An operation and maintenance method, device and access controller

Abstract

The application provides an operation and maintenance method and device and an access controller. The method is applied to the access controller and includes the following steps: obtaining local operation and maintenance data, inputting the local operation and maintenance data into an AI model, obtaining an operation and maintenance decision result, and finally executing an operation and maintenance operation according to the operation and maintenance decision result. The input data of the AI model in the initialization stage is standard operation and maintenance data, and the AI model is updated by the access controller according to historical operation and maintenance data in the operation and maintenance stage. The method embeds the AI model in the access controller, realizes intelligent operation and maintenance of the access controller by using the AI technology, realizes timely response to network problems, avoids user data leakage, and ensures network stability.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and in particular to an operation and maintenance method, apparatus and access controller. Background Technology

[0002] The Access Controller (AC) is the core of a wireless network, responsible for managing Access Points (APs) and forwarding wireless traffic. If the AC experiences overload or unknown problems, it can cause network lag or even prevent users from accessing the internet. With the development of AI (Artificial Intelligence), intelligent network operation and maintenance is becoming increasingly important to ensure stable 24 / 7 network operation.

[0003] Currently, intelligent operation and maintenance (O&M) is generally implemented based on cloud platforms. The main body of cloud-based intelligent O&M consists of two parts: one is the access controller or access point, and the other is the online cloud server. In practical applications, devices in the network periodically report data to the cloud platform, where the cloud server or O&M personnel perform statistical analysis and O&M decisions based on the data, and then distribute subsequent O&M operations to the devices to avoid or resolve problems.

[0004] However, for public clouds, some users worry about exposing their network data and are reluctant to connect, thus rendering intelligent O&M functions unusable. Furthermore, intelligent O&M makes decisions based on reported data, meaning most issues still rely on manual data analysis and judgment by O&M personnel. Additionally, due to the need to alleviate wireless network pressure, the granularity of data reported by various devices is generally large, making it difficult to respond promptly even when problems occur. Summary of the Invention

[0005] To overcome the problems existing in related technologies, this application provides an operation and maintenance method, device and access controller.

[0006] According to a first aspect of the embodiments of this application, an operation and maintenance method is provided, the method being applied to an access controller, the method comprising:

[0007] Obtain local operation and maintenance data;

[0008] The local operation and maintenance data is input into the AI ​​model to obtain the operation and maintenance decision results. The input data of the AI ​​model in the initialization phase is standard operation and maintenance data, and in the operation and maintenance phase, the access controller updates it based on historical operation and maintenance data.

[0009] Perform maintenance operations based on the maintenance decision results.

[0010] According to a second aspect of the embodiments of this application, an operation and maintenance device is provided, the device being applied to an access controller, the device comprising:

[0011] The data acquisition module is used to acquire local operation and maintenance data;

[0012] The decision module is used to input the local operation and maintenance data into the AI ​​model to obtain the operation and maintenance decision results. The input data of the AI ​​model in the initialization phase is standard operation and maintenance data, which is updated by the access controller based on historical operation and maintenance data in the operation and maintenance phase.

[0013] The operation and maintenance module is used to perform operation and maintenance operations based on the operation and maintenance decision results.

[0014] According to a third aspect of the embodiments of this application, an access controller is provided, including: an operating system module, a product driver module, and an AI module;

[0015] The product-driven module is used to obtain local operation and maintenance data from the operating system module and send the local operation and maintenance data to the AI ​​module. The AI ​​module uses an AI model to analyze the local operation and maintenance data and outputs operation and maintenance decision results. The product-driven module calls the operating system module to perform operation and maintenance operations based on the operation and maintenance decision results.

[0016] The AI ​​model's input data during the initialization phase is standard operation and maintenance data, which is then updated by the access controller based on historical operation and maintenance data during the operation and maintenance phase.

[0017] The technical solutions provided by the embodiments of this application may include the following beneficial effects:

[0018] In this embodiment, an AI model is pre-initialized based on standard operation and maintenance data and deployed to the access controller. Subsequently, the access controller iteratively updates the AI ​​model based on historical operation and maintenance data. This application implements an operation and maintenance method based on an access controller integrating an AI model. This method can acquire local operation and maintenance data, input the local operation and maintenance data into the AI ​​model, obtain operation and maintenance decision results, and then execute operation and maintenance operations based on the operation and maintenance decision results.

[0019] This application integrates an AI model into the access controller to achieve intelligent operation and maintenance (O&M) of the access controller. Compared with cloud-based intelligent O&M solutions, firstly, the AI ​​model set up locally on the access controller can acquire richer O&M data with finer granularity, thus enabling further expansion and improvement of cloud-based intelligent O&M functions and enhancing the timeliness of O&M; secondly, this application does not require uploading the access controller's data to the public cloud, thus ensuring the privacy of user data; thirdly, based on the AI ​​model's self-learning capability, this application continuously adjusts and optimizes the AI ​​model using historical O&M data from the access controller in actual application scenarios, ensuring that the O&M solutions based on the AI ​​model are more in line with the needs of actual application scenarios and improving the overall reliability of the O&M solution.

[0020] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0021] The accompanying drawings, which are incorporated in and form part of this application, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0022] Figure 1 A flowchart illustrating the operation and maintenance method provided in the embodiments of this application;

[0023] Figure 2 The AI ​​model update process in a standalone environment provided in the embodiments of this application;

[0024] Figure 3 This is a schematic diagram of the overall structure of the access controller provided in an embodiment of this application;

[0025] Figure 4 This is a schematic diagram of the data flow of the access controller provided in an embodiment of this application;

[0026] Figure 5 This is a schematic diagram illustrating the AI ​​model update process in a multi-machine environment, as provided in an embodiment of this application.

[0027] Figure 6 A schematic diagram of the operation and maintenance device provided in the embodiments of this application. Detailed Implementation

[0028] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0029] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0030] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if” or “suppose” as used herein may be interpreted as “when…” or “when…”.

[0031] The embodiments of this application will now be described in detail.

[0032] This application provides an operation and maintenance method, which is applied to an access controller, such as... Figure 1 As shown, the method includes the following steps:

[0033] Step 110: Obtain local operation and maintenance data;

[0034] Step 120: Input local operation and maintenance data into the AI ​​model to obtain operation and maintenance decision results. The input data of the AI ​​model in the initialization phase is standard operation and maintenance data, which is updated by the access controller based on historical operation and maintenance data in the operation and maintenance phase.

[0035] Step 130: Perform maintenance operations based on the maintenance decision results.

[0036] This embodiment integrates AI computing and analysis functions into the access controller. An AI model is built into the access controller to achieve intelligent operation and maintenance (O&M). Therefore, the access controller no longer needs to upload its O&M data to the cloud platform, thus avoiding user data leakage. Furthermore, the access controller inputs local O&M data into the AI ​​model to make O&M decisions. Compared to cloud-based O&M solutions, the access controller can input richer, more specific, and more timely O&M data into the AI ​​model. Therefore, the AI ​​model-based O&M solution in this embodiment can supplement the O&M functions of cloud-based O&M solutions. Moreover, because the AI ​​model runs locally on the access controller, the access controller has greater flexibility in the timing of O&M data collection. Therefore, the access controller can obtain O&M data with finer time granularity and input it into the AI ​​model for O&M decisions, thereby improving the timeliness of O&M response.

[0037] Specifically, an AI model is pre-created. During the initialization phase, the AI ​​model is initialized using standard operational data, and then deployed to the access controller in the actual user network. In the subsequent operational phase, the access controller iteratively updates and optimizes the AI ​​model based on historical operational data, enabling the AI ​​model to output operational decisions that better meet the actual needs of the user network, thus ensuring the reliability of the overall operational solution.

[0038] In other words, given the vastly different common network environments at different sites, this embodiment first initializes the AI ​​model using standard operation and maintenance data. Then, the access controller uses a linear regression algorithm to continuously adjust and optimize the AI ​​model based on historical operation and maintenance data, thereby enabling the AI ​​model's operation and maintenance decision-making capabilities to better adapt to the site environment.

[0039] The operation and maintenance method in this embodiment can be used to solve various types of problems, including but not limited to uneven network traffic distribution, abnormal terminal traffic, and irreversible device failures. The operation and maintenance process for each type of problem is described below.

[0040] In real-world networks, network traffic is not always evenly distributed. Often, a particular CPU forwarding core experiences excessive utilization, causing lag for some users. To fully utilize the device's CPU performance, this application employs a traffic splitting algorithm to hash network traffic based on the MAC address, IP address, and port number of packets. This distributes network traffic evenly across the access controller's CPUs, thereby maximizing the access controller's forwarding performance.

[0041] Specifically, in this application, the access controller uses an AI model to predict the changing trend of network traffic on the access controller, and determines the target traffic splitting algorithm or target traffic splitting algorithm parameters that have the best splitting effect based on the traffic trend. Then, the access controller changes the splitting algorithm or modifies the splitting algorithm parameters in advance based on the target splitting algorithm or target traffic splitting algorithm parameters output by the AI ​​model, so as to distribute traffic evenly on each CPU as much as possible and maximize forwarding performance.

[0042] In summary, this embodiment provides the following two solutions to the problem of uneven network traffic distribution:

[0043] Solution 1: Pre-configure multiple traffic splitting algorithms on the access controller. When the current traffic splitting algorithm is found to be ineffective, such as when the CPU utilization reaches a threshold, the access controller uses an AI model to predict the target traffic splitting algorithm with the best splitting effect. Then, the access controller performs the corresponding operation and maintenance operation to switch the current traffic splitting algorithm to the target traffic splitting algorithm.

[0044] Specifically, the access controller inputs local network traffic data into the AI ​​model to perform the first traffic splitting operation, obtaining a target traffic splitting algorithm identifier. This first splitting operation includes predicting traffic trends based on local network traffic data and determining the target traffic splitting algorithm identifier based on those trends. Subsequently, the access controller performs a traffic splitting algorithm switching operation based on the target traffic splitting algorithm identifier output by the AI ​​model, switching the current traffic splitting algorithm to the target traffic splitting algorithm.

[0045] Solution 2: When the current traffic splitting algorithm is found to be ineffective, such as when the utilization of a certain CPU reaches a threshold, the access controller uses an AI model to predict the parameters of the target traffic splitting algorithm with the best splitting effect. Then, the access controller performs the corresponding operation and maintenance operation, that is, updates the algorithm parameters of the current traffic splitting algorithm to the parameters of the target traffic splitting algorithm.

[0046] Specifically, the access controller inputs local network traffic data into the AI ​​model to execute the second traffic splitting operation, obtaining the target traffic splitting algorithm parameters. The second traffic splitting operation includes predicting traffic trends based on the local network traffic data and determining the target traffic splitting algorithm parameters based on these trends. Subsequently, the access controller updates the parameters of the current traffic splitting algorithm to the target traffic splitting algorithm parameters output by the AI ​​model.

[0047] In the current network, there is a lot of abnormal traffic caused by accidental operations, as well as a lot of malicious attack traffic. When a large amount of abnormal traffic occurs, it usually takes a certain amount of time to investigate and finally find out which terminals are sending the traffic before taking action. In this application, the access controller uses an AI model to monitor terminal traffic and takes timely action when abnormal traffic and abnormal terminals are detected. For example, it can kick abnormal terminals offline or limit the rate of abnormal traffic based on terminal identifiers such as the five-tuple, so as to ensure that network congestion does not occur.

[0048] In summary, this embodiment provides the following solutions to the problem of abnormal terminal traffic:

[0049] The access controller inputs terminal traffic data into the AI ​​model to perform anomaly detection, obtaining an operation and maintenance (O&M) operation identifier and a terminal identifier. The anomaly detection operation includes determining whether the terminal traffic data is abnormal. If the determination is positive, the O&M operation identifier and the abnormal terminal identifier are determined based on the terminal traffic data. Then, the access controller executes the first O&M operation on the abnormal terminal based on the O&M operation identifier and the terminal identifier. The first O&M operation is either rate limiting or disconnection.

[0050] In terms of device management, the access controller in this embodiment can utilize AI models to achieve fault prevention and fault handling. Specifically, when encountering irreversible problems, the access controller calls the AI ​​model to make operation and maintenance decisions based on device status data, and executes avoidance measures without affecting customer use to ensure stable operation of the device. For example, it restarts processes suspected of memory leaks and records them; for networking environments such as dual-link and cloud clusters, it actively triggers operations such as primary / backup switching or the original primary restarting itself.

[0051] In summary, this embodiment provides the following solutions for irreversible equipment failures:

[0052] The access controller inputs device status data into the AI ​​model to perform an irreversible problem detection operation, obtaining an operation and maintenance (O&M) identifier. This operation includes determining whether an irreversible error has occurred based on the device status data. If the determination is yes, the access controller determines the O&M identifier for a second O&M operation based on the device status data. Finally, the access controller executes the second O&M operation based on the O&M identifier. This second O&M operation is either a restart or a primary / standby switchover.

[0053] Considering the high CPU requirements of the AI ​​model update process, as a preferred implementation, the access controller of this application selectively activates the AI ​​model update function. Specifically, this embodiment provides multiple AI model update methods for different networking environments. Each update method is described below.

[0054] In the first update method, in a relatively simple network environment, such as a standalone environment, the access controller only updates the AI ​​model when it is in an idle state. Specifically, each time the AI ​​model outputs an operational decision based on local operational data, the access controller determines its own device status and then... Figure 2 Update the steps shown:

[0055] Step 210: Access controller determines its own device status;

[0056] Step 220: If the device is idle, update the AI ​​model using historical maintenance data, current local maintenance data, and maintenance decision results.

[0057] Step 230: If the device status is busy, only the local operation and maintenance data and results of this operation and maintenance will be retained as historical operation and maintenance data.

[0058] In the second update method, if the access controller is located in a multi-machine network environment, the access controller first determines its device role within the network. If it is a standby device, the access controller updates its AI model using historical operation and maintenance data and synchronizes the updated AI model to the access controller whose device role is the primary device in the current network. If it is the primary device, it updates its own AI model based on the AI ​​model of the access controller whose device role is standby in the current network. In this way, only the standby devices in the network need to update their AI models; the primary device does not need to update its AI model. Each time the standby device updates its AI model, it synchronizes the updated AI model to the primary device.

[0059] In the third update method, if the access controller implements the above maintenance methods using an external AI chip (i.e., hardware), the AI ​​chip card is generally directly connected to the CPU via PCIe, making management easier and facilitating access to historical maintenance data on the access controller without consuming forwarding performance. Therefore, in this networking environment, device status does not need to be considered, and the access controller can be directly configured to update the AI ​​model periodically.

[0060] Finally, this embodiment can also record operation and maintenance data. Specifically, the access controller records the operation and maintenance decision results locally after each AI model outputs them, and generates a log in the memory flash every 24 hours to facilitate daily inspections by operation and maintenance personnel.

[0061] The operation and maintenance method of this embodiment has been introduced above. The following section uses a practical application as an example to further illustrate the operation and maintenance method provided in this application. This embodiment divides the entire operation and maintenance solution into three parts, each of which will be described below.

[0062] Part 1: Overall Structure and Operation / Maintenance Process of the Access Controller

[0063] The overall structure of the access controller in this embodiment is as follows: Figure 3 As shown, it includes an operating system module, a product driver module, and an AI module.

[0064] The overall operation and maintenance process is as follows: Figure 4 As shown, the product-driven module first collects operation and maintenance data from the operating system module, and then aggregates it together with the operation and maintenance data it compiles to obtain local operation and maintenance data. The product-driven module then sends the local operation and maintenance data to the AI ​​module for operation and maintenance decision-making. The AI ​​module then sends the operation and maintenance decision results back to the product-driven module. The product-driven module executes relevant operation and maintenance operations based on the operation and maintenance decision results to ensure network stability.

[0065] Part Two: Training Methods for AI Models

[0066] Due to the limitations of AC hardware computing power, the AI ​​model in this embodiment mainly consists of multiple rules. For example, to address the issue of uneven network traffic distribution, one rule of the AI ​​model in this embodiment can be set as follows: when the actual value of a certain CPU utilization on the access controller exceeds a threshold, it is determined that the access controller has an uneven network traffic distribution problem, thereby achieving monitoring of network traffic distribution. Specifically, this embodiment first initializes the AI ​​model using standard operation and maintenance data, initializing the CPU utilization threshold. Then, the access controller updates the AI ​​model using historical operation and maintenance data, continuously adjusting the threshold to make its value more reasonable and better suited to the network's actual needs.

[0067] It should be noted that the initialization of the AI ​​model mentioned in this application refers to the process of pre-training the AI ​​model using standard operation and maintenance data, such as laboratory data. This embodiment collects and organizes equipment operation and maintenance data from different industry usage scenarios, and uses this equipment operation and maintenance data as standard operation and maintenance data to initialize the AI ​​model. Alternatively, project test data and acceptance test data can also be used as standard operation and maintenance data to initialize the AI ​​model.

[0068] Part Three: How to Update AI Models

[0069] In this embodiment, the AI ​​model is selectively updated based on different network environments. Specifically, the access controller updates the AI ​​model when the following conditions are met: the device is idle, the device role is a backup device in a network environment with primary and backup devices, and the device type is a device with an external AI chip. In summary, this embodiment divides the network environment into three categories: single-machine environment, multi-machine environment, and access controller containing an AI chip. Each environment will be described below.

[0070] In a standalone environment, when the access controller is idle, it updates the AI ​​model using the retained historical operation and maintenance data.

[0071] For multi-machine environments such as dual-link, IRF (Intelligent Resilient Framework), and cloud clusters, the access controller first determines whether its network environment is a multi-machine environment. If it is, it activates the information transmission module to establish a heartbeat connection with the information transmission modules of other access controllers in the network. Figure 5 As shown. If one party suddenly disconnects, the network environment of the access controller is considered to have changed to a standalone environment, and it will operate in standalone environment mode until the heartbeat is restored.

[0072] Operation and maintenance process in a multi-machine environment, such as Figure 5 As shown, the primary device sends historical operation and maintenance data to the backup device through (1)(2)(3) to update the AI ​​model. The backup device updates its own AI model based on the received historical operation and maintenance data, and then synchronizes the updated AI model to the primary device through (4)(2)(5).

[0073] It is worth mentioning that if the CPU pressure on the backup device is high, the access controller of the backup device can stop updating its own AI model until it is idle. Then, the access controller of the backup device wakes up the information transmission module, obtains the historical operation and maintenance data of the primary device, and updates its own AI model based on the historical operation and maintenance data.

[0074] Access controllers incorporating AI chips are mainly divided into two categories: AI chip plug-in cards and access controllers with integrated AI chips. This type of solution is more expensive, but with the support of AI chip hardware, the access controller can make more complex operational decisions. For example, when users configure commands using command lines, the system can check in real time whether the command combinations configured by the user are incorrect and provide timely prompts. Therefore, for access controllers with AI chips, this embodiment sets the AI ​​model to be updated periodically to ensure optimal AI model performance.

[0075] As can be seen from the above technical solutions, this application applies AI technology to the operation and maintenance solution of the access controller. The network maintenance and optimization based on the AI ​​model has the following advantages: the granularity of operation and maintenance data is controllable, making it easier to respond and handle problems in a timely manner; it can prevent equipment failures or network congestion in advance, ensuring network stability; it does not require uploading device data, thus protecting the privacy of users; and it can better help operation and maintenance personnel understand the status of the equipment in the live network, such as the overall performance of the equipment under the business traffic model.

[0076] Based on the same inventive concept, this application also provides an operation and maintenance device, which is applied to an access controller, and its structural schematic diagram is shown below. Figure 6 As shown, the device includes:

[0077] Data acquisition module 610 is used to acquire local operation and maintenance data;

[0078] The decision module 620 is used to input the local operation and maintenance data into the AI ​​model to obtain the operation and maintenance decision result. The input data of the AI ​​model in the initialization phase is standard operation and maintenance data, and in the operation and maintenance phase, it is updated by the access controller based on historical operation and maintenance data.

[0079] The operation and maintenance module 630 is used to perform operation and maintenance operations based on the operation and maintenance decision results.

[0080] As one specific implementation method, it also includes:

[0081] The model update module is used to determine the device status of the access controller, and when the device status is idle, it updates the AI ​​model using historical operation and maintenance data.

[0082] As one specific implementation, it also includes a model update module, which includes:

[0083] An environment determination unit is used to determine the device role of the access controller in the current network if the network environment in which the access controller is located is a multi-machine environment.

[0084] The update unit is used to update its own AI model using historical operation and maintenance data if the access controller is a backup device in the current network, and to synchronize the updated AI model to the access controller whose device role is the primary device in the current network.

[0085] The synchronization unit is used to update its own AI model based on the AI ​​model of the access controller whose device role is a backup device in the current network if the access controller is a primary device in the current network.

[0086] As a specific implementation method, the decision-making module 620 and the operation and maintenance module 630 are specifically used for:

[0087] Local network traffic data is input into the AI ​​model to perform the first traffic splitting operation and obtain the target traffic splitting algorithm identifier. The first traffic splitting operation includes predicting traffic trends based on local network traffic data and determining the target traffic splitting algorithm identifier based on the traffic trends.

[0088] A traffic splitting algorithm switching operation is performed based on the target traffic splitting algorithm identifier, switching the current traffic splitting algorithm to the target traffic splitting algorithm.

[0089] As a specific implementation method, the decision-making module 620 and the operation and maintenance module 630 are specifically used for:

[0090] Local network traffic data is input into the AI ​​model to perform the second traffic splitting operation and obtain the target traffic splitting algorithm parameters. The second traffic splitting operation includes predicting traffic trends based on local network traffic data and determining the target traffic splitting algorithm parameters based on the traffic trends.

[0091] Based on the target traffic splitting algorithm parameters, the parameters of the current traffic splitting algorithm are updated to the target traffic splitting algorithm parameters.

[0092] As a specific implementation method, the decision-making module 620 and the operation and maintenance module 630 are specifically used for:

[0093] Terminal traffic data is input into an AI model to perform anomaly detection operations, resulting in an operation and maintenance operation identifier and a terminal identifier. The anomaly detection operation includes determining whether the terminal traffic data is abnormal. If the determination result is yes, the operation and maintenance operation identifier of the first operation and maintenance operation and the terminal identifier of the abnormal terminal are determined based on the terminal traffic data. The first operation and maintenance operation is to limit the speed or take the device offline.

[0094] Based on the operation and maintenance operation identifier and the terminal identifier, perform the first operation and maintenance operation on the abnormal terminal.

[0095] As a specific implementation method, the decision-making module 620 and the operation and maintenance module 630 are specifically used for:

[0096] The device status data is input into the AI ​​model to perform an irreversible problem detection operation to obtain an operation and maintenance operation identifier. The irreversible problem detection operation includes determining whether the device has an irreversible error based on the device status data. If the determination result is yes, the operation and maintenance operation identifier of the second operation and maintenance operation is determined based on the device status data. The second operation and maintenance operation is a restart or a primary / backup switch.

[0097] Execute the second maintenance operation based on the maintenance operation identifier.

[0098] This application also provides an access controller, including: an operating system module, a product driver module, and an AI module;

[0099] The product-driven module is used to obtain local operation and maintenance data from the operating system module and send the local operation and maintenance data to the AI ​​module. The AI ​​module uses an AI model to analyze the local operation and maintenance data and outputs operation and maintenance decision results. The product-driven module calls the operating system module to perform operation and maintenance operations based on the operation and maintenance decision results.

[0100] The AI ​​model's input data during the initialization phase is standard operation and maintenance data, which is then updated by the access controller based on historical operation and maintenance data during the operation and maintenance phase.

[0101] The aforementioned access controller can implement the steps of the above operation and maintenance method.

[0102] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. An operation and maintenance method, characterized by, The method is applied to an access controller, and the method comprises: acquiring local operation and maintenance data; inputting the local operation and maintenance data into an AI model to obtain an operation and maintenance decision result, wherein the input data of the AI model in an initialization stage is standard operation and maintenance data, and the AI model is updated by the access controller according to historical operation and maintenance data in an operation and maintenance stage; performing an operation and maintenance operation according to the operation and maintenance decision result to make traffic evenly distributed to each CPU of the access controller; The method specifically performs the operation and maintenance operation in the following manner: inputting local network traffic data into the AI model to perform a first shunting operation to obtain a target shunting algorithm identifier, wherein a plurality of shunting algorithms are pre-set on the access controller, the first shunting operation comprises predicting a traffic trend according to the local network traffic data, and determining the target shunting algorithm identifier according to the traffic trend; and performing a shunting algorithm switching operation according to the target shunting algorithm identifier to switch the current shunting algorithm to the target shunting algorithm; or, inputting local network traffic data into the AI model to perform a second shunting operation to obtain target shunting algorithm parameters, wherein the second shunting operation comprises predicting a traffic trend according to the local network traffic data, and determining the target shunting algorithm parameters according to the traffic trend; and updating the parameters of the current shunting algorithm to the target shunting algorithm parameters according to the target shunting algorithm parameters.

2. The method of claim 1, wherein, The method specifically updates the AI model in the following manner: determining the device state of the self, and updating the AI model by using the historical operation and maintenance data when the device state is idle.

3. The method of claim 1, wherein, The method specifically updates the AI model in the following manner: if the networking environment in which the self is located is a multi-machine environment, determining the device role of the self in the current networking; if the device role of the self in the current networking is a standby device, updating the AI model of the self by using the historical operation and maintenance data, and synchronizing the updated AI model to the access controller whose device role in the current networking is a master device; if the device role of the self in the current networking is a master device, updating the AI model of the self according to the AI model on the access controller whose device role in the current networking is a standby device.

4. The method of claim 1, wherein, The method specifically performs the operation and maintenance operation in the following manner: inputting terminal traffic data into the AI model to perform an anomaly detection operation to obtain an operation and maintenance operation identifier and a terminal identifier, wherein the anomaly detection operation comprises judging whether the terminal traffic data is abnormal traffic, and determining the operation and maintenance operation identifier of a first operation and maintenance operation and the terminal identifier of an abnormal terminal according to the terminal traffic data when the judgment result is yes, wherein the first operation and maintenance operation is speed limiting or offline; performing the first operation and maintenance operation on the abnormal terminal according to the operation and maintenance operation identifier and the terminal identifier.

5. The method of claim 1, wherein, The method specifically performs the operation and maintenance operation in the following manner: The device state data is input into the AI model to perform an irreversible problem detection operation to obtain an operation identifier of a second operation, wherein the irreversible problem detection operation includes determining whether an irreversible error occurs in the device according to the device state data, and determining the operation identifier of the second operation according to the device state data when the determination result is yes, wherein the second operation is a restart or a master-backup switching; According to the operation identifier, the second operation is performed.

6. An operation and maintenance device, characterized by The device is applied to an access controller, and the device includes: A data acquisition module for acquiring local operation data; A decision module for inputting the local operation data into an AI model to obtain an operation decision result, wherein the input data of the AI model in the initialization stage is standard operation data, and the AI model is updated by the access controller according to historical operation data in the operation stage; An operation module for performing an operation according to the operation decision result to make the traffic evenly distributed to each CPU of the access controller; The operation module specifically performs the operation in the following ways: Local network traffic data is input into the AI model to perform a first shunting operation to obtain a target shunting algorithm identifier, wherein a plurality of shunting algorithms are pre-set on the access controller, the first shunting operation includes predicting a traffic trend according to the local network traffic data, and determining the target shunting algorithm identifier according to the traffic trend; and a shunting algorithm switching operation is performed according to the target shunting algorithm identifier to switch the current shunting algorithm to the target shunting algorithm; Or, Local network traffic data is input into the AI model to perform a second shunting operation to obtain a target shunting algorithm parameter, wherein the second shunting operation includes predicting a traffic trend according to the local network traffic data, and determining the target shunting algorithm parameter according to the traffic trend; and the parameters of the current shunting algorithm are updated to the target shunting algorithm parameter according to the target shunting algorithm parameter.

7. The apparatus of claim 6, wherein, Further comprising: A model updating module for determining the device state of the access controller, and updating the AI model of the access controller by using historical operation data when the device state is idle.

8. The apparatus of claim 6, wherein, Further comprising a model updating module, and the model updating module includes: An environment determining unit for determining the device role of the access controller in the current networking environment if the networking environment of the access controller is a multi-machine environment; An updating unit for updating the AI model of the access controller by using historical operation data if the device role of the access controller in the current networking environment is a standby device, and synchronizing the updated AI model to the access controller whose device role in the current networking environment is a master device; A synchronizing unit for updating the AI model of the access controller according to the AI model of the access controller whose device role in the current networking environment is a standby device if the device role of the access controller in the current networking environment is a master device.

9. An access controller, characterized in that Including: An operating system module, a product driver module, and an AI module; The product driving module is configured to obtain local operation and maintenance data from the operating system module and send the local operation and maintenance data to the AI module, the AI module analyzes the local operation and maintenance data by using an AI model and outputs operation and maintenance decision results, and the product driving module calls the operating system module to perform operation and maintenance operations according to the operation and maintenance decision results, so that traffic is evenly distributed to each CPU of the access controller. The input data of the AI model in the initialization stage is standard operation and maintenance data, and the AI model is updated by the access controller according to historical operation and maintenance data in the operation and maintenance stage. The operating system module specifically performs operation and maintenance operations in the following manner: Local network traffic data is input into an AI model to perform a first shunting operation to obtain a target shunting algorithm identifier, wherein a plurality of shunting algorithms are pre-set on the access controller, the first shunting operation includes predicting a traffic trend according to the local network traffic data and determining a target shunting algorithm identifier according to the traffic trend; and a shunting algorithm switching operation is performed according to the target shunting algorithm identifier to switch the current shunting algorithm to a target shunting algorithm. Alternatively, Local network traffic data is input into an AI model to perform a second shunting operation to obtain target shunting algorithm parameters, wherein the second shunting operation includes predicting a traffic trend according to the local network traffic data and determining target shunting algorithm parameters according to the traffic trend; and the parameters of the current shunting algorithm are updated to the target shunting algorithm parameters according to the target shunting algorithm parameters.

Images