AAA system automatic operation and maintenance method and system

Abstract

The invention discloses an AAA system automatic operation and maintenance method and system. The method comprises the following steps: in response to an operation and maintenance instruction initiated by a target object, obtaining first system operation and maintenance data of an AAA system in a first time period; analyzing the first system operation and maintenance data by using a scene recognition model to obtain a target operation and maintenance scene category of the AAA system in the first time period; determining a plurality of target operation and maintenance cases matched with the target operation and maintenance scene category in the strategy knowledge base, and determining an execution effect of an atomic power arrangement strategy in each target operation and maintenance case; and converting the target atomic power arrangement strategy with the optimal execution effect into a corresponding target operation and maintenance operation configuration file, and issuing the target operation and maintenance operation configuration file to an AAA system. According to the method and the device, the technical problems of low configuration efficiency and error proneness caused by performing operation and maintenance configuration change on a plurality of nodes in the AAA system one by one in a manual mode are solved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and more specifically, to an automated operation and maintenance method and system for an AAA system. Background Technology

[0002] In today's rapidly changing technological environment, especially given the increasing complexity of AAA (Authentication, Authorization, and Accounting) systems in business, expansion, and emergency scenarios, deploying AAA systems manually faces a series of significant technical problems and challenges. These problems directly impact the operational efficiency and accuracy of AAA systems:

[0003] First, when deploying and maintaining a scenario manually, maintenance personnel need to log in to each system node of the AAA system one by one to manually make configuration changes and operations. This process is not only time-consuming and labor-intensive, but also prone to operational errors due to the high degree of human intervention, such as incorrect input of configuration parameters or omission of key steps, which can lead to abnormal system operation or service interruption.

[0004] Secondly, when performing tasks, operations and maintenance personnel often need to rely on memory or consult a large number of documents to determine the correct operating procedures and parameter settings. This not only increases the complexity of operations and maintenance work, but also increases the dependence on personnel experience. New employees or employees without experience find it difficult to get started quickly under this mechanism, which affects the overall operations and maintenance capabilities of the team.

[0005] Furthermore, manual operation lacks an automated verification mechanism. After configuration changes, manual secondary checks or functional tests are often required. This process is not only inefficient, but also makes it difficult to ensure that all possible errors are detected and corrected in a timely manner.

[0006] Furthermore, manual operation and maintenance methods make it difficult to achieve intelligent and automated operation and maintenance decisions, resulting in repetitive and inefficient operation and maintenance work. Moreover, it is difficult to make optimal decisions when facing complex operation and maintenance scenarios, which affects the overall operation and maintenance efficiency of the system and the user experience.

[0007] There is currently no effective solution to the above problems. Summary of the Invention

[0008] This application provides an automated operation and maintenance method and system for an AAA system, which at least solves the technical problem of low configuration efficiency and easy error caused by manually changing the operation and maintenance configuration of multiple nodes in an AAA system one by one.

[0009] According to one aspect of the embodiments of this application, an automated operation and maintenance method for an AAA system is provided, comprising: responding to an operation and maintenance instruction initiated by a target object, acquiring first system operation and maintenance data of the AAA system within a first time period, wherein the first system operation and maintenance data includes: operation and maintenance operation logs, operation and maintenance configuration change information, performance monitoring data, and business indicator data; analyzing the first system operation and maintenance data using a pre-trained scenario recognition model to obtain a target operation and maintenance scenario category of the AAA system within the first time period; determining multiple target operation and maintenance cases in a preset strategy knowledge base that match the target operation and maintenance scenario category, and determining the execution effect of the atomic capability orchestration strategy in each target operation and maintenance case, wherein the strategy knowledge base includes triple information of multiple operation and maintenance cases, and the triple information includes: operation and maintenance scenario category, atomic capability orchestration strategy, and execution effect, and the atomic capability orchestration strategy is used to characterize the execution order between multiple atomic capabilities and the parameter information of each atomic capability; converting the target atomic capability orchestration strategy with the best execution effect into a corresponding target operation and maintenance operation configuration file, and distributing the target operation and maintenance operation configuration file to the AAA system.

[0010] Optionally, the training process of the scene recognition model includes: constructing an initial learning model; obtaining a training sample set and a sample label set, wherein the training sample set includes: second system operation and maintenance data of the AAA system in multiple first historical time periods as training samples, and the sample label set includes: operation and maintenance scenario categories of the AAA system in each first historical time period as sample labels for the corresponding training samples, and the operation and maintenance scenario categories include at least one of the following: data verification, service restart, authentication control, interface emergency processing, database restart, and primary / backup node switching; analyzing multiple training samples in the training sample set using the initial learning model to obtain the prediction result of each training sample, and generating multiple adversarial samples based on the sample label and prediction result of each training sample in the training sample set to update the training sample set and the sample label set; and continuing to iteratively train the initial learning model using the updated training sample set and the updated sample label set until a preset iteration termination condition is met to obtain a trained scene recognition model.

[0011] Optionally, multiple adversarial examples are generated based on the sample labels and prediction results of each training sample in the training sample set to update the training sample set and the sample label set, including: constructing a target loss function based on the sample labels and prediction results of each training sample in the training sample set; for each training sample in the training sample set, calculating the gradient of the target loss function relative to the training sample, and adding perturbation to the training sample along the direction of the gradient to obtain adversarial examples, and using the sample labels of the training samples as the sample labels of the adversarial examples respectively; updating the training sample set and the sample label set based on the adversarial examples and sample labels of the multiple training samples respectively.

[0012] Optionally, the construction process of the strategy knowledge base includes: acquiring the second system status data, second system operation and maintenance data, and operation and maintenance scenario categories of the AAA system in multiple first historical time periods; for each first historical time period, analyzing the second system status data and second system operation and maintenance data using a pre-trained strategy decision model to obtain atomic capability orchestration strategies for the operation and maintenance scenario categories within the first historical time period, whose execution effect is higher than a first preset threshold and meets the preset verification mechanism, wherein the execution effect is determined by at least the execution duration, resource consumption, and success probability, and the strategy decision model is trained based on a deep Q-network algorithm; and constructing a strategy knowledge base by consisting of multiple operation and maintenance cases composed of the operation and maintenance scenario categories of the AAA system in each first historical time period, the executed atomic capability orchestration strategies, and the execution effects corresponding to the atomic capability orchestration strategies.

[0013] Optionally, multiple target operation and maintenance cases that match the target operation and maintenance scenario category within a preset strategy knowledge base are determined, including: calculating the similarity between the operation and maintenance scenario category in each operation and maintenance case within the strategy knowledge base and the target operation and maintenance scenario category; and selecting multiple operation and maintenance cases with a similarity higher than a second preset threshold as multiple target operation and maintenance cases that match the target operation and maintenance scenario category.

[0014] Optionally, the target atomic capability orchestration strategy with the best execution effect is converted into a corresponding target operation and maintenance configuration file, including: determining multiple target atomic capabilities corresponding to the target atomic capability orchestration strategy from a preset atomic capability library, wherein the atomic capability library includes: multiple compliant atomic capabilities of the AAA system under different operation and maintenance scenarios; setting parameters and logically orchestrating multiple target atomic capabilities according to the target atomic capability orchestration strategy to obtain the target operation and maintenance configuration file.

[0015] Optionally, after the target operation and maintenance configuration file is distributed to the AAA system, the method further includes: obtaining the execution results of multiple system nodes of the AAA system executing the target operation and maintenance configuration file, wherein the execution results include: execution time, resource consumption, and whether it was successful; determining the execution effect of the target atomic capability orchestration strategy corresponding to the target operation and maintenance configuration file based on the execution results of each system node executing the target operation and maintenance configuration file; using the operation and maintenance scenario category, the target atomic capability orchestration strategy, and the corresponding execution effect as the triple information of the new operation and maintenance case, and storing the triple information of the new operation and maintenance case in the strategy knowledge base.

[0016] According to another aspect of the embodiments of this application, an automated operation and maintenance system for an AAA system is also provided, comprising: an acquisition module, configured to, in response to an operation and maintenance instruction initiated by a target object, acquire first system operation and maintenance data of the Authentication-Authorization-Billing AAA system within a first time period, wherein the first system operation and maintenance data includes: operation and maintenance operation logs, operation and maintenance configuration change information, performance monitoring data, and business indicator data; an identification module, configured to, using a pre-trained scenario identification model to analyze the first system operation and maintenance data to obtain the target operation and maintenance scenario category of the AAA system within the first time period; a determination module, configured to, determine multiple target operation and maintenance cases that match the target operation and maintenance scenario category within a preset strategy knowledge base, and determine the execution effect of the atomic capability orchestration strategy in each target operation and maintenance case, wherein the strategy knowledge base includes triple information of multiple operation and maintenance cases, and the triple information includes: operation and maintenance scenario category, atomic capability orchestration strategy, and execution effect, and the atomic capability orchestration strategy is used to characterize the execution order between multiple atomic capabilities and the parameter information of each atomic capability; and a configuration module, configured to, convert the target atomic capability orchestration strategy with the best execution effect into a corresponding target operation and maintenance operation configuration file, and distribute the target operation and maintenance operation configuration file to the AAA system.

[0017] According to another aspect of the embodiments of this application, a computer program product is also provided, the computer program product comprising: a computer program, wherein the computer program, when executed by a processor, implements the above-described AAA system automated operation and maintenance method.

[0018] According to another aspect of the embodiments of this application, an electronic device is also provided, the electronic device including: a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the above-described AAA system automated operation and maintenance method through the computer program.

[0019] In this embodiment, by responding to the operation and maintenance instructions initiated by the target object, the first system operation and maintenance data of the AAA system is automatically analyzed, and the target operation and maintenance scenario category of the AAA system is determined through a scenario recognition model. Then, the best atomic capability orchestration strategy is selected from the policy knowledge base, converted into a target operation and maintenance configuration file, and distributed to the AAA system for security management. This achieves the operation and maintenance technical effects of intelligent identification and decision-making, and automated orchestration and execution, thereby improving the automation level of AAA system operation and maintenance, reducing human error, improving system stability and operation and maintenance efficiency. This solves the technical problem of low configuration efficiency and easy error caused by manually changing the operation and maintenance configuration of multiple nodes in the AAA system one by one. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0021] Figure 1 This is a flowchart illustrating an optional automated operation and maintenance method for an AAA system according to an embodiment of this application;

[0022] Figure 2 This is a schematic diagram illustrating the formulation of an optional atomic capability orchestration strategy according to an embodiment of this application;

[0023] Figure 3 This is a schematic diagram of the structure of an optional AAA system automated operation and maintenance system according to an embodiment of this application;

[0024] Figure 4 This is a schematic diagram of the hardware structure of a computer terminal for implementing an automated operation and maintenance method for an AAA system, according to an embodiment of this application. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0026] It should be noted that the terms "first," "second," etc., used in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] To better understand the embodiments of this application, the following is a translation and explanation of some nouns or terms that appear in the description of the embodiments of this application:

[0028] AAA (Authentication, Authorization, and Accounting) is a management framework that provides a security mechanism for authorizing certain users to access specified resources and logging their actions. It is widely used due to its good scalability and ease of centralized management of user information.

[0029] Atomic capabilities are capabilities that an enterprise can independently provide, formed by integrating various resources through internal capability encapsulation and external capability introduction. These capabilities are characterized by integrability, reusability, pricing, scalability, licensability, and broad common requirements, serving as common elements and key links supporting various applications and businesses.

[0030] Example 1

[0031] According to an embodiment of this application, an automated operation and maintenance method for an AAA system is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0032] Figure 1 This is a flowchart illustrating an automated operation and maintenance method for an AAA system according to an embodiment of this application, as shown below. Figure 1 As shown, the method includes the following steps:

[0033] Step S102: In response to the operation and maintenance instruction initiated by the target object, obtain the first system operation and maintenance data of the Authentication-Authorization-Billing (AAA) system within the first time period. The first system operation and maintenance data includes: operation and maintenance operation logs, operation and maintenance configuration change information, performance monitoring data, and business indicator data.

[0034] Step S104: Analyze the first system operation and maintenance data using a pre-trained scene recognition model to obtain the target operation and maintenance scene category of the AAA system within the first time period.

[0035] Step S106: Determine the target operation and maintenance cases that match the target operation and maintenance scenario category in the preset strategy knowledge base, and determine the execution effect of the atomic capability orchestration strategy in each target operation and maintenance case. The strategy knowledge base includes triple information of multiple operation and maintenance cases, and the triple information includes: operation and maintenance scenario category, atomic capability orchestration strategy and execution effect. The atomic capability orchestration strategy is used to characterize the execution order between multiple atomic capabilities and the parameter information of each atomic capability.

[0036] Step S108: Convert the target atomic capability orchestration strategy with the best execution effect into the corresponding target operation and maintenance configuration file, and distribute the target operation and maintenance configuration file to the AAA system.

[0037] Based on the scheme defined in steps S102 to S108 above, it can be understood that in this embodiment of the application, by responding to the operation and maintenance instructions initiated by the target object, the first system operation and maintenance data of the AAA system is automatically analyzed, and the target operation and maintenance scenario category of the AAA system is determined by the scenario recognition model. Then, the best atomic capability orchestration strategy is selected from the strategy knowledge base, converted into a target operation and maintenance configuration file, and distributed to the AAA system for security management. This achieves the operation and maintenance technology effects of intelligent identification and decision-making, automated orchestration and execution, and achieves the purpose of improving the automation level of AAA system operation and maintenance, reducing human error, and improving system stability and operation and maintenance efficiency.

[0038] The following section explains each step of the AAA system automated operation and maintenance method in conjunction with the specific implementation process.

[0039] In the technical solution provided in step S102 above, when an operations and maintenance personnel initiate an operations and maintenance task or the system automatically detects an abnormal event occurring in the AAA system, the automated operations and maintenance system can monitor and collect the first system operations and maintenance data of the AAA system within the first time period.

[0040] The first system's operation and maintenance data includes, but is not limited to: operation and maintenance logs, operation and maintenance configuration change information, performance monitoring data, business indicator data, fault reports, and emergency response logs. Operation and maintenance logs record all operations performed by relevant operation and maintenance personnel and the operation and maintenance tasks executed by the system; operation and maintenance configuration change information records detailed information on all configuration changes to the AAA system within the first time period, including change time, change content, and configuration status before and after the change; performance monitoring data records performance indicators of the AAA system within the first time period, such as CPU utilization, memory usage, authentication request rate, and network latency; business indicator data records data directly related to business within the AAA system within the first time period, such as user authentication success rate, number of online users, and average online time; fault reports and emergency response logs record fault events, security incidents, and emergency response actions of the AAA system within the first time period.

[0041] In the technical solution provided in step S104 above, the automated operation and maintenance system uses a pre-trained scene recognition model to analyze the operation and maintenance data of the first system to obtain the target operation and maintenance scenario category of the AAA system in the first time period. The target operation and maintenance scenario category can be one of the following: data verification, service restart, authentication control, interface emergency processing, database restart, primary and backup node switching, daily maintenance and upgrade (such as system maintenance, software upgrade, hardware replacement), and capacity expansion during peak business periods.

[0042] Specifically, the scene recognition model described above can be trained using the following method:

[0043] Step 1: Build the initial learning model.

[0044] Specifically, the initial learning model architecture is selected first. This architecture can be a hybrid of deep learning and machine learning models. Deep learning models (such as convolutional neural networks and long short-term memory networks) are used to handle complex, non-linear data relationships, while machine learning models (such as random forests and support vector machines) are used to handle structured data and perform fast prediction and classification. Next, the parameters of the initial learning model are set, including the number of layers, neurons, and activation functions of the deep learning model, and the decision tree depth and feature selection strategy of the machine learning model.

[0045] Step 2: Obtain the training sample set and sample label set.

[0046] The training sample set includes: the second system operation and maintenance data of the AAA system in multiple first historical time periods as training samples, and the sample label set includes: the operation and maintenance scenario categories of the AAA system in each first historical time period as the sample labels of the corresponding training samples, including but not limited to data verification, service restart, authentication control, interface emergency processing, database restart, master-slave node switching and other scenarios.

[0047] Step 3: Analyze multiple training samples in the training sample set using the initial learning model to obtain the prediction result for each training sample, and generate multiple adversarial samples based on the sample label and prediction result of each training sample in the training sample set to update the training sample set and sample label set.

[0048] In the technical solution provided in the third step above, the method may include: constructing a target loss function, such as a cross-entropy loss function, based on the sample label and prediction result of each training sample in the training sample set; for each training sample in the training sample set, calculating the gradient of the target loss function relative to the training sample through the backpropagation algorithm, and adding perturbation to the training sample along the direction of the gradient to obtain adversarial samples, and using the sample label of the training sample as the sample label of each adversarial sample; updating the training sample set and the sample label set based on the adversarial samples and sample labels of multiple training samples.

[0049] Since the gradient indicates the direction of model parameter adjustment to reduce the value of the target loss function, adding a small perturbation along the gradient direction to the training samples can yield a new training sample, i.e., an adversarial sample. The magnitude of the perturbation needs to be precisely set to ensure that the adversarial sample remains close to the original training sample in the feature space, while simultaneously misleading the model's predictions. Although the adversarial sample created in this way alters some features of the original training sample, its actual operational scenario category remains unchanged; therefore, the original training sample's label is retained as the adversarial sample's label.

[0050] Expanding the training sample set and sample label set using the methods described above can increase the training complexity of the model, enabling it to better handle and distinguish these edge cases or difficult cases during subsequent training, thereby improving the model's generalization ability and robustness.

[0051] It is important to note that when updating the training sample set and sample label set, if there are too many adversarial samples of a certain category, the model may become biased towards that category. Therefore, the training sample set can be balanced to ensure an appropriate proportion of samples from each category and avoid model training bias.

[0052] Step 4: Use the updated training sample set and updated sample label set to continue iteratively training the initial learning model until the preset iteration termination condition is met, and obtain the trained scene recognition model.

[0053] Through the above-mentioned refined model training steps, an accurate, efficient, and robust operation and maintenance scenario identification model can be constructed. This model can intelligently identify and predict the operation and maintenance scenario categories of the AAA system in different time periods, providing decision support for subsequent personalized configuration orchestration and automated operation and maintenance.

[0054] In the technical solution provided in step S106 above, the automated operation and maintenance system can pre-build and maintain a strategy knowledge base. This strategy knowledge base includes multiple operation and maintenance cases, and each operation and maintenance case is stored as a triplet of information, including: operation and maintenance scenario category, atomic capability orchestration strategy, and execution effect. The operation and maintenance scenario category is used to identify the specific operation and maintenance context to which the operation and maintenance case belongs; the atomic capability orchestration strategy records in detail the execution order and parameter information of each atomic capability in the operation and maintenance case; and the execution effect quantifies the system performance, performance data, fault resolution time, resource consumption, and other indicators after the atomic capability orchestration strategy is executed.

[0055] Specifically, the aforementioned strategy knowledge base can be constructed using the following method:

[0056] Step 1: Obtain the secondary system status data, secondary system operation and maintenance data, and operation and maintenance scenario categories of the AAA system within multiple first historical time periods. The secondary system status data includes, but is not limited to: performance metrics, resource utilization, and service health status; the secondary system operation and maintenance data includes, but is not limited to: operation and maintenance logs, operation and maintenance configuration change information, performance monitoring data, and business metric data.

[0057] Step 2: For each first historical time period, the pre-trained strategy decision model is used to analyze the second system status data and the second system operation and maintenance data to obtain atomic capability orchestration strategies that have an execution effect higher than the first preset threshold and meet the preset verification mechanism for the operation and maintenance scenario category within the first historical time period. The execution effect is determined by at least the execution time, resource consumption and success probability. The strategy decision model is trained based on the deep Q-network algorithm.

[0058] Step 3: Construct a strategy knowledge base by consisting of multiple operation and maintenance cases, which are composed of the operation and maintenance scenario categories, the atomic capability orchestration strategies executed, and the corresponding execution effects of the atomic capability orchestration strategies in each first historical time period of the AAA system.

[0059] Specifically, in the technical solution provided in the second step above, the method may include:

[0060] Construct a Q-network and a target network for solving the atomic capability orchestration strategy, and initialize the network weight parameters of the Q-network and the target network;

[0061] Set up the experience pool, determine its capacity, and initialize the priority of each sample in the experience pool.

[0062] After determining the first preset number of iteration periods, the network weight parameters are iteratively solved through the following steps:

[0063] Within each iteration cycle, the vehicle network environment state is initialized, and a second preset number of calculation cycles are determined. The system state includes at least the second system state data and the second system operation and maintenance data corresponding to the vehicle tasks of the AAA system.

[0064] Within each computation cycle, the current system state is input into the Q-network, the Q-values ​​corresponding to different atomic capability orchestration strategies are calculated, and the target atomic capability orchestration strategy with the largest Q-value is selected based on a greedy strategy. The target atomic capability orchestration strategy is executed and the reward is calculated. A new system state is obtained, and the current system state, the target atomic capability orchestration strategy, the reward, and the new system state are treated as a sample and stored in the experience pool. Based on the priority of each sample in the experience pool, multiple samples are sampled and input into a neural network containing a bidirectional long short-term memory network. The probability of each sample being sampled, the mean squared error loss function, and the loss function weight are calculated. Based on the calculation results, the target Q-value is determined and the network weight parameters of the Q-network are updated. The temporal difference error of all samples in the experience pool is calculated and the priority of all samples is updated.

[0065] After a third preset number of calculation cycles, the network weight parameters of the target network are updated based on the network weight parameters of the Q network, wherein the third preset number is less than the second preset number;

[0066] After the iteration is completed, the resulting target network is used as the policy decision model.

[0067] The aforementioned experience pool is a SumTree structure, and the samples in the experience pool are the leaf nodes in the SumTree. Therefore, samples can be replaced by new samples to replace the earliest samples stored in the experience pool according to the first-in-first-out principle.

[0068] Furthermore, regarding the multi-atom capability arrangement strategy, it can be based on... Figure 2 The process shown involves the personalized orchestration and verification of multiple atomic capabilities within a pre-defined atomic capability library. This library includes various compliant atomic capabilities of the AAA system under different operational scenarios, such as database start / stop scripts, port addition scripts, switch start / stop scripts, interface data retransmission scripts, load balancing switching scripts, patch upload scripts, pass-through mode scripts, authentication record scripts, process start / stop scripts, network connectivity scripts, server start / stop scripts, log collection scripts, user table collection scripts, host status collection scripts, dual-machine freeze scripts, online data collection scripts, and authentication record scripts. To ensure the feasibility of atomic capabilities, after personalized orchestration, the resulting orchestration strategy can be automatically verified using methods such as code verification, business verification, data backup, difference display, data comparison, intelligent rollback, process protection, and failure retransmission.

[0069] (1) Code verification includes:

[0070] Static code analysis involves using code analysis tools to check code style and conventions, ensuring that the code meets operational requirements and usage standards.

[0071] Automatic formatting means that the integrated code formatting tool automatically adjusts the code format to improve readability.

[0072] Test coverage refers to writing unit test cases to ensure that critical functional logic is tested.

[0073] Automated execution means integrating code verification into the CI (Continuous Integration) / CD (Continuous Delivery / Continuous Deployment) process to ensure that tests are automatically executed after each commit.

[0074] Rapid feedback means setting up CI / CD tools to provide timely feedback on test results, ensuring that if a run fails, feedback can be provided and the issue fixed immediately.

[0075] (2) Business verification includes:

[0076] Unit testing involves writing unit tests for each verification logic to ensure that it works correctly independently.

[0077] Integration testing involves conducting integration tests within a complete business process environment to ensure the correctness of the collaboration between various business flows.

[0078] Business penetration testing means that the entire business process needs to be tested and verified.

[0079] (3) Data backup includes:

[0080] Confirm the backup content, that is, ensure that the relevant data and configuration files are backed up before each operation, and confirm the backup location.

[0081] Incremental backup records are records of newly added or modified data blocks and their hash values ​​after each incremental backup.

[0082] Incremental verification logic involves periodically checking incremental backup records to ensure that all incremental data passes verification, preventing omissions or corruption.

[0083] (4) The differences are shown in the following ways:

[0084] Incremental difference detection involves recording the data blocks added or modified during each backup and their hash values, and only checking the differences in these incremental parts.

[0085] Detailed logs record the results of each difference check, including the specific location, content, and severity of the difference.

[0086] Classification labels are used to categorize differences into different categories (such as additions, deletions, modifications, damage, etc.) to facilitate subsequent processing and analysis.

[0087] A graphical interface refers to the development or integration of graphical interface tools that intuitively display differences in results and support user interaction for querying and filtering.

[0088] (5) Data comparison includes:

[0089] Byte-level comparison means that for small files or critical data, byte-level comparison can be performed directly to ensure that every byte is completely consistent.

[0090] Metadata verification involves checking the file's metadata (such as creation time, modification time, permissions, etc.) to ensure the consistency of the metadata.

[0091] A graphical interface refers to the development or integration of graphical interface tools that intuitively display data comparison results and support user interaction for querying and filtering.

[0092] (6) Intelligent rollback includes:

[0093] Version control is the use of a version control system to manage code and configuration files, ensuring that every change is traceable.

[0094] A rollback script is a script that is written or optimized to call the selected rollback strategy and execute the rollback operation.

[0095] Scheduling scripts are scripts that are written and used in conjunction with task scheduling tools to periodically check the system's health status and trigger rollbacks when necessary.

[0096] Custom health checks involve defining application-specific health check rules, such as response time and service availability, to promptly identify potential problems.

[0097] The rules engine determines whether a rollback operation needs to be triggered based on preset rules. For example, if the verification process finds that the program has not started, the service is not running normally, or the business test fails, a rollback operation will be performed.

[0098] Hash verification is the process of using a hash algorithm to verify the consistency of data before and after the rollback.

[0099] (7) Process protection includes:

[0100] A daemon script is a script that is written or optimized to call selected monitoring and recovery methods and perform daemon operations.

[0101] Scheduling scripts are scripts that are written and used in conjunction with task scheduling tools to periodically check process status and trigger guardians when necessary.

[0102] Real-time monitoring means using a monitoring tool like Zabbix to continuously monitor the status of critical processes.

[0103] Custom health checks involve defining application-specific health check rules, such as API response time and service availability, to promptly identify potential problems.

[0104] (8) Retransmission upon failure includes:

[0105] Maximum number of retries: This setting determines the maximum number of retries to prevent infinite retrying.

[0106] The resend deadline is the set deadline for the resend operation. If the deadline is exceeded, resending will stop and a failure will be reported.

[0107] Resend priority is set according to the importance of the operation, so as to ensure that critical operations are resent first.

[0108] In step S106, the policy decision model can be loaded into memory. For example, the raw data of the policy decision model can be loaded from non-volatile memory into volatile memory so that the processor can run the policy decision model. The raw data of the policy decision model refers to unprocessed data, which typically includes the parameters and structural data of the policy decision model. The structural data can be the computational relationships based on the parameters, such as the forward propagation computational relationships between intermediate layers and between neurons. Specifically, the structural data can include the structure-related code of the policy decision model, such as code used to perform related calculations between intermediate layers and between neurons.

[0109] In one implementation, a region can be partitioned in memory for loading the policy decision model, which may include a structure data storage area and a parameter storage area. The structure data storage area stores structure-related code, and the parameters referenced by it can be accessed via pointers pointing to the addresses of specific parameters in the parameter storage area. During the training of the policy decision model, frequent parameter updates may be required; in this case, updating the parameter values ​​in the parameter storage area is sufficient.

[0110] Through the above process, a strategy knowledge base that can support future operation and maintenance decisions can be built, significantly improving the automation capabilities and efficiency of the AAA system's operation and maintenance.

[0111] Furthermore, in the technical solution provided in step S106 above, the automated operation and maintenance system can retrieve multiple target operation and maintenance cases that match the target operation and maintenance scenario category from the strategy knowledge base, and determine the execution effect of the atomic capability orchestration strategy in each target operation and maintenance case.

[0112] Specifically, the automated operation and maintenance system can calculate the similarity between the operation and maintenance scenario categories in each operation and maintenance case in the strategy knowledge base and the target operation and maintenance scenario category; and select multiple operation and maintenance cases with similarity higher than a second preset threshold as multiple target operation and maintenance cases that match the target operation and maintenance scenario category.

[0113] In the technical solution provided in step S108 above, the automated operation and maintenance system can convert the target atomic capability orchestration strategy with the best execution effect into the corresponding target operation and maintenance operation configuration file, and distribute the target operation and maintenance operation configuration file to the AAA system.

[0114] Specifically, the automated operations and maintenance system can convert the target atomic capability orchestration strategy with the best execution effect into the corresponding target operations and maintenance configuration file in the following way:

[0115] Step 1: Determine multiple target atomic capabilities from the pre-defined atomic capability library that correspond to the target atomic capability orchestration strategy.

[0116] Step 2: Configure parameters and logically arrange multiple target atomic capabilities according to the target atomic capability orchestration strategy to obtain the target operation and maintenance configuration file.

[0117] The aforementioned atomic capability library includes multiple compliant atomic capabilities of the AAA system under different operation and maintenance scenarios, such as database start / stop scripts, port addition scripts, switch start / stop scripts, interface data retransmission scripts, load balancer switching scripts, patch upload scripts, pass-through mode scripts, authentication record scripts, process start / stop scripts, network connectivity scripts, server start / stop scripts, log collection scripts, user table collection scripts, host status collection scripts, dual-machine freeze scripts, online data collection scripts, authentication record scripts, etc.

[0118] Furthermore, the automated operations and maintenance system can use a secure transmission mechanism to distribute the target operations and maintenance configuration file to multiple system nodes (such as the Radius server) in the AAA system. Each node, upon receiving the configuration file, first verifies its format and content to ensure its integrity. After successful verification, the target operations and maintenance configuration file is stored in a designated directory, ready for subsequent configuration activation.

[0119] Furthermore, the automated operation and maintenance system can also obtain the execution results of the target operation and maintenance operation configuration file executed by multiple system nodes of the AAA system. The execution results include: execution time (including start event and end event), resource consumption (such as CPU and memory usage), abnormal information during the execution process, and whether the configuration change was successfully completed. Based on the execution results of the target operation and maintenance operation configuration file executed by each system node, the execution effect of the target atomic capability orchestration strategy corresponding to the target operation and maintenance operation configuration file is determined. The execution effect can be divided into high efficiency, medium efficiency, and low efficiency, and the evaluation of the execution effect can be determined by the execution results. The operation and maintenance scenario category, the target atomic capability orchestration strategy, and the corresponding execution effect are used as the triple information of the new operation and maintenance case, and the triple information of the new operation and maintenance case is stored in the strategy knowledge base.

[0120] The above methods can be used to continuously update the strategy knowledge base, enabling the atomic capability orchestration strategies in the strategy knowledge base to continuously evolve and optimize themselves to adapt to the ever-changing operation and maintenance environment and technical requirements.

[0121] The aforementioned automated operation and maintenance method for AAA systems has the following technical advantages compared to existing operation and maintenance configuration methods:

[0122] (1) The scene recognition model is used to automatically analyze the first system operation and maintenance data of the AAA system in the first time period, accurately identify the current operation and maintenance scenario category. Compared with traditional manual judgment, it significantly reduces the participation of human in operation and maintenance, reduces errors caused by manual operation, and greatly improves the recognition speed and accuracy.

[0123] (2) Based on the identified scenario category, the most suitable operation and maintenance case in the strategy knowledge base is automatically matched to provide the optimal target atomic capability orchestration strategy for the current scenario. This makes the operation and maintenance operation more targeted and efficient, and avoids the waste of resources and operational errors that may be caused by general configuration.

[0124] (3) The target atomic capability orchestration strategy is converted into a target operation and maintenance configuration file, and then distributed to each node of the AAA system securely and quickly through an automated mechanism. Compared with manual configuration, it can take effect on all relevant nodes faster and more accurately, shortening the time interval from decision to execution, improving business response speed, and greatly reducing operational complexity and chance of errors, thus improving configuration efficiency.

[0125] (4) By collecting and evaluating the execution results of the target operation and maintenance configuration files, the strategy knowledge base can be continuously optimized, becoming a powerful support for operation and maintenance decisions.

[0126] In summary, the embodiments of this application, by introducing intelligent scene recognition, decision recommendation, and automated execution, have achieved comprehensive optimization of the operation and maintenance of the AAA system, which not only improves the efficiency of operation and maintenance and the rationality of resource allocation, but also enhances the stability and security of the system.

[0127] Example 2

[0128] According to embodiments of this application, an automated operation and maintenance system for an AAA system, used to implement the automated operation and maintenance method for an AAA system in Embodiment 1, is also provided, such as... Figure 3 As shown, the AAA system automated operation and maintenance system includes at least: an acquisition module 32, an identification module 34, a determination module 36, and a configuration module 38, wherein:

[0129] The acquisition module 32 is used to respond to the operation and maintenance instructions initiated by the target object and acquire the first system operation and maintenance data of the Authentication-Authorization-Billing AAA system within the first time period. The first system operation and maintenance data includes: operation and maintenance operation logs, operation and maintenance configuration change information, performance monitoring data, and business indicator data.

[0130] The identification module 34 is used to analyze the first system operation and maintenance data using a pre-trained scene identification model to obtain the target operation and maintenance scene category of the AAA system in the first time period.

[0131] The determination module 36 is used to determine multiple target operation and maintenance cases that match the target operation and maintenance scenario category within the preset strategy knowledge base, and to determine the execution effect of the atomic capability orchestration strategy in each target operation and maintenance case. The strategy knowledge base includes triple information of multiple operation and maintenance cases, and the triple information includes: operation and maintenance scenario category, atomic capability orchestration strategy and execution effect. The atomic capability orchestration strategy is used to characterize the execution order between multiple atomic capabilities and the parameter information of each atomic capability.

[0132] Configuration module 38 is used to convert the target atomic capability orchestration strategy with the best execution effect into the corresponding target operation and maintenance configuration file, and to distribute the target operation and maintenance configuration file to the AAA system.

[0133] The following section explains the functions of each module of the AAA system automated operation and maintenance device in conjunction with the specific implementation process.

[0134] Optionally, the scene recognition model described above, including the recognition module 34, can be trained using the following method:

[0135] Step 1: Build the initial learning model.

[0136] Step 2: Obtain the training sample set and the sample label set. The training sample set includes: the second system operation and maintenance data of the AAA system in multiple first historical time periods as training samples. The sample label set includes: the operation and maintenance scenario categories of the AAA system in each first historical time period as the sample labels of the corresponding training samples, including but not limited to scenarios such as data verification, service restart, authentication control, interface emergency handling, database restart, and primary / backup node switching.

[0137] Step 3: Analyze multiple training samples in the training sample set using the initial learning model to obtain the prediction result for each training sample, and generate multiple adversarial samples based on the sample label and prediction result of each training sample in the training sample set to update the training sample set and sample label set.

[0138] In the technical solution provided in the third step above, the method may include: constructing a target loss function, such as a cross-entropy loss function, based on the sample label and prediction result of each training sample in the training sample set; for each training sample in the training sample set, calculating the gradient of the target loss function relative to the training sample through the backpropagation algorithm, and adding perturbation to the training sample along the direction of the gradient to obtain adversarial samples, and using the sample label of the training sample as the sample label of each adversarial sample; updating the training sample set and the sample label set based on the adversarial samples and sample labels of multiple training samples.

[0139] Step 4: Use the updated training sample set and updated sample label set to continue iteratively training the initial learning model until the preset iteration termination condition is met, and obtain the trained scene recognition model.

[0140] Alternatively, module 36 can construct the strategy knowledge base as follows:

[0141] Step 1: Obtain the secondary system status data, secondary system operation and maintenance data, and operation and maintenance scenario categories of the AAA system within multiple first historical time periods. The secondary system status data includes, but is not limited to: performance metrics, resource utilization, and service health status; the secondary system operation and maintenance data includes, but is not limited to: operation and maintenance logs, operation and maintenance configuration change information, performance monitoring data, and business metric data.

[0142] Step 2: For each first historical time period, the pre-trained strategy decision model is used to analyze the second system status data and the second system operation and maintenance data to obtain atomic capability orchestration strategies that have an execution effect higher than the first preset threshold and meet the preset verification mechanism for the operation and maintenance scenario category within the first historical time period. The execution effect is determined by at least the execution time, resource consumption and success probability. The strategy decision model is trained based on the deep Q-network algorithm.

[0143] Step 3: Construct a strategy knowledge base by consisting of multiple operation and maintenance cases, which are composed of the operation and maintenance scenario categories, the atomic capability orchestration strategies executed, and the corresponding execution effects of the atomic capability orchestration strategies in each first historical time period of the AAA system.

[0144] Specifically, in the technical solution provided in the second step above, the method may include:

[0145] Construct a Q-network and a target network for solving the atomic capability orchestration strategy, and initialize the network weight parameters of the Q-network and the target network;

[0146] Set up the experience pool, determine its capacity, and initialize the priority of each sample in the experience pool.

[0147] After determining the first preset number of iteration periods, the network weight parameters are iteratively solved through the following steps:

[0148] Within each iteration cycle, the vehicle network environment state is initialized, and a second preset number of calculation cycles are determined. The system state includes at least the second system state data and the second system operation and maintenance data corresponding to the vehicle tasks of the AAA system.

[0149] Within each computation cycle, the current system state is input into the Q-network, the Q-values ​​corresponding to different atomic capability orchestration strategies are calculated, and the target atomic capability orchestration strategy with the largest Q-value is selected based on a greedy strategy. The target atomic capability orchestration strategy is executed and the reward is calculated. A new system state is obtained, and the current system state, the target atomic capability orchestration strategy, the reward, and the new system state are treated as a sample and stored in the experience pool. Based on the priority of each sample in the experience pool, multiple samples are sampled and input into a neural network containing a bidirectional long short-term memory network. The probability of each sample being sampled, the mean squared error loss function, and the loss function weight are calculated. Based on the calculation results, the target Q-value is determined and the network weight parameters of the Q-network are updated. The temporal difference error of all samples in the experience pool is calculated and the priority of all samples is updated.

[0150] After a third preset number of calculation cycles, the network weight parameters of the target network are updated based on the network weight parameters of the Q network, wherein the third preset number is less than the second preset number;

[0151] After the iteration is completed, the resulting target network is used as the policy decision model.

[0152] The aforementioned experience pool is a SumTree structure, and the samples in the experience pool are the leaf nodes in the SumTree. Therefore, samples can be replaced by new samples to replace the earliest samples stored in the experience pool according to the first-in-first-out principle.

[0153] Furthermore, the determining module 36 determines multiple target operation and maintenance cases that match the target operation and maintenance scenario category according to the following method: calculating the similarity between the operation and maintenance scenario category in each operation and maintenance case in the strategy knowledge base and the target operation and maintenance scenario category; and taking multiple operation and maintenance cases with similarity higher than the second preset threshold as multiple target operation and maintenance cases that match the target operation and maintenance scenario category.

[0154] Optionally, configuration module 38 can convert the target atomic capability orchestration strategy with the best execution effect into the corresponding target operation and maintenance configuration file in the following manner:

[0155] Step 1: Determine multiple target atomic capabilities from the pre-defined atomic capability library that correspond to the target atomic capability orchestration strategy.

[0156] Step 2: Configure parameters and logically arrange multiple target atomic capabilities according to the target atomic capability orchestration strategy to obtain the target operation and maintenance configuration file.

[0157] The aforementioned atomic capability library includes multiple compliant atomic capabilities of the AAA system under different operation and maintenance scenarios, such as database start / stop scripts, port addition scripts, switch start / stop scripts, interface data retransmission scripts, load balancer switching scripts, patch upload scripts, pass-through mode scripts, authentication record scripts, process start / stop scripts, network connectivity scripts, server start / stop scripts, log collection scripts, user table collection scripts, host status collection scripts, dual-machine freeze scripts, online data collection scripts, authentication record scripts, etc.

[0158] Furthermore, the configuration module 38 can distribute the target operation and maintenance configuration file to multiple system nodes (such as the Radius server) in the AAA system through a secure transmission mechanism. After receiving the operation and maintenance configuration file, each node first verifies its format and content to ensure its integrity. After successful verification, the target operation and maintenance configuration file is stored in a specified directory, ready for subsequent configuration activation.

[0159] Furthermore, the configuration module 38 can also obtain the execution results of the target operation and maintenance configuration file executed by multiple system nodes of the AAA system. The execution results include: execution time (including start event and end event), resource consumption (such as CPU and memory usage), exception information during execution, and whether the configuration change was successfully completed. Based on the execution results of the target operation and maintenance configuration file executed by each system node, the execution effect of the target atomic capability orchestration strategy corresponding to the target operation and maintenance configuration file is determined. The execution effect can be divided into high efficiency, medium efficiency, and low efficiency, and the evaluation of the execution effect can be determined by the execution results. The operation and maintenance scenario category, the target atomic capability orchestration strategy, and the corresponding execution effect are used as the triple information of the new operation and maintenance case, and the triple information of the new operation and maintenance case is stored in the strategy knowledge base.

[0160] It should be noted that each module in the AAA system automated operation and maintenance system in this application embodiment corresponds one-to-one with each implementation step of the AAA system automated operation and maintenance method in embodiment 1. Since embodiment 1 has been described in detail, some details not shown in this embodiment can be referred to embodiment 1, and will not be elaborated further here.

[0161] Example 3

[0162] According to an embodiment of this application, a computer program product is also provided, which includes a computer program, wherein when the computer program is executed by a processor, it implements the AAA system automated operation and maintenance method in Embodiment 1.

[0163] According to an embodiment of this application, a non-volatile storage medium is also provided, which includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the AAA system automated operation and maintenance method in Embodiment 1 by running the computer program.

[0164] According to an embodiment of this application, a processor is also provided for running a computer program, wherein the computer program executes the AAA system automated operation and maintenance method in Embodiment 1 during runtime.

[0165] According to an embodiment of this application, an electronic device is also provided, comprising: a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the AAA system automated operation and maintenance method of Embodiment 1 through the computer program.

[0166] Specifically, the computer program executes the following steps during runtime: In response to the operation and maintenance instructions initiated by the target object, it acquires the first system operation and maintenance data of the AAA system within a first time period. This first system operation and maintenance data includes: operation and maintenance logs, operation and maintenance configuration change information, performance monitoring data, and business indicator data. It then analyzes the first system operation and maintenance data using a pre-trained scenario recognition model to obtain the target operation and maintenance scenario category of the AAA system within the first time period. Next, it identifies multiple target operation and maintenance cases within a preset strategy knowledge base that match the target operation and maintenance scenario category, and determines the execution effect of the atomic capability orchestration strategy in each target operation and maintenance case. The strategy knowledge base includes triple information for multiple operation and maintenance cases, and the triple information includes: operation and maintenance scenario category, atomic capability orchestration strategy, and execution effect. The atomic capability orchestration strategy is used to characterize the execution order between multiple atomic capabilities and the parameter information of each atomic capability. Finally, it converts the target atomic capability orchestration strategy with the best execution effect into a corresponding target operation and maintenance configuration file and distributes the target operation and maintenance configuration file to the AAA system.

[0167] As an alternative implementation, the above-mentioned electronic device may exist in the form of a mobile terminal, a computer terminal, or a similar computing device. Figure 4 A hardware block diagram of a computer terminal for implementing an automated operation and maintenance method for an AAA system is shown. Figure 4 As shown, the computer terminal 40 may include one or more processors 402 (shown as 402a, 402b, ..., 402n in the figure) 402 (processor 402 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 404 for storing data, and a transmission device 406 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 4 The structure shown is for illustrative purposes only and does not limit the structure of the computer terminal described above. For example, computer terminal 40 may also include... Figure 4 The more or fewer components shown, or having the same Figure 4 The different configurations shown.

[0168] It should be noted that the aforementioned one or more processors 402 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 40. As involved in the embodiments of this application, the data processing circuits serve as processor control (e.g., selection of a variable resistor termination path connected to an interface).

[0169] The memory 404 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the AAA system automated operation and maintenance method in this embodiment. The processor 402 executes various functional applications and data processing by running the software programs and modules stored in the memory 404, thereby implementing the above-mentioned application vulnerability detection method. The memory 404 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 404 may further include memory remotely located relative to the processor 402, and these remote memories can be connected to the computer terminal 40 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0170] The transmission device 406 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 40. In one example, the transmission device 406 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 406 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0171] The display can be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 40.

[0172] The sequence numbers of the above embodiments are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0173] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0174] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0175] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0176] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0177] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0178] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for automatic operation and maintenance of an AAA system, characterized in that, The method comprises the steps of: in response to an operation and maintenance instruction initiated by a target object, obtaining first system operation and maintenance data of an authentication, authorization and accounting (AAA) system in a first time period, wherein the first system operation and maintenance data comprises operation and maintenance operation logs, operation and maintenance configuration change information, performance monitoring data and service index data; analyzing the first system operation and maintenance data by using a pre-trained scene recognition model to obtain a target operation and maintenance scene category of the AAA system in the first time period; determining a plurality of target operation and maintenance cases in a preset policy knowledge base that match the target operation and maintenance scene category, and determining the execution effect of an atomic capability arrangement strategy in each target operation and maintenance case, wherein the policy knowledge base comprises triple information of a plurality of operation and maintenance cases, and the triple information comprises an operation and maintenance scene category, an atomic capability arrangement strategy and an execution effect, and the atomic capability arrangement strategy is used to represent the execution order between a plurality of atomic capabilities and the parameter information of each atomic capability; converting the target atomic capability arrangement strategy with the best execution effect into a corresponding target operation and maintenance operation configuration file, and issuing the target operation and maintenance operation configuration file to the AAA system.

2. The method of claim 1, wherein, The training process of the scene recognition model comprises: constructing an initial learning model; obtaining a training sample set and a sample label set, wherein the training sample set comprises second system operation and maintenance data of the AAA system in a plurality of first historical time periods as training samples, and the sample label set comprises an operation and maintenance scene category of the AAA system in each first historical time period as a sample label of a corresponding training sample, and the operation and maintenance scene category comprises at least one of data verification, service restart, authentication control, interface emergency handling, database restart and master-slave node switching; analyzing a plurality of training samples in the training sample set by using the initial learning model to obtain a prediction result of each training sample, and generating a plurality of adversarial samples according to the sample label and the prediction result of each training sample in the training sample set to update the training sample set and the sample label set; continuing to iteratively train the initial learning model by using the updated training sample set and the updated sample label set until a preset iteration termination condition is met, and obtaining a trained scene recognition model.

3. The method of claim 2, wherein, Generating a plurality of adversarial samples according to the sample label and the prediction result of each training sample in the training sample set to update the training sample set and the sample label set comprises: constructing a target loss function according to the sample label and the prediction result of each training sample in the training sample set; for each training sample in the training sample set, calculating the gradient of the target loss function with respect to the training sample, and adding a disturbance to the training sample in the direction of the gradient to obtain an adversarial sample, and taking the sample label of the training sample as the sample label of the adversarial sample; updating the training sample set and the sample label set according to the adversarial samples and the sample labels of the plurality of training samples.

4. The method of claim 1, wherein, The construction process of the policy knowledge base comprises: obtaining second system state data, second system operation and maintenance data and operation and maintenance scene categories of the AAA system in a plurality of first historical time periods; for each of the first historical time periods, analyzing the second system state data and the second system operation and maintenance data using a pre-trained policy decision model to obtain an atomic capability arrangement strategy for the operation and maintenance scene category in the first historical time period, the execution effect of which is higher than a first preset threshold and meets a preset verification mechanism, wherein the execution effect is determined at least by execution time, resource consumption and success probability, and wherein the policy decision model is obtained based on a deep Q network algorithm training; a plurality of operation and maintenance cases composed of the operation and maintenance scene category in each of the first historical time periods, the executed atomic capability arrangement strategy and the execution effect corresponding to the atomic capability arrangement strategy, are used to construct the policy knowledge base.

5. The method of claim 1, wherein, determining a plurality of target operation and maintenance cases in the preset policy knowledge base that match the target operation and maintenance scene category, including: calculating the similarity of the operation and maintenance scene category in each operation and maintenance case in the policy knowledge base to the target operation and maintenance scene category; the plurality of operation and maintenance cases with a similarity higher than a second preset threshold are taken as the plurality of target operation and maintenance cases that match the target operation and maintenance scene category.

6. The method of claim 1, wherein, converting the target atomic capability arrangement strategy with the best execution effect into a corresponding target operation and maintenance operation configuration file, including: determining a plurality of target atomic capabilities corresponding to the target atomic capability arrangement strategy from a preset atomic capability library, wherein the atomic capability library includes a plurality of compliant atomic capabilities of the AAA system under different operation and maintenance scenes; performing parameter setting and logical arrangement on the plurality of target atomic capabilities according to the target atomic capability arrangement strategy to obtain the target operation and maintenance operation configuration file.

7. The method of claim 1, wherein, after the target operation and maintenance operation configuration file is issued to the AAA system, the method further includes: obtaining execution results of the target operation and maintenance operation configuration file executed by a plurality of system nodes of the AAA system, wherein the execution results include execution time, resource consumption, and success or not; determining the execution effect of the target atomic capability arrangement strategy corresponding to the target operation and maintenance operation configuration file according to the execution results of the target operation and maintenance operation configuration file executed by each of the system nodes; using the operation and maintenance scene category, the target atomic capability arrangement strategy and the corresponding execution effect as the triple information of a new operation and maintenance case, and storing the triple information of the new operation and maintenance case into the policy knowledge base.

8. A AAA system automatic operation and maintenance system, characterized in that, includes: an obtaining module configured to obtain first system operation and maintenance data of an authentication-authorization-charging (AAA) system in a first time period in response to an operation and maintenance instruction initiated by a target object, wherein the first system operation and maintenance data includes operation and maintenance operation logs, operation and maintenance configuration change information, performance monitoring data and service index data; an identification module configured to analyze the first system operation and maintenance data using a pre-trained scene identification model to obtain a target operation and maintenance scene category of the AAA system in the first time period; A determining module is configured to determine a plurality of target operation and maintenance cases in a preset policy knowledge base that match the target operation and maintenance scenario category, and determine an execution effect of an atomic capability arrangement strategy in each target operation and maintenance case, wherein the policy knowledge base includes triple information of a plurality of operation and maintenance cases, and the triple information includes an operation and maintenance scenario category, an atomic capability arrangement strategy, and an execution effect, and the atomic capability arrangement strategy is used to represent an execution sequence between a plurality of atomic capabilities and parameter information of each atomic capability. A configuring module is configured to convert the target atomic capability arrangement strategy with the best execution effect into a corresponding target operation and maintenance operation configuration file, and issue the target operation and maintenance operation configuration file to the AAA system.

9. A computer program product, characterised in that, The computer program is executed by a processor to implement the AAA system automatic operation and maintenance method in any one of claims 1 to 7. The memory stores a computer program, and the processor is configured to execute the AAA system automatic operation and maintenance method in any one of claims 1 to 7 through the computer program.

10. An electronic device, comprising: The memory stores a computer program, and the processor is configured to execute the AAA system automatic operation and maintenance method in any one of claims 1 to 7 through the computer program. ​

Images