the local path decisions to the relevant switching units in the distributed switch cluster.

[0141] In this invention, distributed control specifically refers to a working mode that complements centralized control. It is not entirely decentralized, but rather delegates path selection authority for non-critical service flows, allowing switching units to make more timely forwarding decisions based on their own perceived local network conditions. Local link state information refers to the switching unit's "first-person" observation data of its directly connected links, such as the link's current real-time load, latency, and remaining available bandwidth.

[0142] The comprehensive link quality assessment result is a quantitative score that integrates link status data from multiple dimensions into a single metric, allowing for a direct comparison of the merits of different links. Local path decision-making is the final product of the distributed control process; it is a set of instructions specifically designed for non-critical business flows, including path selection strategies and specific forwarding rules.

[0143] Optionally, S104 may specifically include the following steps:

[0144] S1041. Obtain the local link status information of each switching unit through a distributed control process. The local link status information includes link utilization, transmission delay, and available bandwidth data.

[0145] S1042. Correlate the local link status information with the utilization data to generate a comprehensive link quality assessment result. Based on the comprehensive link quality assessment result and the transmission characteristics of non-critical service flows, implement dynamic path selection for non-critical service flows.

[0146] S1043. Based on the results of dynamic path selection, generate local path decisions that include path selection strategies and forwarding rules, ensure that local path decisions are coordinated with flow table instructions, and distribute local path decisions to relevant switching units in the distributed switch cluster.

[0147] In this embodiment of the application, the execution process of S104 above first gathers the local "views" of each switching unit through distributed data interaction; then, it comprehensively evaluates these local information to "select the best" among available links for non-critical business flows; finally, it forms a specific execution decision and, on the premise of ensuring that it does not "compete" with critical business, issues the decision for execution.

[0148] As an example, this process can be implemented as follows:

[0149] First, the local link status is obtained via S1041. The central control node maintains a lightweight communication connection with switching units A, B, and C, and polls each unit at a short interval (e.g., 6 seconds) to request them to report the local status of their directly connected links. For example, switching unit A reports the real-time utilization, transmission delay, and available bandwidth of its connected link 1; B and C similarly report the corresponding data for links 2 and 3. This data is aggregated at the control node to form a real-time dataset reflecting the status of the network's "capillaries."

[0150] Secondly, evaluation and path selection are performed via S1042. The control node retrieves the recently collected local link status information and, combined with the network-wide utilization data obtained in S101, initiates a multi-dimensional evaluation algorithm to calculate a comprehensive quality score for each link. The algorithm formula is as follows:

[0151] (2)

[0152] in, The overall quality score representing the link; , , These represent the weighting coefficients for utilization, latency, and available bandwidth (the sum of the three is 1). , , These are the normalized utilization, latency, and available bandwidth values, respectively.

[0153] Assuming that for non-critical business flows such as file backup, bandwidth and low utilization are more important, the weight is set to... =0.4, =0.2, =0.4. Based on the collected data, the scores are calculated as follows: Link 1 score is 0.75, Link 2 score is 0.68, and Link 3 score is 0.32.

[0154] At this point, the system retrieves the transmission characteristics requirements of the non-critical business flow Flow-03 (file backup flow) (meeting a minimum bandwidth of 10Mbps is sufficient), and finds that both link 1 and link 2 meet the requirements. Based on the "optimal selection" principle, the system dynamically selects link 1, which currently has the highest overall quality score, as the transmission path for this file backup flow.

[0155] Finally, local route decisions are generated and coordinated via S1043. Based on the selection results from the previous step, the system generates a local route decision. This decision includes:

[0156] Path selection strategy: "Use link 1 first; when the QS of link 1 is lower than 0.7, automatically switch to link 2."

[0157] Forwarding rule: "For data packets that match the characteristics of the file backup stream, they enter from port 2 of switching unit B and are forwarded to the core unit from port 1."

[0158] Before issuing this decision, the system performs a crucial coordination step: checking for conflicts between the local path decision and the flow table instructions generated for critical business flows in S103. The check finds that the file backup flow using link 1 will not cause the load on link 1 to exceed the safety threshold reserved for critical business flows, nor will it consume its backup path resources. After confirming there are no conflicts, this local path decision is issued to the path-related switching unit B and core unit, guiding them to forward non-critical business flows.

[0159] The above example is only one specific implementation of this application. In practical applications, the weights of the evaluation algorithm, the switching threshold in the decision-making process, etc. can be dynamically adjusted, and this application does not limit them.

[0160] This application, through the aforementioned scheme, decentralizes the responsibility for path management of a large number of non-critical service flows. Through real-time local status awareness and dynamic evaluation, it flexibly matches these service flows with the optimal transmission resources available at the moment. This scheme, while ensuring the stable operation of critical service flows, achieves precise and efficient utilization of network resources, effectively avoiding resource waste and network congestion caused by a "one-size-fits-all" approach, and improving the overall network throughput and intelligence level.

[0161] S105 coordinates the execution of flow table instructions and local path decisions, dynamically adjusts the allocation of service flow paths among multi-level switching units according to network status, and realizes the coordinated operation of centralized control and distributed execution.

[0162] Optionally, S105 may specifically include the following steps:

[0163] S1051. Receive flow table instructions and local path decisions, and coordinate the flow table instructions and local path decisions.

[0164] Coordination refers to a final review and alignment of strategies developed for different business types before any path allocation is formally implemented, to ensure that they do not conflict with each other.

[0165] S1052. Dynamically adjust the allocation of service flow paths among multi-level switching units according to network status, prioritize the service quality requirements of critical service flows, and optimize the transmission efficiency of non-critical service flows. Ensure that critical service flows are forwarded strictly according to the path specified by the flow table instructions, and that non-critical service flows flexibly select paths based on local path decisions to avoid path conflicts.

[0166] Path conflict refers to the contradiction in resource usage that occurs when different service flows are routed, such as two paths competing for the bandwidth of the same narrowband link.

[0167] S1053. Monitor the execution effect of path adjustment, optimize the decision-making process based on feedback information, and achieve coordinated operation of centralized control and distributed execution.

[0168] Feedback information refers to the actual performance data collected by the system after the path adjustment is executed. It serves as the basis for evaluating the effectiveness of the decision and for self-optimization.

[0169] In this embodiment of the application, the execution process of S105 firstly involves a preliminary coordination process to check and calibrate the conflicts of the instructions and decisions to be issued; then, during operation, the path guarantee for critical business flows is strictly implemented, and non-critical business flows are flexibly scheduled to achieve differentiated services; finally, through a closed-loop monitoring and feedback mechanism, the entire decision-making system is continuously evaluated and optimized to achieve system self-adaptation and evolution.

[0170] As an example, this process can be implemented as follows:

[0171] First, coordination is performed via S1051. Before issuing any instructions, the central control node initiates a conflict detection algorithm. This algorithm takes the flow table instructions generated in S103 (for critical business flows) and the local path decisions generated in S104 (for non-critical business flows) as input.

[0172] Taking the aforementioned industrial network scenario as an example, the flow table instruction planned the primary path "C→B→Core Unit" for the critical business flow Flow-01 (voice scheduling) and declared that some bandwidth needed to be reserved for link 2. However, the local path decision selected the initial path "B→Core Unit" for the non-critical business flow Flow-03 (file backup), which also passed through link 2. The conflict detection algorithm found that there was a potential risk of resource contention between the two paths on link 2. To avoid conflict, the system automatically adjusted the local path decision, modifying the path of the file backup flow to the then suboptimal "B→A→Core Unit", thereby ensuring that the dedicated resources of the critical business flow were not infringed.

[0173] Secondly, dynamic path allocation is performed via S1052. During network operation, each switching unit forwards data according to coordinated instructions and decisions.

[0174] For the critical business flow Flow-01, switching units C and B strictly follow the flow table instructions, forwarding data via the path "C→B→core unit". The system continuously monitors this path, and once it detects that the performance (such as latency) of link 2 deteriorates and reaches the switching threshold defined in the flow table instructions, the switching unit will automatically and seamlessly activate the preset backup path, without requiring further intervention from the central controller.

[0175] For non-critical business flow Flow-03, switching units B and A forward the data via the "B→A→core unit" path based on their local path decisions. Simultaneously, switching unit B continuously evaluates the overall quality score of the "B→A→core unit" path based on the strategy in its local decision. If, at any point, the quality of this path deteriorates, and another alternative path (such as link 3, which has now become idle) has better quality, switching unit B will autonomously switch the file backup flow to the better path to improve transmission efficiency.

[0176] Finally, feedback and optimization are achieved through S1053. The system's monitoring module continuously collects data on the performance of the adjusted paths, generating feedback information. For example, monitoring revealed that the average transmission rate of the non-critical business flow Flow-03 was still lower than expected after switching to the new path.

[0177] Upon receiving this feedback, the central control node analyzed it and found that although the new path had sufficient bandwidth, occasional and brief load fluctuations affected the stability of file transfers. Based on this feedback, the system optimized the decision-making process: it updated the multi-dimensional evaluation algorithm in S104, added consideration of "link stability," and adjusted the weights. Simultaneously, a new, more stable alternative path was added for Flow-03 in the local path decision-making process. The optimized decision was redistributed to the relevant switching units, enabling better path selection in the next decision cycle, forming a complete "execution-monitoring-feedback-optimization" closed loop.

[0178] The above example is only one specific implementation of this application. In practical applications, the rules for conflict detection, the triggering conditions for path adjustment, and the strategies for feedback optimization can all be customized according to the actual needs of the business. This application does not limit them.

[0179] This application, through the aforementioned scheme, ensures that service flows of different priorities do not interfere with each other in path planning, providing a fundamental guarantee for stable network operation. This scheme, through differentiated path allocation and adjustment strategies, achieves rigid guarantees for the quality of service of critical service flows and flexible optimization of the transmission efficiency of non-critical service flows.

[0180] Figure 2 This is a schematic diagram illustrating a specific implementation of a collaborative communication system for a distributed switch cluster provided in this application. (Refer to...) Figure 2 The system may include:

[0181] Module 21 is used to collect active data flow information of all switching units in the distributed switch cluster to generate a global flow table and monitor the utilization data of multi-level interconnection links in the distributed switch cluster to construct network status information of the distributed switch cluster.

[0182] Module 22 is established to identify congested areas based on network status information, generate path optimization schemes for critical business flows based on the identification results, and establish corresponding redundant path scheduling strategies based on the path optimization schemes.

[0183] The distribution module 23 is used to parse the global flow table and path optimization scheme through a centralized control process, calculate the optimal forwarding path of the critical business flow based on the parsing results, generate flow table instructions that are compatible with the redundant path scheduling strategy, and distribute the flow table instructions to the target switching unit in the distributed switch cluster.

[0184] The second distribution module 24 is used to obtain the local link status information of the switching unit through the distributed control process, combine the local link status information with the utilization data to perform dynamic path selection for non-critical business flows, generate local path decisions, and distribute the local path decisions to the relevant switching units in the distributed switch cluster.

[0185] The coordination module 25 is used to coordinate the execution of flow table instructions and local path decisions, and dynamically adjust the allocation of service flow paths among multi-level switching units according to network status, so as to realize the coordinated operation of centralized control and distributed execution.

[0186] The distributed switch cluster cooperative communication system of this application embodiment is used to implement the aforementioned distributed switch cluster cooperative communication method. Therefore, the specific implementation of the distributed switch cluster cooperative communication system can be found in the embodiment section of the distributed switch cluster cooperative communication method above. The specific implementation can be referred to the description of the corresponding embodiment, which will not be repeated here.

[0187] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the cooperative communication method of any of the above-described distributed switch clusters.

[0188] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the cooperative communication method of any of the above-described distributed switch clusters.

[0189] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0190] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the cooperative communication method for a distributed switch cluster.

[0191] 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 implementations should not be considered beyond the scope of this invention.

[0192] The above provides a detailed description of the collaborative communication method and system for a distributed switch cluster provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A collaborative communication method for a distributed switch cluster, characterized in that, include: Collect active data flow information of all switching units in the distributed switch cluster to generate a global flow table, and monitor the utilization data of multi-level interconnection links in the distributed switch cluster to construct the network status information of the distributed switch cluster. Based on the network status information, congested areas are identified, and path optimization schemes for critical service flows are generated according to the identification results. A corresponding redundant path scheduling strategy is then established based on the path optimization schemes. The global flow table and the path optimization scheme are parsed through a centralized control process. The optimal forwarding path for critical service flows is calculated based on the parsing results. Flow table instructions are generated based on the calculated optimal forwarding path and the path switching rules defined in the redundant path scheduling strategy. The flow table instructions contain detailed forwarding rules for the primary path and backup path. The flow table instructions are then distributed to the target switching units in the distributed switch cluster. The local link status information of the switching unit is obtained through a distributed control process. The local link status information and the utilization data are combined to perform dynamic path selection for non-critical service flows, generate local path decisions, and distribute the local path decisions to the relevant switching units in the distributed switch cluster. The system coordinates the execution of the flow table instructions and the local path decisions, and dynamically adjusts the path allocation of service flows among multi-level switching units based on the network status information, thereby achieving coordinated operation of centralized control and distributed execution.

2. The method according to claim 1, characterized in that, Based on the network status information, congested areas are identified. Based on the identification results, path optimization schemes for critical service flows are generated. Corresponding redundant path scheduling strategies are established based on the path optimization schemes, including: Analyze the utilization data of multi-level interconnected links in the network status information, identify link segments whose utilization rate continuously exceeds a predetermined threshold, and mark the identified high-utilization link segments as congested areas; Identify all critical business flows passing through the congested area, analyze the service quality requirements and timeliness constraints of the critical business flows, and generate path optimization schemes based on the analysis results; Based on the backup path information determined in the path optimization scheme, a corresponding redundant path scheduling strategy is established, wherein the redundant path scheduling strategy includes: path switching conditions, execution timing requirements, and fault handling mechanisms.

3. The method according to claim 1, characterized in that, The global flow table and the path optimization scheme are parsed through a centralized control process. Based on the parsing results, the optimal forwarding path for critical service flows is calculated. According to the calculated optimal forwarding path and the path switching rules defined in the redundant path scheduling strategy, flow table instructions are generated. These flow table instructions contain detailed forwarding rules for the primary and backup paths. The flow table instructions are then distributed to the target switching units in the distributed switch cluster, including: The global flow table is parsed to obtain the feature information of the key business flow. The service quality target and path selection constraints of the key business flow are determined by combining the feature information of the key business flow and the optimization requirements defined in the path optimization scheme. Based on the service quality objectives and path selection constraints, and combined with the current network topology and utilization data, the optimal forwarding path for the critical service flow in the distributed switch cluster is calculated. Based on the calculated optimal forwarding path and the path switching rules defined in the redundant path scheduling strategy, flow table instructions are generated. The flow table instructions contain detailed forwarding rules for the primary path and backup path, and the flow table instructions are distributed to the target switching units in the distributed switch cluster.

4. The method according to claim 1, characterized in that, The local link status information of the switching unit is obtained through a distributed control process. Combined with the local link status information and utilization data, dynamic path selection is performed on non-critical service flows to generate local path decisions. These local path decisions are then distributed to the relevant switching units in the distributed switch cluster, including: The local link status information of each switching unit is obtained through a distributed control process. The local link status information includes link utilization, transmission delay and available bandwidth data. The local link status information is correlated with the utilization data to generate a comprehensive link quality assessment result. Based on the comprehensive link quality assessment result and the transmission characteristics requirements of non-critical service flows, dynamic path selection is implemented for non-critical service flows. Based on the results of dynamic path selection, a local path decision containing path selection strategy and forwarding rules is generated to ensure that the local path decision is coordinated with the flow table instructions, and the local path decision is distributed to the relevant switching units in the distributed switch cluster.

5. The method according to claim 4, characterized in that, The local link status information of each switching unit is obtained through a distributed control process. This local link status information includes link utilization, transmission delay, and available bandwidth data, including: A data acquisition agent is deployed on the switching unit through a distributed control process to establish a communication connection with the controller; The data acquisition agent periodically collects local link operation status data, including the number of communication data packets transmitted, latency measurement information, and bandwidth usage. The collected operation status data is preprocessed to extract link utilization indicators, transmission delay indicators, and available bandwidth indicators. The extracted indicator data are standardized in format and combined to generate structured local link status information, which is then transmitted to the controller through the established communication connection. The received local link status information is stored in a distributed database for use in subsequent path decision-making processes.

6. The method according to claim 1, characterized in that, The system coordinates the execution of the flow table instructions and the local path decisions, dynamically adjusts the allocation of service flow paths among multi-level switching units based on the network status information, and achieves coordinated operation of centralized control and distributed execution, including: Receive the flow table instruction and the local path decision, and coordinate the flow table instruction and the local path decision; Based on the network status information, the service flow path allocation is dynamically adjusted among the multi-level switching units to prioritize the service quality requirements of the critical service flows, while optimizing the transmission efficiency of the non-critical service flows. This ensures that the critical service flows are forwarded strictly according to the path specified by the flow table instructions, while the non-critical service flows flexibly select paths based on the local path decision to avoid path conflicts. Monitor the effectiveness of path adjustments, optimize the decision-making process based on feedback, and achieve coordinated operation of centralized control and distributed execution.

7. The method according to claim 1, characterized in that, Collect active data flow information from all switching units in the distributed switch cluster to generate a global flow table, monitor the utilization data of multi-level interconnection links within the distributed switch cluster, and construct network status information for the distributed switch cluster, including: Periodically collect active data flow information of local forwarding from each switching unit in the distributed switch cluster, including the source address, destination address and service type information of the data packets, and integrate and process the collected active data flow information; Extract the characteristic parameters of each active data stream, including traffic size, duration and quality of service requirements, and generate a unified global flow table based on the extracted characteristic parameters; The utilization data of each link on the multi-level interconnection link is continuously monitored, the bandwidth usage and load status of each link are recorded, and the utilization data is associated with and stored in the global flow table. Based on the correlated data, a network status information system is constructed that includes information on the distribution of data flows across the entire network and the link load status. This network status information reflects the network's operational status.

8. A collaborative communication system for a distributed switch cluster, characterized in that, include: The construction module is used to collect active data flow information of all switching units in the distributed switch cluster to generate a global flow table, and to monitor the utilization data of multi-level interconnection links in the distributed switch cluster to construct the network status information of the distributed switch cluster. A module is established to identify congested areas based on the network status information, generate path optimization schemes for critical service flows based on the identification results, and establish corresponding redundant path scheduling strategies based on the path optimization schemes. The distribution module is used to parse the global flow table and the path optimization scheme through a centralized control process, calculate the optimal forwarding path of the critical service flow based on the parsing results, generate flow table instructions based on the calculated optimal forwarding path and the path switching rules defined in the redundant path scheduling strategy, the flow table instructions contain detailed forwarding rules for the primary path and the backup path, and distribute the flow table instructions to the target switching units in the distributed switch cluster. The second distribution module is used to obtain the local link status information of the switching unit through a distributed control process, combine the local link status information with the utilization data to perform dynamic path selection for non-critical service flows, generate local path decisions, and distribute the local path decisions to the relevant switching units in the distributed switch cluster. The coordination module is used to coordinate the execution of the flow table instructions and the local path decision, and dynamically adjust the allocation of service flow paths among multi-level switching units according to the network status information, so as to realize the coordinated operation of centralized control and distributed execution.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the cooperative communication method for a distributed switch cluster as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the collaborative communication method of a distributed switch cluster as described in any one of claims 1 to 7.

Images