Network equipment configuration scheme checking method and device and computer equipment

Abstract

The invention relates to a network equipment configuration scheme checking method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring network topology information and equipment configuration information of a network equipment configuration scheme; formatting the network topology information and the equipment configuration information to obtain structural body data of a network equipment configuration scheme; based on the structural body data, constructing prompt text information of a network equipment configuration scheme; and performing scheme checking processing on the prompt text information through the scheme checking big language model to obtain fault reason information and repair information of the network equipment configuration scheme. According to the method and the device, the input network equipment configuration scheme is subjected to automatic troubleshooting processing through the large language model, and the network equipment configuration scheme can be effectively checked in the scheme forming stage, so that the troubleshooting efficiency in the network equipment configuration process is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for verifying network device configuration schemes. Background Technology

[0002] With the development of computer technology, network technology related to computers is also constantly evolving. Network devices are a crucial component for achieving network interconnection, serving as the cornerstone of network infrastructure. Network devices play a vital role in modern society. They not only connect devices globally but also provide various critical network services. Network devices refer to the hardware used to connect network devices such as computers, servers, printers, routers, and switches. Their primary function is to enable communication and data transmission between different devices. In network device networking, after configuring all devices and enabling actual communication, situations often arise where improper configuration leads to network failures. Each large network consists of dozens or even hundreds of network devices, and each device's configuration items typically number in the hundreds. Therefore, troubleshooting network configurations is a highly time-consuming, tedious, and extremely difficult task.

[0003] However, current methods for verifying network configuration solutions typically involve checking the reliability of the configuration solution based on simulation or actual performance after the network device configuration information is set and the solution is configured. Adjustments are then made based on the results, resulting in low configuration efficiency for network device solutions. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for verifying network device configuration schemes that can effectively improve the efficiency of network device configuration, in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a method for verifying network device configuration schemes, including:

[0006] Obtain network topology information and device configuration information for the network device configuration scheme;

[0007] The network topology information and the device configuration information are formatted to obtain the structure data of the network device configuration scheme.

[0008] Based on the structure data, construct the prompt text information for the network device configuration scheme;

[0009] The prompt text information is processed by a scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification big language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information of different network troubleshooting cases.

[0010] Secondly, this application also provides a network device configuration scheme verification device, including:

[0011] The data acquisition module is used to acquire network topology information and device configuration information of the network device configuration scheme;

[0012] The data construction module is used to format the network topology information and the device configuration information to obtain the structure data of the network device configuration scheme;

[0013] The prompt text construction module is used to construct prompt text information for the network device configuration scheme based on the structure data;

[0014] The scheme verification module is used to perform scheme verification processing on the prompt text information through a scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification big language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information of different network troubleshooting cases.

[0015] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0016] Obtain network topology information and device configuration information for the network device configuration scheme;

[0017] The network topology information and the device configuration information are formatted to obtain the structure data of the network device configuration scheme.

[0018] Based on the structure data, construct the prompt text information for the network device configuration scheme;

[0019] The prompt text information is processed by a scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification big language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information of different network troubleshooting cases.

[0020] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0021] Obtain network topology information and device configuration information for the network device configuration scheme;

[0022] The network topology information and the device configuration information are formatted to obtain the structure data of the network device configuration scheme.

[0023] Based on the structure data, construct the prompt text information for the network device configuration scheme;

[0024] The prompt text information is processed by a scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification big language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information of different network troubleshooting cases.

[0025] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0026] Obtain network topology information and device configuration information for the network device configuration scheme;

[0027] The network topology information and the device configuration information are formatted to obtain the structure data of the network device configuration scheme.

[0028] Based on the structure data, construct the prompt text information for the network device configuration scheme;

[0029] The prompt text information is processed by a scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification big language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information of different network troubleshooting cases.

[0030] The aforementioned network device configuration scheme verification method, apparatus, computer equipment, storage medium, and computer program product first acquire the network topology information and device configuration information of the network device configuration scheme to obtain the specific information of the network device configuration scheme to be verified. Then, the network topology information and device configuration information are formatted to obtain the structure data of the network device configuration scheme. The network topology information and device configuration information are automatically configured in the form of structure data. Then, based on the structure data, prompt text information for the network device configuration scheme is constructed to obtain the input data for model prediction. Finally, the prompt text information is processed by a scheme verification big language model to obtain fault cause information and repair information for the network device configuration scheme. This application constructs corresponding structure data and prompt text data for the network device configuration scheme based on network topology information and device configuration information, thereby automating the troubleshooting of the input network device configuration scheme through a big language model. This allows for effective checking of the network device configuration scheme during the scheme construction stage, thereby improving the troubleshooting efficiency of the network device configuration process. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying 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.

[0032] Figure 1 This is an application environment diagram of the network device configuration scheme verification method in one embodiment;

[0033] Figure 2 This is a flowchart illustrating a network device configuration scheme verification method in one embodiment;

[0034] Figure 3 This is a schematic diagram of the structure data in one embodiment;

[0035] Figure 4 This is a schematic diagram of the structure of the device field in a structured data example.

[0036] Figure 5 This is a schematic diagram illustrating the connection relationship of router devices in one embodiment;

[0037] Figure 6 This is a schematic diagram of the overall structure of the topology editor in one embodiment;

[0038] Figure 7 This is a schematic diagram of a network topology rendered in one embodiment.

[0039] Figure 8 This is a schematic diagram of a device configuration information input interface in one embodiment;

[0040] Figure 9 This is a schematic diagram of a verification result report in one embodiment;

[0041] Figure 10 This is a flowchart illustrating a network device configuration scheme verification method in another embodiment;

[0042] Figure 11 This is a structural block diagram of a network device configuration scheme verification device in one embodiment;

[0043] Figure 12 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0045] This application relates to Artificial Intelligence (AI) technology. AI is the theory, methods, technology, and application systems that utilize digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine capable of reacting in a manner similar to human intelligence. AI studies the design principles and implementation methods of various intelligent machines, enabling them to possess the functions of perception, reasoning, and decision-making.

[0046] Artificial intelligence (AI) is a comprehensive discipline encompassing a wide range of fields, including both hardware and software technologies. Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, pre-trained model technology, operating / interactive systems, and mechatronics. Pre-trained models, also known as large-scale models or foundational models, can be widely applied to downstream tasks across various AI fields after fine-tuning. AI software technologies primarily include computer vision, speech processing, natural language processing, and machine learning / deep learning. This application focuses on natural language processing and machine learning technologies within the broader field of AI.

[0047] Natural Language Processing (NLP) is an important field within computer science and artificial intelligence. It studies the theories and methods for enabling effective communication between humans and computers using natural language. NLP deals with natural language, the language people use in daily life, and is closely related to linguistics; it also involves computer science and mathematics. Pre-trained models, a crucial technique for model training in artificial intelligence, evolved from large language models in NLP. After fine-tuning, large language models can be widely applied to downstream tasks. NLP techniques typically include text processing, semantic understanding, machine translation, question answering, and knowledge graphs.

[0048] Machine Learning (ML) is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and pre-trained learning. Pre-trained models are the latest development in deep learning, integrating all of these techniques.

[0049] Pre-trained models, also known as foundational models or large models, refer to deep neural networks (DNNs) with a large number of parameters. These DNNs are trained on massive amounts of unlabeled data, leveraging the function approximation capabilities of large-parameter DNNs to enable Programmable Transformers (PTMs) to extract common features from the data. Through fine-tuning, efficient parameter fine-tuning (PEFT), and prompt-tuning techniques, they are suitable for downstream tasks. Therefore, pre-trained models can achieve ideal results in few-shot or zero-shot scenarios. PTMs can be categorized according to the data modalities they process, such as language models (ELMO, BERT, GPT), visual models (Swin-transformer, ViT, V-MOE), speech models (VALL-E), and multimodal models (ViBERT, CLIP, Flamingo, Gato). Multimodal models refer to models that establish feature representations for two or more data modalities. Pre-trained models are important tools for outputting AI-generated content (AIGC) and can also serve as a general interface connecting multiple specific task models.

[0050] Distributed training refers to splitting and sharing the workload of training a model across multiple microprocessors. Large models have large parameters and large training data, exceeding the capacity of a single machine, thus requiring distributed parallelism for speedup. Parallel mechanisms include data parallelism (DP), model parallelism (MP), pipeline parallelism (PP), and hybrid parallelism (HP). Architecture designs include parameter server-based, reduction-based, and MPI-based structures.

[0051] Model compression and quantization refer to techniques used to reduce model size and accelerate model inference, thereby lowering the storage and computational costs of the model. Model compression typically includes pruning, low-rank decomposition, and knowledge distillation, while model quantization involves converting floating-point parameters in the model into fixed-point or integer parameters, thus reducing model size and accelerating model inference.

[0052] Adaptive computation refers to automatically adjusting the computational load and accuracy of a model based on different input data, aiming to improve computational efficiency while maintaining model accuracy. Adaptive computation can flexibly adjust the computational load and accuracy of a model on different input data, thereby better balancing computational efficiency and accuracy.

[0053] Model parallel computing refers to distributing the computational tasks of a model to multiple computing devices (such as CPUs, GPUs, TPUs, etc.) for simultaneous computation, thereby accelerating model training and inference. Model parallel computing can effectively utilize computing resources and improve the computational efficiency and training speed of the model.

[0054] The network device configuration scheme verification method provided in this application embodiment can be applied to, for example, Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located in the cloud or on other network servers. Users can edit network device configuration schemes on the front-end application of terminal 102, setting up network devices and their connections, and providing device configuration information. This information is then submitted to server 104 via terminal 102, where server 104 verifies the network device configuration scheme to confirm whether there are any unreasonable settings. First, after receiving the information submitted by terminal 102, server 104 directly retrieves the network topology information and device configuration information of the network device configuration scheme from the information. It then formats the network topology information and device configuration information to obtain the structure data of the network device configuration scheme. Based on the structure data, it constructs the prompt text information for the network device configuration scheme. Finally, it performs scheme verification processing on the prompt text information using a scheme verification language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information from different network troubleshooting cases. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Server 104 can be implemented using a standalone server or a server cluster composed of multiple servers.

[0055] In one exemplary embodiment, such as Figure 2 As shown, a method for verifying network device configuration schemes is provided, which can be applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 201 to 207. Wherein:

[0056] Step 201: Obtain the network topology information and device configuration information of the network device configuration scheme.

[0057] Network device configuration schemes refer to the device connection and configuration schemes used in the network formation process. Network devices are hardware devices used to connect network equipment such as computers, servers, printers, routers, and switches. Their main function is to enable communication and data transmission between different devices. Common network devices include routers, switches, bridges, gateways, firewalls, and wireless access points. A router is a network device used to connect multiple networks, transmitting data packets from one network to another. Routers can select the optimal path for data transmission based on IP address and destination address. A switch is a hardware device used to connect network devices of the same type, sending data packets to the port where the target device is located. Switches improve network transmission efficiency and reduce data packet collisions and loss. A bridge is a device used to connect networks of the same or different types, transmitting data packets and performing signal shaping and forwarding between networks. Bridges can connect multiple networks together, expanding network coverage. A gateway is a device used to connect different types of networks, converting data packets from one network protocol to another. Gateways act as intermediaries between networks, enabling data transmission between networks with different protocols. A firewall is a device or software used to protect network security, blocking unauthorized access and data packet transmission. Firewalls can filter and block network traffic based on its characteristics and packet content, improving network security. A wireless access point is a device used in wireless networks, providing access to wireless devices and enabling them to connect to wired networks. Wireless access points provide Wi-Fi signals, allowing mobile phones, laptops, and other devices to access the internet wirelessly. Network topology information refers to the structural relationships between entities (network devices) in a network. It is the first step in building a computer network and the foundation for implementing various network protocols. It has a significant impact on network performance, system reliability, and communication costs. In computer networks, topology refers to the form and method of connecting nodes. Device configuration information refers to a set of data used to define and control the functions and operations of system devices. This information includes device name, interface information, IP address, subnet mask, gateway address, routing table, etc. Network administrators use configuration information to control the behavior of network devices, such as setting connection methods, determining communication paths, and configuring security policies. Configuration information is usually stored in the device's non-volatile memory and loaded into memory when the device starts up.

[0058] For example, in network device networking, after completing all device configurations and enabling actual communication, network instability often occurs due to improper configuration. Each large network consists of dozens or hundreds of network devices, and each device's configuration items typically number in the hundreds. Therefore, troubleshooting network configurations is a highly time-consuming, tedious, and extremely difficult task. To address this pain point, the applicant proposes a network device configuration scheme verification method to automate the verification process. During verification, the target object first needs to submit specific information about the network device configuration scheme. Therefore, the target object can edit the network topology and device configuration information of the scheme on terminal 102 and then submit this information to server 104. Server 104 then verifies the scheme after obtaining the network topology and device configuration information. In one embodiment, server 104 can collect relevant information about the network device configuration scheme by setting up a web service, i.e., a network topology editor. The target object can access the network topology editor interface on terminal 102 to edit and submit the network topology and device configuration information.

[0059] Step 203: Format the network topology information and device configuration information to obtain the structure data of the network device configuration scheme.

[0060] Formatting refers to the process of adjusting or arranging data for better presentation or processing, specifically including data type conversion, data structure conversion, and data layout conversion. In this application's solution, data formatting refers to converting input network topology information and device configuration information into structured data in a fixed format, using fields within the structure to display various information about the network device configuration scheme. A structure is structured data composed of a batch of data. Each piece of data that makes up the structured data is called a "member" of the structured data, describing the size and meaning of a memory region. In one embodiment, the structured data of the network device configuration scheme can specifically be as follows: Figure 3 The DSL JSON structure is shown below. The descriptions of each field are shown in Table 1.

[0061]

[0062] The most important field in the structure data is Devices, which contains all relevant information about the network devices, most importantly network connection information and configuration details. In one embodiment, the specific structure of the Devices field can be found by referring to... Figure 4 As shown, the descriptions of each field can be found in Table 2 below:

[0063]

[0064] The most important information in the Devices data mentioned above consists of three fields: PrevInstIds, NextInstIds, and Configs. PrevInstIds and NextInstIds describe the current network connection of the devices. Combining the PrevInstIds and NextInstIds of all devices describes the overall network topology. For example... Figure 5 As shown, assume that the InstID of router device A is user-0001-0001, the InstID of router device B is user-0001-0002, and the InstID of router device C is user-0001-0003. Then, the network connections of A can be seen in Table 3 below:

[0065]

[0066] Similarly, the network relationships of B are shown in Table 4 below:

[0067]

[0068] The network relationships of C are shown in Table 5 below:

[0069]

[0070] The Configs field stores detailed configuration information for the network device, which is typically obtained from the network device's control panel using commands such as "show config". Network device configuration information can be categorized as follows:

[0071] Basic configuration information includes device name, device model, device serial number, device IP address, device management username and password, etc. This information serves as the device's basic identification and access control information, used to identify the device and ensure its security.

[0072] Interface configuration information: This includes the configuration information of the device's physical and logical interfaces, such as IP address, subnet mask, Media Access Control Address (MAC), and port number. This information is used to configure the device's connection method and communication parameters with the network.

[0073] Routing configuration information: This includes the device's routing table and routing policies, used to determine the transmission path of data packets. This information typically includes a list of IP address prefixes, routing protocols, and routing priorities.

[0074] Access control configuration information: This includes the device's access control rules and security policies, used to control network access logic. This information may include IP address filtering, port filtering, protocol filtering, etc.

[0075] Advanced traffic control information includes the device's traffic classification method, the priority of tagged traffic, and QoS (Quality of Service) rules.

[0076] For example, this application specifically utilizes a large language model to learn network device configuration-related knowledge, thereby achieving the function of scheme verification. Therefore, after obtaining network topology information and device configuration information, these two types of information can be formatted and transformed into structured data with a fixed format. Through structured data with fixed fields, the large language model can better understand the relevant information of the network device configuration scheme, thereby achieving effective scheme verification processing. In one embodiment, when the target object edits the network topology information and device configuration information of the network device configuration scheme using a network topology editor, after completing the network device configuration scheme, the network device configuration scheme can be tested. After passing the test, the corresponding structured data of the network device configuration scheme can be directly generated on server 104.

[0077] Step 205: Based on the structure data, construct the prompt text information for the network device configuration scheme.

[0078] The prompt text construction process generates prompt text, which is a text fragment or hint provided to initiate a large language model. This hint can be a word, a phrase, a sentence, a paragraph, or even an entire article. The large language model generates a complete text based on this hint, striving to match the theme, tone, and style of the hint as closely as possible. In practical applications, prompt text is usually carefully designed based on user needs and application scenarios to ensure that the generated text meets specific requirements. In this application, different prompt text information needs to be constructed for network topology and device configuration, and then fed in together. This allows the large language model to have a global understanding of all relevant information during the network device configuration verification process, resulting in a more accurate and complete question-and-answer effect.

[0079] For example, in order to input the network device configuration scheme into the large language model for processing, after obtaining the structure data, it is also necessary to construct the prompt text based on the specific content of the structure data. For the network topology information and device configuration information contained in the structure data, different prompt text information can be constructed separately through prompt text construction, and these prompt text information can be summarized.

[0080] Step 207: The prompt text information is processed by the scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification big language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information of different network troubleshooting cases.

[0081] In simple terms, the working principle of a large language model involves inputting text data, preprocessing it, and then training it using a neural network to ultimately obtain a model capable of generating new text. This model can be applied to fields such as natural language generation, machine translation, and speech recognition. The large language model for scheme verification in this application refers to a big data model specifically designed for verifying network device configuration schemes. It can understand the input prompt text and output corresponding response information according to the instructions. Specifically, the scheme verification large language model is trained using historical network device configuration text as training data on a pre-trained general large language model. This historical network device configuration text includes texts from historical network configuration cases and network troubleshooting cases. The texts from historical network configuration cases can include positive network configuration case texts collected through various channels, such as network device configuration manuals and network configuration scheme descriptions. The network device configuration manual texts come from configuration manuals and user manuals of different types of network devices from various security vendors, such as routers, switches, bridges, gateways, firewalls, and wireless access points. These manuals contain standard and complete device configuration instructions. Since these manuals are from official vendor sources, the accuracy and authority of the information are sufficiently guaranteed. The network topology scheme description text is obtained by collecting text data from the official network topology schemes of large enterprises. These schemes include detailed equipment configurations, network topology, and network specifications. Network topology schemes used in practice by large enterprises also possess high accuracy and usability. The network troubleshooting case text is derived from verification cases accumulated by network engineers through manual troubleshooting. These cases include fault phenomena, configuration error information, troubleshooting approaches, and corrective methods, allowing for the extraction of fault cause information and repair information. Using these two types of case information to construct model training data effectively ensures the validity of the corpus.

[0082] For example, the solution of this application can pre-collect text information of historical networking cases and network troubleshooting cases, and construct model training data based on these two types of information, which can effectively ensure the reliability of the model training corpus. Then, the large language model is trained using this model training data to obtain a usable solution verification large language model. After training, the model can make predictions. For the input prompt text information, the prompt text information can be converted into a numerical sequence and input into the model for calculation. If the prediction is that there is no problem with the network device configuration solution, a message of verification passed and no anomalies found will be output. However, if the prediction is that there is indeed a problem with the network device configuration solution, information on the cause of the fault and repair information will be output. In one embodiment, the training process of the solution verification large language model can be improved by means of distributed training, model compression and quantization, and learning rate adjustment, etc., to improve training efficiency and effectiveness.

[0083] The aforementioned network device configuration scheme verification method first obtains the network topology information and device configuration information of the network device configuration scheme to obtain the specific information of the network device configuration scheme to be verified. Then, the network topology information and device configuration information are formatted to obtain the structure data of the network device configuration scheme. The network topology information and device configuration information are automatically configured in the form of structure data. Then, based on the structure data, prompt text information of the network device configuration scheme is constructed to obtain the input data for model prediction. Finally, the prompt text information is processed by a scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. This application constructs the corresponding structure data and prompt text data of the network device configuration scheme based on the network topology information and device configuration information, and then uses a big language model to automatically troubleshoot the input network device configuration scheme. It can complete the effective check of the network device configuration scheme at the scheme construction stage, thereby improving the troubleshooting efficiency of the network device configuration process.

[0084] In an exemplary embodiment, step 201 includes: obtaining a network device configuration scheme generation request; providing a scheme orchestration canvas based on the network device configuration scheme generation request; and receiving configuration information based on the scheme orchestration canvas to obtain network topology information and device configuration information of the network device configuration scheme.

[0085] The network device configuration scheme generation request refers to a request submitted by the target object through a terminal to a server implementing the method of this application, requesting the server to verify the specified network device configuration scheme. The scheme arrangement canvas refers to the interface used for editing network device configuration schemes, allowing the target object to construct a complete network topology on a single page by dragging and dropping lines. Configuration information comprises various information about the network device scheme edited by the target object on the scheme arrangement canvas, including selected network devices, the connection relationships between network devices, and the connection information entered for the network devices. Based on this configuration information, the server can process and obtain the network topology information and device configuration information of the network device configuration scheme.

[0086] For example, when a target on the terminal side wants to verify a specified network device configuration scheme, it can submit a network device configuration scheme generation request to the server, thereby submitting the corresponding network topology information and device configuration information to begin the verification process. Specifically, the network topology information and device configuration information can be submitted via image orchestration. The server provides a scheme orchestration canvas to the target on the terminal side, and the target constructs a complete network topology on the scheme orchestration canvas by dragging and dropping lines. The server then receives the configuration information based on the scheme orchestration canvas to obtain the network topology information and device configuration information of the network device configuration scheme. In the construction process of the network topology information and device configuration information, network device identifiers can be used to represent network devices on the scheme orchestration canvas. The target can drag and drop network device identifiers on the canvas, and after dragging, can also set the connection information between the various network device identifiers on the canvas. The server can then receive the device identifier dragging information and device identifier connection information from the target based on the scheme orchestration canvas to obtain the network topology information of the network device configuration scheme; and then perform graphical rendering processing based on the network topology information to obtain graphical network topology information. The target user can fill in the corresponding device configuration information for each device in the rendered interface, and the server can then receive the device configuration information based on the graphical network topology information feedback. In this embodiment, collecting network topology information and device configuration information through a scheme arrangement canvas in a graphical arrangement manner can effectively improve data collection efficiency.

[0087] In an exemplary embodiment, formatting network topology information and device configuration information to obtain structure data of network device configuration scheme includes: performing debugging processing on network topology information to obtain debugging results; if the debugging results indicate that the debugging is successful, formatting processing on network topology information and device configuration information to determine the structure fields of network device configuration scheme, and obtaining structure data composed of structure fields.

[0088] Debugging refers to the process of checking for connection problems in the network topology information. By debugging, network topology information can be discovered, thereby avoiding the generation of network device configuration schemes containing connection errors.

[0089] For example, after the target object submits network topology information and device configuration information through the scheme arrangement canvas, to avoid erroneous schemes generated due to wiring errors, an automated check can be performed on the network topology information in a debugging manner before formatting. This check aims to determine whether it contains low-level errors such as wiring mistakes. Only after the debugging results indicate successful debugging should the network topology information and device configuration information be formatted to determine the structure fields of the network device configuration scheme and obtain the structure data of the network device configuration scheme. In one embodiment, the overall structure of the topology editor containing the scheme arrangement canvas can be referenced... Figure 6 As shown, this includes the scheme orchestration canvas (flow orchestration canvas), application selector, graphics rendering engine, application parameter configurator, debugger, version controller, etc. The scheme orchestration canvas is the main part of the topology editor, used to display the entire topology. The device selector is responsible for the application positioning component when dragging network devices into the canvas, ensuring that the correct network device module is selected. The graphics rendering engine is responsible for rendering the applications and connections graphically on the canvas, and also needs to ensure that the entire network topology is displayed. The specific rendered network topology structure can be found in [reference needed]. Figure 7 As shown. The parameter configurator is a graphical configuration tool required for each application. Users input network device configuration information in text format into the configurator. To access this process, the target user can click on the network device in the rendered network topology and then view the pop-up window. Figure 8The configuration page shown allows users to input configuration information in text format obtained from the device management backend to complete the submission process. Network device configuration information is crucial for the normal operation of the network. Incorrect configuration can lead to network connectivity problems, data transmission failures, and even security risks. The network configuration information input by the target user must be ensured to be correct and complete so that the system can accurately identify configuration problems, pinpoint the cause of the failure, and provide solutions. The debugger is used for real-time debugging after the network topology is orchestrated, to identify topology connectivity issues. The version controller allows different versions of the same topology, enabling switching between versions and recording information on each change. Data storage is used to convert and store network topology information and device configuration information. After the network topology is graphically orchestrated in the visual editor and tested successfully, the entire topology is converted into a DSL JSON structure, which can be saved to the backend data module for subsequent prompt text generation. In this embodiment, debugging the network topology information automates the check, effectively improving the verification results of subsequent network device configuration scheme checks.

[0090] In an exemplary embodiment, step 205 includes: extracting the network topology information field and the device configuration information field from the structure data; filling a first text template based on the network topology information field to generate network topology prompt text information; filling a second text template based on the device configuration information field to generate device configuration prompt text information; and summarizing the network topology prompt text information and the device configuration prompt text information to obtain prompt text information for the network device configuration scheme.

[0091] For example, the first text template and the second text template are used to construct prompt text information for network topology information and device configuration information, respectively. Each template contains general text content and content to be filled. The corresponding prompt text information can be generated by filling the text templates with the field information corresponding to the network topology information and device configuration information. During the filling process, the network topology information field and device configuration information field required for the construction process are first extracted from the structure data. Then, the two types of fields are filled into their respective text templates, and the complete prompt text information is obtained by combining the two filled templates. In one embodiment, for the structure of network topology information, the prompt corpus for network topology information can be constructed as follows: "There is a data structure of network topology relationships that needs to be understood. The PrevInstIds and NextInstIds fields describe the current device's network connection relationship. InstID represents the current device ID, PrevInstIds represents the upstream device ID, and NextInstIds represents the downstream device ID." The corresponding Prompt corpus for device configuration information is as follows: There is configuration information for a batch of network devices. Their topology is determined by the previously mentioned network topology data structure. Please use the network topology data structure and this configuration information to discover faults, find the cause of network connectivity issues, and provide a solution. The configuration information for this batch of network devices is as follows: XXXXXXXXXX. In this embodiment, information fields are used to fill in text templates to construct prompt text information, which can effectively improve the accuracy and efficiency of prompt text information construction, and improve the overall processing efficiency of the network device configuration scheme verification process.

[0092] In an exemplary embodiment, the method further includes: acquiring text information of historical networking cases and network troubleshooting cases; constructing model training data based on the text information of historical networking cases and network troubleshooting cases; and fine-tuning the pre-trained large language model using the model training data to obtain a solution verification large language model.

[0093] For example, in the process of verifying network device configuration schemes based on a large language model, the quality of the training corpus related to network device configuration verification is particularly critical, as it directly affects the output quality of the large language model. To ensure training results, this application uses text information from historical networking cases and network troubleshooting cases as training corpora. During training, this text data can be collected first, and then model training data with a similar prompt text information structure can be constructed based on the text data. This data can then be used to fine-tune the pre-trained large language model to obtain the scheme verification large language model. The construction process of training data can be divided into the construction of positive samples and the construction of negative samples. For the text information from historical networking cases and network troubleshooting cases, the network device configuration manuals and network networking scheme descriptions in historical networking cases generally contain accurate and error-free device configuration schemes, while the texts in network troubleshooting cases may contain schemes with configuration errors. Therefore, when constructing samples, original positive sample data can be constructed based on the text information from historical networking cases, and original negative sample data can be constructed based on the text information from network troubleshooting cases. After constructing different types of samples, data augmentation can be performed on the original positive and negative sample data to obtain model training data. Data augmentation techniques can increase the diversity of training data, thereby improving model performance. For example, synonym replacement, random insertion, and random deletion can be used to generate new training samples. In this embodiment, different types of training samples are constructed by collecting text content related to various network device networking cases, thereby enabling fine-tuning of the pre-trained large language model. Using a pre-trained model can significantly reduce training time and computational resources while improving model performance, and high-quality training corpus can also improve the accuracy of scheme verification for the large language model.

[0094] In an exemplary embodiment, fine-tuning the pre-trained large language model using model training data to obtain a scheme verification large language model includes: inputting model training data into the pre-trained large language model in batches to obtain prediction result data for each model; obtaining model loss based on the data labels of the model training data and the model prediction result data; determining the model gradient based on the model loss; performing gradient accumulation processing on the model gradients of different batches of model training data to obtain the accumulated gradient; and updating the parameters of the pre-trained large language model according to the accumulated gradient to obtain the scheme verification large language model.

[0095] In machine learning model training, the model parameters are adjusted using the output results; this process is called backpropagation. To adjust the model parameters, the loss value of the current forward propagation needs to be calculated first: the difference between the model output and the actual result (the labels of the training data). The loss value is generally calculated using a loss function, and the gradient is the derivative or partial derivative of the loss function with respect to the model parameters. When the gradient is positive, it means that moving (increasing) the parameter in the positive direction will increase the value of the loss function at the current parameter value. When the gradient is negative, it means that moving (decreasing) the parameter in the negative direction will increase the value of the loss function at the current parameter value. Therefore, if the model update direction is the opposite of the gradient, it is the direction in which the loss function decreases fastest with respect to the weights, thus causing the model's loss function to decrease faster, i.e., the model's performance to improve faster. Gradient accumulation is a technique that uses multiple mini-batch gradients to achieve model updates.

[0096] For example, since training large language models requires significant computational resources and time, this application accelerates the training process by using gradient accumulation techniques. This technique accumulates gradients from multiple mini-batches and then updates the model parameters all at once. During model training, the training data is first processed in batches, with one batch of training data input at a time. Then, the model loss and gradient corresponding to each batch of training data are obtained based on the model prediction results. Finally, gradient accumulation is used to update the parameters of the pre-trained large language model all at once, resulting in a large language model for verification. In one embodiment, the model parameter update process can also incorporate learning rate adjustment and pruning / compression techniques. For learning rate adjustment, learning rate adjustment techniques can be used during training to optimize model performance. These techniques dynamically adjust the learning rate based on the training data and the model's convergence, thereby improving the model's convergence speed and performance. Pruning / compression, on the other hand, aims to reduce the storage and computational resource consumption of large language models by removing redundant parameters or introducing sparsity to reduce the model's size, thereby improving its efficiency and performance. Therefore, during the parameter update process, the parameters of the pre-trained large language model can be updated first based on the accumulated gradient to obtain the parameter-updated model. Then, the model convergence data of the parameter-updated model is determined, and the learning rate of the parameter-updated model is updated based on the model convergence data. Finally, the pre-trained large language model is pruned and compressed to obtain the scheme verification large language model. In this embodiment, by combining learning rate adjustment techniques and pruning / compression techniques during the model parameter update process, the training efficiency and performance of the model can be improved.

[0097] In an exemplary embodiment, the method further includes: generating a verification result report based on fault cause information and repair information; finding the report submission path of the network device configuration scheme; and pushing the verification result report based on the report submission path.

[0098] The verification result report is used to visually display the verification results of the network device configuration scheme, while the report submission path is the way to inform the target of the verification result report, such as email, message or software notification.

[0099] For example, after obtaining the fault cause information and corresponding repair information, in order to more intuitively display the verification results of the network device configuration scheme to the target audience, a verification result report can be generated based on the fault cause information and repair information output by the large language model. Then, the report submission path for the network device configuration scheme is located, and the verification result report is pushed based on the report submission path. The specific details of the finally pushed verification result report can be found in [reference needed]. Figure 9 As shown, for each type of verification information input, the corresponding verification result is output. If no abnormality is found, the verification is passed. If an abnormality is found, the corresponding fault cause information and repair information are output so that the target object can repair the detected fault.

[0100] In an exemplary embodiment, step 201 includes: obtaining a network device configuration scheme verification request submitted by the terminal through a network interface; parsing the network device configuration scheme verification request to obtain network topology information and device configuration information of the network device configuration scheme. Pushing the verification result report based on the report submission path includes: pushing the verification result report in a response manner based on the report submission path.

[0101] For example, the submission process of network topology information and device configuration information can specifically obtain network device configuration scheme verification requests through a network interface. Specifically, a trained large language model can be encapsulated and provided as a network interface, such as a Hypertext Transfer Protocol (HTTP) interface. When requesting network device configuration scheme verification, a corresponding request can be constructed, with input parameters such as network topology information and device configuration information written in the request body of the request message. The server equipped with the large language model will parse the request upon receiving it, extract the parameters from the message, process, understand, and predict them, and then return the processed fault cause information and repair information to the caller through a response message. In this embodiment, receiving requests and providing feedback results through a network interface can improve request processing efficiency.

[0102] In one embodiment, a logging process is also included. Logs are recorded to document various faults discovered during the verification and processing of network device configuration schemes. Logs can be graded and categorized based on fault information. For example, they can be categorized into ERROR, WARN, INFO, and DEBUG levels, and further categorized into configuration logs, monitoring logs, alarm logs, and operational logs. The ERROR log is the highest-level error record, indicating a very serious system fault that directly leads to malfunction. ERROR logs require close attention from administrators to ensure timely resolution and maintain normal business system operation. The WARN log is a low-level exception log, indicating that an abnormal process was triggered during system operation, but it does not affect normal system operation, and the next stage of business processes can be executed normally. WARN logs require sufficient attention from administrators, typically indicating a certain risk to system operation and a potential system failure. The INFO log usually records critical system information and retains key operational data during normal system operation; administrators also need to pay attention to this in daily maintenance work. The DEBUG log mainly records various detailed system information, serving a debugging purpose, including detailed parameter information, debugging details, operational return information, and other types of information. Configuration logs record actions such as adding, deleting, and modifying configurations for target objects. Monitoring logs record monitoring activities during system operation. Alarm logs record alarm information during system operation. Runtime logs record the behavior of the entire system during background operation.

[0103] This application also provides an application scenario, using the network device configuration scheme verification method described above as an example. The network device configuration scheme verification method specifically includes:

[0104] When users wish to determine the device configuration scheme for a newly deployed network system, they can log in to the front-end page of the network device configuration scheme verification method in this application on their terminal, generate a network device configuration scheme on the front-end page, and submit the generated scheme to the back-end server of the network device configuration scheme verification method. The back-end server will then use a large language model to verify the scheme and determine its feasibility.

[0105] The overall process for verifying network device configuration schemes in this application can be referred to as follows: Figure 10As shown, the first step is to construct the network topology. Users need to add network devices to the orchestration canvas by dragging and dropping on the front-end page, and set the network relationships between devices using lines. Next, configuration information is entered by double-clicking a device in the network topology diagram and filling in the text-formatted device configuration information in the pop-up dialog box. This includes basic configuration information, interface configuration information, routing configuration information, access control configuration information, and advanced flow control information. After saving this network topology and device configuration information, it is automatically formatted, converting the graphical network topology into a JSON data structure. This data structure contains all network connections, defined using PrevInstlds, NextInstlds, and Instld. After obtaining the structure data, effective prompt text content can be constructed for the network topology and configuration content, ready for calling the large language model.

[0106] For large language models, it is necessary to construct and train them in advance. Specifically, corpora can be collected from equipment manuals, large-scale network topology schemes, and typical troubleshooting cases to construct samples. Then, through steps such as dictionary building, model construction, and model training, a usable large language model is obtained. Finally, the constructed prompt text content is input into the large language model, which performs the prediction processing. During model training, techniques such as pre-training, data augmentation, optimal model architecture, gradient accumulation, learning rate adjustment, model pruning, and compression can be used to improve model performance. After detection, the results of the detailed detection items of the four major compliance detection categories can be summarized and statistically analyzed to generate the final results. Finally, the final detection results are notified to the user via message, and relevant log information is recorded.

[0107] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0108] Based on the same inventive concept, this application also provides a network device configuration scheme verification device for implementing the network device configuration scheme verification method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more network device configuration scheme verification device embodiments provided below can be found in the limitations of the network device configuration scheme verification method described above, and will not be repeated here.

[0109] In one exemplary embodiment, such as Figure 11 As shown, a network device configuration scheme verification device is provided, comprising:

[0110] The data acquisition module 1102 is used to acquire network topology information and device configuration information of the network device configuration scheme.

[0111] The data construction module 1104 is used to format network topology information and device configuration information to obtain the structure data of the network device configuration scheme.

[0112] The prompt text construction module 1106 is used to construct prompt text information for network device configuration schemes based on structure data.

[0113] The solution verification module 1108 is used to process the prompt text information through the solution verification big language model to obtain the fault cause information and repair information of the network device configuration solution. The training samples of the solution verification big language model include positive configuration solution samples extracted from historical networking cases and negative configuration solution samples extracted from network troubleshooting cases. The negative configuration solution samples include fault cause information and repair information of different network troubleshooting cases.

[0114] In one embodiment, the data acquisition module 1102 is specifically used to: acquire a network device configuration scheme generation request; provide a scheme arrangement canvas based on the network device configuration scheme generation request; receive configuration information based on the scheme arrangement canvas, and obtain network topology information and device configuration information of the network device configuration scheme.

[0115] In one embodiment, the configuration information includes device identifier drag information and device identifier connection information. The data acquisition module 1102 is further configured to: receive the device identifier drag information and device identifier connection information fed back from the scheme arrangement canvas to obtain the network topology information of the network device configuration scheme; perform graphical rendering processing based on the network topology information to obtain graphical network topology information; and receive device configuration information fed back based on the graphical network topology information.

[0116] In one embodiment, the data construction module 1104 is specifically used to: perform debugging processing on the network topology information to obtain debugging results; if the debugging results indicate that the debugging has passed, perform formatting processing on the network topology information and device configuration information to determine the structure fields of the network device configuration scheme and obtain structure data composed of structure fields.

[0117] In one embodiment, the prompt text construction module 1106 is specifically used to: extract the network topology information field and the device configuration information field from the structure data; fill the first text template based on the network topology information field to generate network topology prompt text information; fill the second text template based on the device configuration information field to generate device configuration prompt text information; and summarize the network topology prompt text information and the device configuration prompt text information to obtain prompt text information for the network device configuration scheme.

[0118] In one embodiment, a model training module is further included, which is used to: acquire text information of historical networking cases and network troubleshooting cases; construct model training data based on the text information of historical networking cases and network troubleshooting cases; and fine-tune the pre-trained large language model using the model training data to obtain a solution verification large language model.

[0119] In one embodiment, the model training module is further configured to: construct original positive sample data based on text information of historical networking cases, and construct original negative sample data based on text information of network troubleshooting cases; perform data augmentation processing on the original positive sample data and the original negative sample data to obtain model training data.

[0120] In one embodiment, the model training module is further configured to: input model training data into a pre-trained large language model in batches to obtain prediction result data for each model; obtain model loss based on the data labels of the model training data and the model prediction result data; determine model gradient based on model loss; perform gradient accumulation processing on the model gradients of different batches of model training data to obtain accumulated gradient; and perform parameter update processing on the pre-trained large language model according to the accumulated gradient to obtain a scheme verification large language model.

[0121] In one embodiment, the model training module is further configured to: perform parameter update processing on the pre-trained large language model according to the accumulated gradient to obtain a parameter update model; determine the model convergence data of the parameter update model, update the learning rate of the parameter update model according to the model convergence data, and perform pruning and compression processing on the pre-trained large language model to obtain a scheme verification large language model.

[0122] In one embodiment, a report generation module is further included, which is used to: generate a verification result report based on fault cause information and repair information; find the report submission path of the network device configuration scheme; and push the verification result report based on the report submission path.

[0123] In one embodiment, the data acquisition module 1102 is specifically used to: acquire a network device configuration scheme verification request submitted by the terminal through a network interface; parse the network device configuration scheme verification request to obtain the network topology information and device configuration information of the network device configuration scheme. The report generation module is specifically used to: push the verification result report in a response manner based on the report submission path.

[0124] Each module in the aforementioned network device configuration verification device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0125] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 12 As shown, the computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data related to network device configuration scheme verification. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a network device configuration scheme verification method.

[0126] Those skilled in the art will understand that Figure 12 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0127] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0128] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0129] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0130] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0131] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0132] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0133] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for verifying network device configuration schemes, characterized in that, The method includes: Obtain network topology information and device configuration information for the network device configuration scheme; The network topology information and the device configuration information are formatted to obtain the structure data of the network device configuration scheme. Based on the structure data, construct the prompt text information for the network device configuration scheme; The prompt text information is processed by a scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification big language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information of different network troubleshooting cases.

2. The method according to claim 1, characterized in that, The network topology information and device configuration information for obtaining the network device configuration scheme include: Obtain the network device configuration scheme generation request; Generate a request feedback scheme arrangement canvas based on the network device configuration scheme; Receive configuration information fed back based on the scheme orchestration canvas to obtain network topology information and device configuration information of the network device configuration scheme.

3. The method according to claim 2, characterized in that, The configuration information includes device identifier drag-and-drop information and device identifier connection information; The step of receiving configuration information based on the scheme orchestration canvas feedback to obtain the network topology information and device configuration information of the network device configuration scheme includes: Receive device identifier drag information and device identifier connection information based on the scheme arrangement canvas to obtain the network topology information of the network device configuration scheme; Based on the network topology information, graphical rendering processing is performed to obtain graphical network topology information; Receive device configuration information based on the graphical network topology information.

4. The method according to claim 1, characterized in that, The process of formatting the network topology information and the device configuration information to obtain the structure data of the network device configuration scheme includes: The network topology information is processed to obtain debugging results; If the debugging result indicates that the debugging is successful, the network topology information and the device configuration information are formatted to determine the structure fields of the network device configuration scheme, and the structure data composed of the structure fields is obtained.

5. The method according to claim 1, characterized in that, The prompt text information for constructing the network device configuration scheme based on the structure data includes: Extract the network topology information field and device configuration information field from the structure data; Based on the network topology information field, the first text template is filled to generate network topology prompt text information; Based on the device configuration information fields, a second text template is filled to generate device configuration prompt text information; By summarizing the network topology prompt text information and the device configuration prompt text information, the prompt text information for the network device configuration scheme is obtained.

6. The method according to claim 1, characterized in that, The method further includes: Obtain text information of historical networking cases and network troubleshooting cases; Model training data is constructed based on the text information of the historical networking cases and the text information of the network troubleshooting cases; The pre-trained large language model is fine-tuned using the model training data to obtain the scheme verification large language model.

7. The method according to claim 6, characterized in that, The training data for building the model based on the text information of the historical networking cases and the text information of the network troubleshooting cases includes: Original positive sample data is constructed based on the text information of the historical networking cases, and original negative sample data is constructed based on the text information of the network troubleshooting cases. Data augmentation processing is performed on the original positive sample data and the original negative sample data to obtain model training data.

8. The method according to claim 6, characterized in that, The step of fine-tuning the pre-trained large language model using the model training data to obtain the scheme verification large language model includes: The training data of the model is input into the pre-trained large language model in batches to obtain the prediction results data of each model. Based on the data labels of the model training data and the model prediction results data, the model loss is obtained; The model gradient is determined based on the model loss. The gradients of the model training data from different batches are accumulated to obtain the accumulated gradient. The parameters of the pre-trained large language model are updated based on the accumulated gradient to obtain the scheme verification large language model.

9. The method according to claim 8, characterized in that, The step of updating the parameters of the pre-trained large language model based on the accumulated gradient to obtain the scheme verification large language model includes: The parameters of the pre-trained large language model are updated based on the accumulated gradient to obtain the parameter-updated model. Determine the model convergence data of the parameter update model, update the learning rate of the parameter update model based on the model convergence data, and perform pruning and compression processing on the pre-trained large language model to obtain the scheme verification large language model.

10. The method according to any one of claims 1 to 9, characterized in that, The method further includes: A verification result report is generated based on the fault cause information and repair information; Find the report submission path for the network device configuration scheme; The verification result report is pushed based on the report submission path.

11. The method according to claim 10, characterized in that, The network topology information and device configuration information for obtaining the network device configuration scheme include: Obtain the network device configuration scheme verification request submitted by the terminal through the network interface; Parse the network device configuration scheme verification request to obtain the network topology information and device configuration information of the network device configuration scheme; The process of pushing the verification result report based on the report submission path includes: The verification result report is pushed out in a response manner based on the report submission path.

12. A network device configuration scheme verification device, characterized in that, The device includes: The data acquisition module is used to acquire network topology information and device configuration information of the network device configuration scheme; The data construction module is used to format the network topology information and the device configuration information to obtain the structure data of the network device configuration scheme. The prompt text construction module is used to construct prompt text information for the network device configuration scheme based on the structure data; The scheme verification module is used to perform scheme verification processing on the prompt text information through a scheme verification big language model to obtain the fault cause information and repair information of the network device configuration scheme. The training samples of the scheme verification big language model include positive configuration scheme samples extracted from historical networking cases and negative configuration scheme samples extracted from network troubleshooting cases. The negative configuration scheme samples include fault cause information and repair information of different network troubleshooting cases.

13. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 11.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 11.

15. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 11.

Images