Network self-healing method and device based on SDN controller, and medium

Abstract

The invention discloses a network self-healing method and device based on an SDN controller and a medium. The method comprises the steps that a historical knowledge base and an atomic operation and maintenance command base are embedded into a self-healing large language model; the historical knowledge base comprises fault operation and maintenance commands corresponding to historical fault types, and the atomic operation and maintenance command base comprises atomic operation and maintenance commands corresponding to different fault types; the alarm information of the SDN controller is retrieved and matched in a historical knowledge base and an atomic operation and maintenance command base through a self-healing large language model; processing the target fault operation and maintenance command and the target atomic operation and maintenance command through a self-healing large language model to obtain an operation and maintenance command with the highest self-healing success rate; and determining the operation and maintenance command with the highest self-healing success rate as a self-healing execution strategy of the alarm information, and issuing the self-healing execution strategy to the SDN controller. And in combination with a customized self-healing large language model, the self-healing capability of the SDN is more flexible and the adaptability is higher.

Description

Technical Field

[0001] The present application relates to the technical field of SDN controller fault handling, and in particular to a network self-healing method, device, and medium based on an SDN controller. Background Art

[0002] As enterprises continue to expand, the number of servers and switches has exploded, often reaching hundreds or even thousands. Managing this scale of equipment increases complexity. Relying on traditional manual methods for daily updates, deployments, and management not only consumes valuable time but is also prone to operational errors and omissions due to negligence. This not only impacts network system stability but also significantly increases operational costs and management complexity. Therefore, introducing SDN-based fault recovery technology to improve enterprise network system availability, reduce troubleshooting complexity and manpower, and achieve a shift from manual to unattended intelligent operations has become an inevitable choice for enterprise network management.

[0003] Currently, fixed rules for resolving network issues are typically developed and then hard-coded into the system. The core logic is based on manually defined rules and heuristics, using simple descriptive analysis to drive optimization of the entire network system. For example, when signal A exceeds threshold B within C seconds, self-healing action X is directly triggered.

[0004] However, this approach is essentially a hard-coded model, with rules that cannot be freely adjusted or flexibly assembled. This makes it difficult to adapt to dynamic changes in the network environment and effectively address increasingly complex network situations. Furthermore, these rules are limited by the boundaries of human thinking, and even the most experienced experts cannot exhaustively consider all factors that affect network performance. More importantly, this approach cannot achieve real-time optimization of network fault repair, making it difficult to intelligently select the optimal self-healing strategy based on specific warning conditions. This makes it difficult to meet the needs of modern enterprises for efficient and intelligent network management. Summary of the Invention

[0005] The embodiments of the present application provide a network self-healing method, device, and medium based on an SDN controller, which are used to solve the problems of poor flexibility, poor adaptability, and low efficiency of the network self-healing method based on the SDN controller.

[0006] The embodiments of this application adopt the following technical solutions: On the one hand, an embodiment of the present application provides a network self-healing method based on an SDN controller, the method comprising: embedding a preset historical knowledge base and a preset atomic operation and maintenance command library into a pre-trained self-healing large language model; the preset historical knowledge base includes fault operation and maintenance commands corresponding to historical fault types, and the preset atomic operation and maintenance command library includes atomic operation and maintenance commands corresponding to different fault types; the alarm information of the SDN controller is retrieved and matched in the preset historical knowledge base and the preset atomic operation and maintenance command library through the self-healing large language model; when there are matching target fault operation and maintenance commands and target atomic operation and maintenance commands, the target fault operation and maintenance commands and the target atomic operation and maintenance commands are processed through the self-healing large language model to obtain the operation and maintenance command with the highest self-healing success rate; the operation and maintenance command with the highest self-healing success rate is determined as the self-healing execution strategy of the alarm information, and the self-healing execution strategy is sent down to the SDN controller.

[0007] In one example, before embedding the preset historical knowledge base and the preset atomic operation and maintenance command library into the pre-trained self-healing large language model, the method also includes: obtaining sample alarm information and atomic operation and maintenance commands of the sample alarm information; using the sample alarm information as sample input and the atomic operation and maintenance commands of the sample alarm information as sample output, training the initial self-healing large language model to obtain an initial trained self-healing large language model; training the initial trained self-healing large language model according to the self-healing success rate of the sample fault operation and maintenance commands and the sample fault operation and maintenance commands, as well as the self-healing success rate of the sample atomic operation and maintenance commands and the sample atomic operation and maintenance commands to obtain a self-healing large language model.

[0008] In one example, the method further includes: when there is no matching target atomic operation and maintenance command, inputting the alarm information into the self-healing large language model to generate an updated atomic operation and maintenance command for the alarm information; and determining the updated atomic operation and maintenance command as the target atomic operation and maintenance command.

[0009] In one example, the method further includes: when there is no matching target fault operation and maintenance command, searching the Internet for an operation and maintenance command of the fault type in the alarm information; and determining the searched operation and maintenance command as the target atomic operation and maintenance command.

[0010] In one example, before searching and matching the alarm information of the SDN controller in a preset historical knowledge base and a preset atomic operation and maintenance command library through the self-healing large language model, the method also includes: performing data cleaning on the initial alarm information uploaded by the SDN controller; performing noise reduction processing on the cleaned initial alarm information to obtain noise-reduced alarm information; converting the noise-reduced alarm information format into a format recognizable by the self-healing large language model to obtain the alarm information.

[0011] In one example, after the self-healing execution policy is sent to the SDN controller, the method further includes: the SDN controller sends the self-healing execution policy to the network device through the MCP protocol; receives feedback results from the network device, and when the feedback result is that the fault is not resolved, generates an operation and maintenance work order for the alarm information, and sends the operation and maintenance work order to the operation and maintenance client; when the feedback result is that the fault is resolved, adds the alarm information and the self-healing execution policy to the preset historical knowledge base.

[0012] In one example, the method further includes: processing the initial alarm information of the SDN controller through a front-end interceptor to obtain alarm information, and sending the alarm information to the self-healing large language model; and performing log recording according to a back-end interceptor.

[0013] In one example, the method further includes: verifying the legitimacy of operation and maintenance commands for fault types in alarm information searched from the Internet through a front-end interceptor; after the verification is passed, determining the searched operation and maintenance commands as target atomic operation and maintenance commands.

[0014] On the other hand, an embodiment of the present application provides a network self-healing device based on an SDN controller, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any one of the above-described network self-healing methods based on an SDN controller.

[0015] On the other hand, an embodiment of the present application provides a network self-healing non-volatile computer storage medium based on an SDN controller, which stores computer-executable instructions, and the computer-executable instructions can execute any of the above-mentioned network self-healing methods based on an SDN controller.

[0016] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: On the one hand, a complete software-defined networking (SDN) operations and maintenance system has been built, enabling centralized management and dynamic allocation of network resources. From network status monitoring to fault diagnosis and self-healing execution, the entire process is seamlessly integrated, forming an efficient, closed-loop operations and maintenance ecosystem. This systematic operations and maintenance architecture not only improves network stability and availability, but also significantly reduces operations and maintenance costs and complexity.

[0017] On the other hand, users can freely combine different types of atomic operation and maintenance commands according to their needs, achieving highly customized operation and maintenance operations. This flexibility gives users great autonomy and provides them with a path for expansion. Whether it is simple tasks or complex scenarios, they can flexibly combine atomic commands to meet specific needs, further improving operation and maintenance efficiency and adaptability. In addition, when the atomic command library is insufficient, atomic operation and maintenance commands can be generated in a timely manner for the current alarm information, further improving the flexibility and adaptability of the self-healing execution strategy.

[0018] On the other hand, the customized self-healing large language model significantly improves the success rate and accuracy of network self-healing.

[0019] Specifically, the system not only relies on real-time alarm and fault information but also builds a historical knowledge base and atomic operation and maintenance command library, providing a rich source of information for reasoning in the self-healing large language model. Furthermore, it can obtain additional knowledge (such as solutions to fault types) through online searches, further expanding the information dimension.

[0020] In addition, the traditional hard-coded rigid model has been abandoned, and the model's intelligent decision-making logic has been introduced. It can match different operation and maintenance commands based on real-time network status and warning information, and predict the self-healing success rate of different operation and maintenance commands, thereby dynamically selecting the most appropriate and optimal self-healing execution strategy.

[0021] Obviously, through the flexible deployment and assembly of operation and maintenance commands, the system's fault tolerance and adaptability have been enhanced, enabling it to flexibly respond to various complex network environments, thereby achieving more efficient and intelligent network self-healing, and realizing the leap from fixed rules to intelligent decision-making, ensuring that operation and maintenance operations are always accurately matched with the actual needs of the network. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solution of the present application, some embodiments of the present application will be described in detail below with reference to the accompanying drawings, in which: Figure 1 A flowchart of a network self-healing method based on an SDN controller provided in an embodiment of the present application; Figure 2 A flowchart of another network self-healing method based on an SDN controller provided in an embodiment of the present application; Figure 3 A schematic diagram of a network self-healing system based on an SDN controller provided in an embodiment of the present application; Figure 4 A flowchart of a rule-based network self-healing method in the prior art provided in an embodiment of the present application; Figure 5A schematic diagram of the structure of a network self-healing device based on an SDN controller provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] To make the objectives, technical solutions, and advantages of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0024] Some embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0025] Figure 1 This is a flow chart of a network self-healing method based on an SDN controller, provided in an embodiment of the present application. This method can be applied to various business areas, such as internet finance, e-commerce, instant messaging, gaming, and government affairs. Certain input parameters or intermediate results in this process can be manually adjusted to help improve accuracy.

[0026] The analysis method involved in the embodiments of the present application can be implemented by a terminal device or a server, and the present application does not impose any special restrictions on this. For ease of understanding and description, the following embodiments are described in detail using a server as an example.

[0027] It should be noted that the server can be a single device or a system composed of multiple devices, that is, a distributed server, and this application does not make any specific restrictions on this.

[0028] Figure 1 The process in includes the following steps: S101: Embed a preset historical knowledge base and a preset atomic operation and maintenance command library into a pre-trained self-healing large language model; the preset historical knowledge base includes fault operation and maintenance commands corresponding to historical fault types, and the preset atomic operation and maintenance command library includes atomic operation and maintenance commands corresponding to different fault types.

[0029] It should be noted that the preset historical knowledge base can be a private internal business knowledge base of the enterprise. Specifically, the historical knowledge base contains fault records from the enterprise's deployed software network. This serves as a knowledge base of past faults and fault alarm recovery cases, providing a rich source of information for the self-healing large language model to reason about. Furthermore, a fault type can correspond to one or more atomic operation and maintenance commands.

[0030] In summary, by setting up an atomic operation and maintenance command library, users can freely combine different types of atomic operation and maintenance commands according to their needs, achieving highly customized operation and maintenance operations. This flexibility gives users great autonomy. Whether it is simple tasks or complex scenarios, they can flexibly combine atomic commands to meet specific needs, further improving operation and maintenance efficiency and adaptability.

[0031] In some embodiments of the present application, the training process of the self-healing large language model is as follows: First, obtain sample alarm information and atomic operation and maintenance commands for the sample alarm information.

[0032] Then, the sample alarm information is used as sample input, and the atomic operation and maintenance command of the sample alarm information is used as sample output to train the initial self-healing large language model to obtain the initial trained self-healing large language model.

[0033] The initial training self-healing large language model is trained according to the self-healing success rates of the sample fault operation and maintenance commands and the self-healing success rates of the sample atomic operation and maintenance commands to obtain the self-healing large language model.

[0034] In summary, the self-healing large language model has the ability to generate atomic operation and maintenance commands for alarm information, thereby making up for the lack of fault types in the atomic operation and maintenance command library. It can generate atomic operation and maintenance commands more flexibly based on actual network alarm information.

[0035] In addition, it has the ability to predict the self-healing success rate of fault operation and maintenance instructions, as well as the ability to predict the self-healing success rate of atomic operation and maintenance commands.

[0036] The self-healing large language model derived from the aforementioned training process can perceive changes in network status in real time and, with the help of AI, dynamically trigger corresponding O&M strategies based on complex and ever-changing warning signals. This flexible triggering mechanism breaks through the limitations of traditional solutions, enabling a transition from fixed rules to intelligent decision-making, ensuring that O&M operations are always precisely matched to the actual needs of the network.

[0037] S102: Search and match the alarm information of the SDN controller in a preset historical knowledge base and a preset atomic operation and maintenance command library using the self-healing large language model.

[0038] In some embodiments of the present application, the AI service is specifically divided into a front-end interceptor, an AI large language model, and a post-interceptor.

[0039] Based on this, in the front-end interceptor, the initial alarm information is preprocessed and denoised to be assembled into a data format that can be recognized by the self-healing large language model.

[0040] Therefore, first, the initial alarm information uploaded by the SDN controller is cleaned. Then, the cleaned initial alarm information is subjected to noise reduction to obtain the noise-reduced alarm information. The noise-reduced alarm information is then converted into a format recognizable by the self-healing large language model to obtain the aforementioned alarm information.

[0041] S103: When there are matching target fault operation and maintenance commands and target atomic operation and maintenance commands, the target fault operation and maintenance commands and the target atomic operation and maintenance commands are processed by the self-healing large language model to obtain an operation and maintenance command with the highest self-healing success rate.

[0042] The target fault O&M command and the target atomic O&M command are input into the self-healing large language model to obtain the self-healing success rate of the target fault O&M command and the self-healing success rate of the target atomic O&M command. These commands are then sorted by self-healing success rate to obtain the O&M command with the highest self-healing success rate. In other words, among all target fault O&M commands and target atomic O&M commands, the command with the highest self-healing success rate is selected. This command type is either the target fault O&M command or the target atomic O&M command.

[0043] In some embodiments of the present application, when there is no matching target atomic operation and maintenance command, the alarm information is input into the self-healing large language model to generate an update atomic operation and maintenance command for the alarm information, and the update atomic operation and maintenance command is determined as the target atomic operation and maintenance command.

[0044] When there is no matching target fault operation and maintenance command, the operation and maintenance command of the fault type in the alarm information is searched on the Internet, and the searched operation and maintenance command is determined as the target atomic operation and maintenance command.

[0045] It should be noted that the pre-interceptor can pre-process the searched operation and maintenance commands and intercept illegal content. That is, the pre-interceptor verifies the legitimacy of the operation and maintenance commands of the fault type in the alarm information searched from the Internet. After passing the verification, the searched operation and maintenance command is determined as the target atomic operation and maintenance command.

[0046] S104: Determine the operation and maintenance command with the highest self-healing success rate as the self-healing execution strategy of the alarm information, and send the self-healing execution strategy to the SDN controller.

[0047] In some embodiments of the present application, the SDN controller sends the self-healing execution policy to the network device through the MCP protocol.

[0048] Receive feedback from network devices. If the feedback indicates an unresolved fault, generate an operation and maintenance work order with alarm information and send the operation and maintenance work order to the operation and maintenance client.

[0049] When the feedback result is to resolve the fault, the alarm information and self-healing execution strategy are added to the preset historical knowledge base.

[0050] SDN controllers can deeply integrate with cloud computing, big data, and artificial intelligence technologies to achieve dynamic resource allocation and management, enhancing the intelligence of network services. Therefore, the combination of SDN and large AI models can be used to resolve alarms or failures in network communication equipment. During operation, SDN controllers detect alarms in real time, conduct pre-diagnostic analysis, intelligently self-recover from faults, and connect to surrounding systems for rapid recovery.

[0051] It should be noted that although the embodiments of this application are based on Figure 1 Steps S101 to S104 are described in sequence, but this does not mean that steps S101 to S104 must be performed in a strict order. Figure 1 The order shown in FIG1 is to introduce and explain step S101 to step S104 in order to facilitate those skilled in the art to understand the technical solution of the embodiment of the present application. In other words, in the embodiment of the present application, the order between step S101 to step S104 can be appropriately adjusted according to actual needs.

[0052] pass Figure 1 This approach, on the one hand, builds a complete Software-Defined Network (SDN) operations and maintenance system, enabling centralized management and dynamic allocation of network resources. From network status monitoring to fault diagnosis and self-healing execution, the entire process seamlessly integrates, forming an efficient, closed-loop operations and maintenance ecosystem. This systematic operations and maintenance architecture not only improves network stability and availability, but also significantly reduces operations and maintenance costs and complexity.

[0053] On the other hand, users can freely combine different types of atomic operation and maintenance commands according to their needs, achieving highly customized operation and maintenance operations. This flexibility gives users great autonomy and provides them with a path for expansion. Whether it is simple tasks or complex scenarios, they can flexibly combine atomic commands to meet specific needs, further improving operation and maintenance efficiency and adaptability. In addition, when the atomic command library is insufficient, atomic operation and maintenance commands can be generated in a timely manner for the current alarm information, further improving the flexibility and adaptability of the self-healing execution strategy.

[0054] On the other hand, the customized self-healing large language model significantly improves the success rate and accuracy of network self-healing.

[0055] Specifically, the system not only relies on real-time alarm and fault information but also builds a historical knowledge base and atomic operation and maintenance command library, providing a rich source of information for reasoning in the self-healing large language model. Furthermore, it can obtain additional knowledge (such as solutions to fault types) through online searches, further expanding the information dimension.

[0056] In addition, the traditional hard-coded rigid model has been abandoned, and the model's intelligent decision-making logic has been introduced. It can match different operation and maintenance commands based on real-time network status and warning information, and predict the self-healing success rate of different operation and maintenance commands, thereby dynamically selecting the most appropriate and optimal self-healing execution strategy.

[0057] Obviously, through the flexible deployment and assembly of operation and maintenance commands, the system's fault tolerance and adaptability have been enhanced, enabling it to flexibly respond to various complex network environments, thereby achieving more efficient and intelligent network self-healing, and realizing the leap from fixed rules to intelligent decision-making, ensuring that operation and maintenance operations are always accurately matched with the actual needs of the network.

[0058] The network self-healing method based on SDN controllers, with its efficient and intelligent features, has demonstrated significant application value in numerous key sectors. In data centers, it enables dynamic resource allocation and rapid fault remediation, ensuring the stable operation of critical services. In the IoT, it monitors device status in real time and automatically fixes connectivity issues, ensuring the smooth operation of infrastructure such as smart cities. In industrial automation, this method helps optimize production processes and reduce equipment maintenance costs. In cloud computing environments, it supports elastic resource management and ensures multi-tenant network isolation and security. In 5G networks, it optimizes slice management and supports efficient edge computing. In the financial sector, it ensures high availability of transaction systems and ensures data security and compliance. In healthcare, it supports telemedicine and ensures stable equipment operation. In education, it optimizes online education platforms and ensures campus network stability. In traffic management, it ensures the stability and efficiency of intelligent transportation systems and the Internet of Vehicles. In the energy sector, it ensures the stable operation of smart grids and energy facility monitoring systems. Overall, this network self-healing method can significantly improve network stability and business continuity, providing strong support for the stable development of various industries.

[0059] More intuitively, Figure 2 A flowchart of another network self-healing method based on an SDN controller provided in an embodiment of the present application.

[0060] exist Figure 2In the process, data sources, whether manually processed or source data, are stored in the database, providing a rich repository of data experience for subsequent self-healing models. The acquired data sources undergo preprocessing and noise reduction to assemble them into a data format recognizable by the self-healing large language model. The self-healing model (self-healing large language model) intelligently selects a self-healing solution and executes it on the SDN controller. The results are fed back to the self-healing large language model, and if successful, the fault is recovered. Faults and alarms that cannot be resolved by the self-healing large language model are further generated as maintenance tickets and notified to the corresponding maintenance engineers via email or other means for manual resolution.

[0061] In summary, the self-healing large language model is constructed by labeling and classifying multiple built-in atomic O&M commands, combining an expert system with a classification algorithm. When a real-time alarm is received, it is matched with the atomic O&M command, atomic O&M command set, or fault O&M command with the highest self-healing success rate. This meets users' basic daily O&M needs while also supporting the configuration of user-specific policies. By combining and reusing atomic policies, policy templates are created for different O&M scenarios, meeting the unique needs of different devices in different O&M scenarios.

[0062] More intuitively, Figure 3 A schematic diagram of a network self-healing system based on an SDN controller provided in an embodiment of the present application.

[0063] exist Figure 3 This method provides defense and recovery services when SDN network failures and alarms occur. The system utilizes a layered architecture, primarily comprising network self-healing applications at the client, server, and application layers. The underlying layer of the network self-healing application is supported by AI services, giving the system powerful self-healing decision-making capabilities. The integration of the SDN controller and AI services is the core of this application.

[0064] AI services are specifically divided into a front-end interceptor, an AI big model (a self-healing big language model), and a back-end interceptor. The front-end interceptor is responsible for pre-processing data obtained from web crawling, uploads from SDN controller terminals, and online searches. It also provides security protection to block illegal content. The AI big model utilizes an open source AI big model, deployed locally or by calling the big model interface. Combined with expert system knowledge, it infers solutions to network alarms or faults. To better adapt the AI big model to SDN network self-healing solutions, various optimizations have been implemented. For example, RAG technology is used to search the knowledge base to reduce the big model's hallucination problem. The knowledge base retrieved can be the enterprise's internal business private knowledge base.

[0065] Furthermore, as a management and control protocol, MCP allows the AI service to detect network failures and alarms. The MCP client can then request the latest self-healing strategies from the server and execute commands and operations locally. The AI service module also utilizes a post-processor interceptor for its own storage and logging. External dependencies primarily include vector databases, the internet, and API services.

[0066] More intuitively, Figure 4 A flowchart of a rule-based network self-healing method in the prior art provided in an embodiment of the present application is provided.

[0067] exist Figure 4 The process begins with capturing real-time, accurate alerts, then performs noise reduction on the alert data to accurately determine the alert type and severity. For routine alerts, the corresponding self-healing rules are directly triggered, and the platform automatically executes recovery operations, resolving the issue efficiently and quickly. For severe and complex alerts, operations and maintenance managers are promptly notified through alert notifications and the generation of maintenance work orders, enabling them to quickly intervene and address the situation manually.

[0068] Obviously, most of the above solutions pre-set fixed rules for various warnings or faults, making it difficult to intelligently select the optimal self-healing strategy based on specific warning situations. Furthermore, the existing rules have low fault tolerance, making it difficult to freely adjust and flexibly assemble rules, making them difficult to adapt to complex and changing network environments. Finally, the information source for faults and alarms is limited, making it impossible to compare with similar faults and alarms, thus hindering the ability to make optimal self-healing decisions.

[0069] It should be noted that if the existing technology simply uses complex algorithm models to improve the ability to predict and respond to potential failures, however, the overall prediction and response capabilities are weak, not only does it fail to improve the self-healing ability of the entire network, but it also increases additional workload.

[0070] Based on the same idea, some embodiments of the present application also provide devices and non-volatile computer storage media corresponding to the above methods.

[0071] Figure 5 A schematic diagram of the structure of a network self-healing device based on an SDN controller provided in an embodiment of the present application includes: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform any one of the above-mentioned network self-healing methods based on the SDN controller.

[0072] Some embodiments of the present application provide a network self-healing non-volatile computer storage medium based on an SDN controller, storing computer-executable instructions, wherein the computer-executable instructions can execute any of the above-mentioned network self-healing methods based on an SDN controller.

[0073] The various embodiments in this application are described in a progressive manner. Similar portions between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the device and medium embodiments are generally similar to the method embodiments, so their descriptions are relatively simple. For relevant portions, refer to the descriptions of the method embodiments.

[0074] The devices and media provided in the embodiments of the present application correspond one-to-one to the methods. Therefore, the devices and media also have similar beneficial technical effects to their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.

[0075] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0076] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0077] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1The function specified in one or more boxes.

[0078] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0079] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0080] Memory may include non-permanent storage in a computer-readable medium, in the form of random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0081] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can be implemented using any method or technology to store information. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change RAM (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0082] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0083] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the technical principles of the present application should fall within the scope of protection of the present application.

Claims

1. A network self-healing method based on an SDN controller, characterized in that: The method comprises: Embed a preset historical knowledge base and a preset atomic operation and maintenance command library into a pre-trained self-healing large language model; the preset historical knowledge base includes fault operation and maintenance commands corresponding to historical fault types, and the preset atomic operation and maintenance command library includes atomic operation and maintenance commands corresponding to different fault types; The alarm information of the SDN controller is retrieved and matched in a preset historical knowledge base and a preset atomic operation and maintenance command library through the self-healing large language model; When there are matching target fault operation and maintenance commands and target atomic operation and maintenance commands, the target fault operation and maintenance commands and the target atomic operation and maintenance commands are processed by the self-healing large language model to obtain the operation and maintenance command with the highest self-healing success rate; The operation and maintenance command with the highest self-healing success rate is determined as the self-healing execution strategy of the alarm information, and the self-healing execution strategy is sent to the SDN controller.

2. The method according to claim 1, characterized in that Before embedding the preset historical knowledge base and the preset atomic operation and maintenance command library into the pre-trained self-healing large language model, the method further includes: Obtain sample alarm information and atomic operation and maintenance commands for sample alarm information; The sample alarm information is used as a sample input, the atomic operation and maintenance command of the sample alarm information is used as a sample output, and an initial self-healing large language model is trained to obtain an initial trained self-healing large language model; The initial training self-healing large language model is trained according to the self-healing success rates of the sample fault operation and maintenance commands and the self-healing success rates of the sample atomic operation and maintenance commands to obtain the self-healing large language model.

3. The method according to claim 2, characterized in that The method further comprises: When there is no matching target atomic operation and maintenance command, inputting the alarm information into the self-healing large language model to generate an update atomic operation and maintenance command for the alarm information; The update atomic operation and maintenance command is determined as the target atomic operation and maintenance command.

4. The method according to claim 1, wherein The method further comprises: When there is no matching target fault operation and maintenance command, searching the Internet for an operation and maintenance command of the fault type in the alarm information; The searched operation and maintenance command is determined as the target atomic operation and maintenance command.

5. The method according to claim 1, characterized in that Before searching and matching the alarm information of the SDN controller in a preset historical knowledge base and a preset atomic operation and maintenance command library using the self-healing large language model, the method further includes: Clean the initial alarm information uploaded by the SDN controller; Performing noise reduction processing on the cleaned initial alarm information to obtain noise-reduced alarm information; The noise reduction alarm information format is converted into a format recognizable by the self-healing large language model to obtain the alarm information.

6. The method according to claim 1, characterized in that After sending the self-healing execution policy to the SDN controller, the method further includes: The SDN controller sends the self-healing execution policy to the network device through the MCP protocol; receiving a feedback result of the network device, and generating an operation and maintenance work order for the alarm information when the feedback result indicates an unresolved fault, and sending the operation and maintenance work order to an operation and maintenance client; When the feedback result is that the fault is resolved, the alarm information and the self-healing execution strategy are added to the preset historical knowledge base.

7. The method according to claim 1, characterized in that The method further comprises: The initial alarm information of the SDN controller is processed by the front-end interceptor to obtain the alarm information, and the alarm information is sent to the self-healing large language model; Logging based on post-interceptors.

8. The method according to claim 4, characterized in that The method further comprises: Through the front-end interceptor, the legitimacy of the operation and maintenance commands for fault types in the alarm information searched on the Internet is verified; After verification, the searched operation and maintenance command is determined as the target atomic operation and maintenance command.

9. A network self-healing device based on an SDN controller, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the network self-healing method based on the SDN controller according to any one of claims 1 to 8.

10. A network self-healing non-volatile computer storage medium based on an SDN controller, storing computer executable instructions, characterized in that: The computer-executable instructions can execute the network self-healing method based on the SDN controller according to any one of claims 1 to 8.

Images