Network policy adjustment method and device, electronic equipment and storage medium

Abstract

The invention provides a network policy adjustment method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring equipment condition data and user behavior data of network equipment; calling a preset user behavior model to process the equipment condition data and the user behavior data, and adjusting a network strategy according to a processing result; and issuing the adjusted network strategy to the network equipment. According to the scheme, the equipment condition data and the user behavior data of the network equipment are processed through the user behavior model, and the network strategy of the network equipment is adjusted according to the processing result. By adjusting the network strategy of the network equipment, on one hand, the stable operation of the network equipment can be ensured to ensure the normal operation of network services, and on the other hand, the dynamically changing network environment and user requirements can be flexibly handled; on the premise that manual intervention is not needed, network optimization can be carried out timely and accurately to guarantee normal operation of network services, so that the processing efficiency is improved, the response delay is reduced, and the misjudgment rate is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method, device, electronic device and storage medium for adjusting network policies. Background Art

[0002] To ensure network performance and user experience, it is necessary to guarantee the normal operation of network services. Currently, the traditional method for guaranteeing network services is to use static predefined rules and manual intervention for network resource management and fault handling.

[0003] However, the traditional method for guaranteeing network services cannot flexibly respond to dynamic network environments and user requirements, and has problems such as low processing efficiency, delayed response, and high misjudgment rate. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method, device, electronic device and storage medium for adjusting network policies to solve the problems of low processing efficiency, delayed response, and high misjudgment rate existing in the traditional method for guaranteeing network services.

[0005] To achieve the above object, embodiments of the present invention provide the following technical solutions:

[0006] A first aspect of an embodiment of the present invention discloses a method for adjusting network policies, the method comprising:

[0007] Obtaining device status data and user behavior data of a network device;

[0008] Invoking a preset user behavior model to process the device status data and the user behavior data, and adjusting the network policy according to the processing result, the user behavior model being obtained by training a specified model based on sample data;

[0009] Issuing the adjusted network policy to the network device.

[0010] Preferably, the device status data is real-time device status data or historical device status data, and the user behavior data is real-time user behavior data or historical user behavior data;

[0011] Obtaining device status data and user behavior data of a network device comprises:

[0012] When triggering a real-time resource adjustment mechanism, obtaining real-time device status data and real-time user behavior data of the network device regularly uploaded by a wireless controller;

[0013] When triggering a prediction mechanism, obtaining historical device status data and historical user behavior data of the network device collected within a historical time period.

[0014] Preferably, the device status data is real-time device status data, and the user behavior data is real-time user behavior data;

[0015] Call a preset user behavior model to process the device status data and the user behavior data, and adjust the network policy according to the processing result, including:

[0016] Input the real-time device status data and the real-time user behavior data into the preset user behavior model for fitting judgment to obtain a model analysis result;

[0017] Based on the policy database and the model analysis result, adjust the network policy.

[0018] Preferably, the device status data is historical device status data, and the user behavior data is historical user behavior data;

[0019] Call a preset user behavior model to process the device status data and the user behavior data, and adjust the network policy according to the processing result, including:

[0020] Input the historical device status data and the historical user behavior data into the preset user behavior model for prediction to obtain a model prediction result;

[0021] Adjust the network policy according to the model prediction result.

[0022] Preferably, adjusting the network policy according to the model prediction result includes:

[0023] Use the decision tree algorithm to convert the model prediction result into an adjusted network policy.

[0024] Preferably, the process of training a specified model based on sample data to obtain a user behavior model includes:

[0025] Load a pre-trained source model;

[0026] Use the sample data and the transfer learning algorithm to adjust or retrain the source model to obtain a user behavior model.

[0027] Preferably, it further includes:

[0028] The user behavior model is updated regularly.

[0029] A second aspect of the embodiments of the present invention discloses a device, and the device includes:

[0030] An acquisition unit, configured to acquire device status data and user behavior data of a network device;

[0031] A processing unit, configured to call a preset user behavior model to process the device status data and the user behavior data, and adjust the network policy according to the processing result, where the user behavior model is obtained by training a specified model based on sample data;

[0032] A distribution unit, configured to distribute the adjusted network policy to the network device.

[0033] A third aspect of an embodiment of the present invention discloses an electronic device, including: a processor, a memory, and a communication interface. Among them, the processor is configured to call and execute a program stored in the memory; the memory is configured to store a program, and the program is used to implement the network policy adjustment method disclosed in the first aspect of the embodiment of the present invention; the communication interface is configured to implement communication with a network device.

[0034] A fourth aspect of an embodiment of the present invention discloses a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, it implements the network policy adjustment method disclosed in the first aspect of the embodiment of the present invention.

[0035] Based on the network policy adjustment method, device, electronic device, and storage medium provided in the above embodiments of the present invention, the method is as follows: obtaining device status data and user behavior data of a network device; calling a preset user behavior model to process the device status data and the user behavior data, and adjusting the network policy according to the processing result; distributing the adjusted network policy to the network device. This solution processes the device status data and user behavior data of the network device through a user behavior model, and adjusts the network policy of the network device according to the processing result. By adjusting the network policy of the network device, on the one hand, it can ensure the stable operation of the network device, thereby ensuring the normal operation of network services. On the other hand, it can flexibly respond to the dynamically changing network environment and user needs, and can perform network optimization in a timely and accurate manner without manual intervention to ensure the normal operation of network services, thereby improving processing efficiency, reducing response latency, and reducing the misjudgment rate. Description of the Drawings

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to the provided drawings.

[0037] Figure 1 It is a flowchart of a network policy adjustment method provided by an embodiment of the present invention;

[0038] Figure 2Flowchart for adjusting network policies provided by an embodiment of the present invention;

[0039] Figure 3 Another flowchart for adjusting network policies provided by an embodiment of the present invention;

[0040] Figure 4 Example diagram of the workflow for adjusting network policies provided by an embodiment of the present invention;

[0041] Figure 5 Block diagram of the structure of a device provided by an embodiment of the present invention. Detailed implementation manners

[0042] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0043] In this application, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, the element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or device including the said element.

[0044] It should be noted here that subsequent to this solution, data collection behaviors such as obtaining user behavior data will be involved. Before collecting user behavior data and other data in this solution, users have been informed in advance and user authorization has been obtained, that is, this solution conducts data collection behaviors on the premise of compliance with relevant regulations.

[0045] As can be seen from the background art, in order to ensure network performance and user experience, it is necessary to ensure the normal operation of network services. Currently, the traditional network service guarantee method is: using static predefined rules and manual intervention for network resource management and fault handling.

[0046] It has been found through research that the traditional network service guarantee method has the following main disadvantages:

[0047] 1. Dependence on predefined rules and manual intervention: The traditional network service guarantee method usually relies on static predefined rules and manual intervention for network resource management and fault handling, and cannot flexibly respond to the dynamically changing network environment and user needs. Moreover, manual intervention not only has low efficiency, but also is prone to misjudgment and delayed response.

[0048] 2. Unable to cope with sudden network failures and performance bottlenecks: Since traditional network service assurance methods lack instant data analysis and decision support, they are unable to effectively predict and handle sudden network failures and performance bottlenecks, resulting in service interruptions and degraded user experiences.

[0049] 3. Lack of in-depth analysis and prediction capabilities for user behavior: Traditional network service assurance methods usually only manage network devices and resources and lack the ability to deeply analyze and predict user behavior. This makes it impossible to accurately understand user needs and behavior patterns, affecting the personalization and optimization of services.

[0050] To address the above drawbacks, this solution proposes a method, device, electronic device, and storage medium for adjusting network policies. By using a user behavior model to process device status data and user behavior data of network devices and adjusting the network policies of network devices according to the processing results, that is, dynamically adjusting network policies by analyzing user behavior data. On the one hand, it can ensure the stable operation of network devices, thereby guaranteeing the normal operation of network services. On the other hand, it can flexibly cope with the dynamically changing network environment and user needs, and can perform network optimization in a timely and accurate manner without manual intervention to ensure the normal operation of network services, thereby improving processing efficiency, reducing response latency, reducing misjudgment rates, improving network performance, and enhancing user experiences.

[0051] See Figure 1 , which shows a flowchart of a method for adjusting network policies provided by an embodiment of the present invention. The method for adjusting network policies includes the following steps:

[0052] Step S101: Obtain device status data and user behavior data of network devices.

[0053] In the specific process of implementing step S101, device status data and user behavior data of network devices are obtained. Among them, network devices specifically refer to each constituent node in the network topology. For example, network devices are terminals, wireless access points (APs), switches, and controllers, etc.

[0054] User behavior data includes, but is not limited to, network data such as terminal portal (portal website) authentication frequency, network access frequency, network usage duration, accessed content, and hot functions that can reflect user behavior characteristics.

[0055] Device status data includes, but is not limited to, information such as wireless access point channel utilization rate, terminal signal strength, AP radio frequency transmission power, switch load, and controller CPU utilization rate that can reflect the actual status of network devices.

[0056] Among them, the network access frequency is used to record the number of accesses by the user within a specific time period, which can help analyze the user activity.

[0057] The network usage duration is used to calculate the stay time of the user in various accesses, and is used to evaluate the attractiveness of different contents.

[0058] The accessed content is used to record in detail information such as the website address accessed by the user, file download, etc., providing rich context data for subsequent analysis.

[0059] The hot functions are used to record the business functions mainly used by the user within a specific time period, such as portal authentication and data auditing, etc.

[0060] In some embodiments, data verification technology is used to perform real-time analysis and verification on the obtained device status data and user behavior data to ensure that the obtained data is valid and accurate.

[0061] In other embodiments, the obtained device status data and user behavior data can be real-time data or historical data. That is, the device status data is real-time device status data or historical device status data, and the user behavior data is real-time user behavior data or historical user behavior data.

[0062] This solution provides a real-time resource adjustment mechanism and a prediction mechanism, which can trigger the real-time resource adjustment mechanism or the prediction mechanism according to the preset conditions.

[0063] Among them, under the real-time resource adjustment mechanism, the real-time device status data and real-time user behavior data are used to adjust the network policy of the network device in real time. Under the prediction mechanism, the historical device status data and historical user behavior data of the network device are used to predict the device requirements of the network device, and then the network policy of the network device is adjusted in advance.

[0064] In some specific embodiments, the specific ways to obtain the device status data and user behavior data of the network device are as follows:

[0065] When triggering the real-time resource adjustment mechanism, obtain the real-time device status data and real-time user behavior data of the network device regularly uploaded by the Wireless Access Controler (WAC).

[0066] When triggering the prediction mechanism, obtain the historical device status data and historical user behavior data of the network device collected within the historical time period.

[0067] It can be understood that when the WAC uploads the real-time device status data and real-time user behavior data, the WAC performs cleaning and formatting processing on the collected real-time device status data and real-time user behavior data.

[0068] Step S102: Invoke a preset user behavior model to process device status data and user behavior data, and adjust the network policy according to the processing result.

[0069] It should be noted that the network policy of the network device is the network resource allocation policy and the service policy. More specifically, the network policy refers to the configuration items of the network device. The user behavior model is obtained by training a specified model based on sample data, and the sample data includes at least sample device status data and sample user behavior data.

[0070] In some embodiments, the method for training the user behavior model is: load a pre-trained source model (or main model), and use the sample data and transfer learning algorithm to adjust or retrain the source model to obtain the user behavior model.

[0071] In practical applications, the user behavior model can also be trained in the following way: select a machine learning algorithm suitable for the scenario (such as a deep learning neural network) to construct the user behavior model, extract features and recognize patterns from the sample data, and then train and evaluate the user behavior model to optimize the algorithm parameters, so as to ensure the accuracy and generalization ability of the trained user behavior model.

[0072] It should be noted that the user behavior model constructed by this solution does not refer to a single specific type of model. Certain types of models can be selected for training according to actual needs to obtain the user behavior model.

[0073] For example: for scenarios of predicting continuous values such as channel utilization rate, user traffic demand, and network latency, the user behavior model can be trained using a multiple linear regression model and a ridge regression model, and the trained user behavior model can predict future traffic demands.

[0074] Another example: for the need to identify which areas (networks composed of single or multiple access points and switches) have traffic peaks, the user behavior model can be trained using a classification model (such as a decision tree model) or a clustering model (such as K-means).

[0075] Another example: for the need to analyze the non-linear relationship of multi-dimensional data, the user behavior model can be trained using a reinforcement learning model or a deep neural network model.

[0076] The above is the related description of training the user behavior model.

[0077] In the process of specifically implementing step S102, invoke the trained user behavior model to process device status data and user behavior data, and adjust the network policy of the network device according to the processing result.

[0078] In some embodiments, the user behavior model is updated regularly. Specifically, the device status data and user behavior data of each network device can be obtained regularly according to a set time interval, and then the user behavior model can be updated regularly according to the regularly obtained device status data and user behavior data.

[0079] Step S103: Send the adjusted network policy to the network device.

[0080] In the process of specifically implementing step S103, after invoking the network policy of the network device, the adjusted network policy is sent to the network device to ensure reasonable resource allocation and maximization of network performance.

[0081] In practical applications, the adjusted network policy can be first sent to the WAC, and then the WAC centrally manages and sends the adjusted network policy to other network devices.

[0082] In the embodiments of the present invention, the device status data and user behavior data of the network device are processed through the user behavior model, and the network policy of the network device is adjusted according to the processing results. On the one hand, it can ensure the stable operation of the network device, thereby ensuring the normal operation of network services. On the other hand, it can flexibly respond to the dynamically changing network environment and user requirements, and can perform network optimization in a timely and accurate manner without manual intervention to ensure the normal operation of network services, thereby improving the processing efficiency, reducing the response delay, and reducing the misjudgment rate.

[0083] It should be noted that according to the obtained device status data and user behavior data, the user behavior model of this solution is invoked to determine whether the network policy needs to be optimized, or the user behavior model is invoked to perform certain prediction tasks.

[0084] For the above embodiments of the present invention Figure 1 Regarding the invocation of the user behavior model to process the device status data and user behavior data involved in step S102, refer to Figure 2 , which shows the flowchart of adjusting the network policy provided by the embodiments of the present invention. Among them, the device status data is real-time device status data, and the user behavior data is real-time user behavior data. Figure 2 It includes the following steps:

[0085] Step S201: Input the real-time device status data and real-time user behavior data into a preset user behavior model for fitting judgment to obtain a model analysis result.

[0086] In the process of specifically implementing step S201, when the obtained device status data is real-time device status data and the user behavior data is real-time user behavior data (i.e., triggering the real-time resource adjustment mechanism), the real-time device status data and the real-time user behavior data are input into a preset user behavior model for fitting judgment to obtain a model analysis result (such as a classification result).

[0087] Judge whether it is necessary to adjust the current network policy of the network device according to the model analysis result.

[0088] For example: According to the model analysis result, if it is judged that the channel utilization rate of an access point in a certain area continues to increase (caused by terminal aggregation), the current network policy of the network device can be adjusted at this time (such as increasing the channel power, roaming to other APs, etc.), so as to avoid the channel utilization rate being too high and affecting the network of the entire access point.

[0089] Another example: Through multi-day data analysis, if the portal authentication concurrency of a certain controller is large in the morning, then it can be judged according to the model analysis result whether other service resources of the controller can be released and whether the requirements are met after the release.

[0090] If it is judged according to the model analysis result that it is necessary to adjust the current network policy of the network device, execute step S202.

[0091] Step S202: Adjust the network policy based on the policy database and the model analysis result.

[0092] In the process of specifically implementing step S202, if it is judged according to the model analysis result that it is necessary to adjust the current network policy of the network device, adjust the network policy based on the policy database and the model analysis result.

[0093] It should be noted that when adjusting the network policy of the network device, it will involve specific configuration items of the network device, and most network policies are obtained from the built-in policy database.

[0094] When adjusting the network policy, determine which network policies need to be adjusted (specifically, adjust each configuration item) according to the policy database and the model analysis result. Among them, the adjustment of some network policies will also involve the assignment of specific values.

[0095] The above Figure 2 is the related description of how to adjust the network policy when triggering the real-time resource adjustment mechanism.

[0096] For the above embodiments of the present invention Figure 1 Regarding the invocation of the user behavior model to process the device status data and the user behavior data involved in step S102, refer to Figure 3, showing the flowchart of adjusting network policies provided by the embodiments of the present invention. Among them, the device status data is historical device status data, and the user behavior data is historical user behavior data. Figure 3 It includes the following steps:

[0097] Step S301: Input the historical device status data and historical user behavior data into a preset user behavior model for prediction to obtain a model prediction result.

[0098] In the process of specifically implementing step S301, when the obtained device status data is historical device status data and the user behavior data is historical user behavior data (i.e., triggering the prediction mechanism), input the historical device status data and historical user behavior data into the user behavior model for prediction to obtain a model prediction result, which can indicate that the network policy needs to be adjusted.

[0099] Step S302: Adjust the network policy according to the model prediction result.

[0100] In the process of specifically implementing step S302, use the decision tree algorithm to convert the model prediction result into an adjusted network policy.

[0101] For example: If the model prediction result indicates that the pressure of portal authentication of network devices is generally high during the morning rush hour, then it is necessary to adjust resources such as concurrent connections and memory in the network policy of network devices, so as to ensure that the adjusted network policy can guarantee the smooth operation of the authentication service.

[0102] The above Figure 3 is the relevant description of how to adjust the network policy when triggering the prediction mechanism.

[0103] In practical applications, this solution can be implemented through "Data Collection and Storage Service", "Machine Learning Analysis Service", "Real-time Resource Adjustment Service" and "User Behavior Prediction Service". The following uses Figure 4 the workflow example diagram of the network policy adjustment method shown to explain the working principles and interactions of the foregoing "Data Collection and Storage Service" and other objects.

[0104] 1. Combining Figure 4 As can be seen from the content shown, the Data Collection and Storage Service uses WAC to regularly report device status data and user behavior data to the Real-time Resource Adjustment Service. Among them, the data collected by WAC is batch forwarded to the Real-time Resource Adjustment Service after preliminary cleaning and formatting processing, ensuring that the Real-time Resource Adjustment Service can receive the latest device status data and user behavior data, so as to provide a basis for subsequent network policy adjustment and decision-making.

[0105] 2. After the real-time resource adjustment service receives the device status data and user behavior data, it analyzes the device status data and user behavior data of each network device, and uses the pre-constructed user behavior model for fitting judgment. According to the model analysis results, it determines whether it is necessary to adjust the current network policy of the network device.

[0106] If it is necessary to adjust the current network policy, the real-time resource adjustment service generates the adjusted network policy and distributes the adjusted network policy to each network device, so as to ensure reasonable resource allocation and maximize network performance.

[0107] 3. The machine learning analysis service is mainly responsible for obtaining the latest device status data and user behavior data from the data collection and storage service, and conducting in-depth analysis based on the device status data and user behavior data, so as to train a personalized user behavior model. In addition, the machine learning analysis service also regularly obtains the device status data and user behavior data of each network device at set time intervals, extracts features and trains the model for the device status data and user behavior data by using machine learning algorithms, so as to continuously update the user behavior model and improve the prediction accuracy and practicality of the user behavior model.

[0108] 4. The user behavior prediction service calls this user behavior model to predict the future device requirements of the network device. If it is predicted that the network policy of the network device needs to be adjusted, the user behavior prediction service will distribute the adjusted network policy to each network device through the real-time resource adjustment service.

[0109] For example: If it is predicted that the pressure of portal authentication is generally high during the morning rush hour for network devices, then the user behavior prediction service will pre-adjust resources such as concurrent connections and memory in the network policy of the network device, so as to ensure that the adjusted network policy can guarantee the smooth operation of the authentication service; after the peak of the working day, the resources will be adjusted back to services such as network access to achieve efficient use of resources.

[0110] It should be noted that from Figure 4 the content shown, when the real-time resource adjustment mechanism is triggered, the real-time resource adjustment service will distribute the adjusted network policy; when the prediction mechanism is triggered, the user behavior prediction service will distribute the adjusted network policy;

[0111] When the real-time resource adjustment mechanism and the prediction mechanism are triggered simultaneously (or triggered within a short period of time) and generate adjusted network policies, and the policy parameters of the adjusted network policies generated by the real-time resource adjustment mechanism and the prediction mechanism are inconsistent, then the adjusted network policy generated by the real-time resource adjustment mechanism shall prevail, that is, the network device adopts the adjusted network policy generated by the real-time resource adjustment mechanism, which can guarantee the current service.

[0112] For data collection and storage services, the data collection and storage services can use network proxy technology to capture various user behavior data of users in the network in real time and obtain device status data.

[0113] Among them, user behavior data includes but is not limited to terminal portal (portal website) authentication frequency, network access frequency, network usage duration, accessed content, hot functions, etc.

[0114] Device status data includes wireless access point channel utilization rate, terminal signal strength, ap radio frequency transmission power, switch load, controller CPU utilization rate, memory, disk space, concurrency, etc.

[0115] During the process of data collection and storage, the data collection and storage services mainly focus on aspects such as "data integrity and real-time", "high-concurrency data collection", "efficient data storage structure", and "data retrieval and large-scale storage support". The following will explain the content of these aspects respectively.

[0116] Data integrity and real-time: To ensure data integrity, the data collection and storage services will use data verification technology to analyze and verify the collected data in real time to ensure that all obtained data is valid and accurate. In addition, timestamps and unique identifiers are used to mark each data record for convenient subsequent tracking and auditing. At the same time, a fault tolerance mechanism is designed for possible data loss and errors, such as redundant data storage and real-time backup, to ensure data security and reliability.

[0117] High-concurrency data collection: When implementing the functions of data collection and storage services, concurrency issues need to be considered to ensure reliability and performance under high load. Concurrency issues can be solved from aspects such as load balancing, horizontal scaling, request rate limiting, and queue systems.

[0118] Among them, a load balancer is used to evenly distribute the incoming requests to multiple data collection instances to avoid excessive single-point load.

[0119] Horizontal scaling is allowed by adding more collection nodes to dynamically add data collection instances to cope with traffic fluctuations.

[0120] Rate limiting strategies such as the token bucket algorithm and the leaky bucket algorithm are used to control the data collection rate in real time to ensure that the system will not crash due to sudden load.

[0121] A message queue is introduced to put high-concurrency requests into the queue and be asynchronously processed by the backend service, which can effectively solve the problem of instantaneous traffic impact.

[0122] Efficient Data Storage Structure: To support large-scale data storage, this solution adopts a hybrid storage structure that combines a relational database and a non-relational database to achieve flexible data management. Among them, the relational database is used to store structured data, such as user behavior summaries, access statistics, etc. The non-relational database is used for high-concurrency read and write scenarios to store raw logs and unstructured data.

[0123] Data Retrieval and Large-Scale Storage Support: To meet the high-speed data retrieval requirements, means such as caching mechanisms, query optimization, data compression, and archiving are adopted to meet the data retrieval needs.

[0124] Among them, the caching mechanism means: using a distributed cache (such as Redis) to cache popular query results, thereby reducing database access latency.

[0125] Query optimization means: through technical means such as SQL optimization, materialized views, and summary tables, improve the performance of analytical queries and support complex queries required for large-scale data analysis.

[0126] Data compression and archiving means: To reduce the data storage pressure, historical data is regularly compressed and archived to free up storage space, enabling more efficient storage management of active data and historical data.

[0127] The above is the description of the data collection and storage service.

[0128] For the machine learning analysis service, the machine learning analysis service selects a machine learning algorithm suitable for the scenario (such as a deep learning neural network) to build a user behavior model, extracts features and performs pattern recognition on historical data, and then conducts model training and evaluation to optimize algorithm parameters to ensure model accuracy and generalization ability, and improve the accuracy of user behavior prediction.

[0129] In addition, it is also possible to use the source model and differential data to train different user behavior models. Specifically, through transfer learning, a pre-trained source model can be fine-tuned or retrained by combining data from different users or task-specific data, thereby obtaining a user behavior model for a specific user or task.

[0130] Among them, transfer learning has advantages such as saving resources, accelerating the training speed, improving performance, and personalized customization.

[0131] Saving resources means: Since the source model has already been trained, the overall consumption of computing resources can be reduced, and only a small amount of fine-tuning or retraining of the source model is required instead of training a completely new model from scratch.

[0132] Speeding up training means that by leveraging the general features learned by the source model, the training time of the user behavior model can be shortened, and the training efficiency can be improved.

[0133] Improving performance means that since the source model has learned some general features and knowledge, the user behavior model may converge faster and achieve better performance when learning specific tasks.

[0134] Personalized customization means that according to the needs of different users or tasks, different differential data can be used to fine-tune the source model, thereby obtaining a customized user behavior model.

[0135] The above is the description of the machine learning analysis service.

[0136] For the real-time resource adjustment service, the real-time resource adjustment service monitors the changes in user behavior data and device status data, and adjusts the network policy (such as adjusting configuration items such as bandwidth and cache allocation) in real time according to the model analysis results of the user behavior model to ensure the efficient execution of the network policy adjusted in real time, so as to quickly respond to changes in user needs and improve network performance and user experience.

[0137] For the user behavior prediction service, based on historical user behavior data, historical device status data, and the user behavior model, it predicts the future behavior trends of users and adjusts the network policy in advance to adapt to future demand changes, realize intelligent network operation, and improve network performance and stability.

[0138] Generally speaking, based on the content of the above various embodiments, this solution can achieve the following beneficial effects: improve network resource utilization rate, reduce resource waste, and optimize network performance; dynamically adapt to changes in user needs, enhance user experience, and increase user satisfaction; improve the ability of fault prediction and handling, reduce service interruptions, and ensure service continuity; enhance the level of network intelligence, reduce manual intervention, and improve network operation efficiency.

[0139] Corresponding to the method for adjusting a network policy provided by the above embodiments of the present invention, see Figure 5 , the embodiments of the present invention also provide a structural block diagram of a device, and the device includes: an acquisition unit 100, a processing unit 200, and a distribution unit 300;

[0140] The acquisition unit 100 is used to acquire the device status data and user behavior data of the network device.

[0141] In some embodiments, the device condition data is real-time device condition data or historical device condition data, and the user behavior data is real-time user behavior data or historical user behavior data; the acquisition unit 100 is specifically configured to: when triggering the real-time resource adjustment mechanism, acquire the real-time device condition data and real-time user behavior data of the network device regularly uploaded by the wireless controller; when triggering the prediction mechanism, acquire the historical device condition data and historical user behavior data of the network device collected within the historical time period.

[0142] The processing unit 200 is configured to call a preset user behavior model to process the device condition data and user behavior data, and adjust the network policy according to the processing result. The user behavior model is obtained by training a specified model based on sample data.

[0143] In some embodiments, the user behavior model is updated regularly.

[0144] In some embodiments, the device condition data is real-time device condition data, and the user behavior data is real-time user behavior data; the processing unit 200 is specifically configured to: input the real-time device condition data and real-time user behavior data into a preset user behavior model for fitting judgment to obtain a model analysis result; based on the policy database and the model analysis result, adjust the network policy.

[0145] In other embodiments, the device condition data is historical device condition data, and the user behavior data is historical user behavior data; the processing unit 200 is specifically configured to: input the historical device condition data and historical user behavior data into a preset user behavior model for prediction to obtain a model prediction result; adjust the network policy according to the model prediction result.

[0146] Among them, the process of adjusting the network policy according to the model prediction result includes: using the decision tree algorithm to convert the model prediction result into an adjusted network policy.

[0147] The distribution unit 300 is configured to distribute the adjusted network policy to the network device.

[0148] In the embodiments of the present invention, the device condition data and user behavior data of the network device are processed through the user behavior model, and the network policy of the network device is adjusted according to the processing result. On the one hand, it can ensure the stable operation of the network device, and thus ensure the normal operation of the network service. On the other hand, it can flexibly respond to the dynamically changing network environment and user requirements, and can perform network optimization in a timely and accurate manner without manual intervention to ensure the normal operation of the network service, thereby improving the processing efficiency, reducing the response delay and reducing the misjudgment rate.

[0149] Preferably, in combination with Figure 5As shown in the content, the processing unit 200 includes a loading subunit and a training subunit, and the execution principles of each subunit are as follows:

[0150] The loading subunit is used to load a pre-trained source model.

[0151] The training subunit is used to adjust or re-train the source model by using sample data and a transfer learning algorithm to obtain a user behavior model.

[0152] Preferably, an embodiment of the present invention further provides an electronic device, including: a processor, a memory, and a communication interface. Among them, the processor is used to call and execute a program stored in the memory; the memory is used to store the program, and the program is used to implement the network policy adjustment method provided in the above method embodiment; the communication interface is used to communicate with a network device.

[0153] Preferably, an embodiment of the present invention further provides a computer-readable storage medium. A computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, it implements the network policy adjustment method provided in the above method embodiment.

[0154] In summary, an embodiment of the present invention provides a network policy adjustment method, device, electronic device, and storage medium. The device status data and user behavior data of a network device are processed through a user behavior model, and the network policy of the network device is adjusted according to the processing result. By adjusting the network policy of the network device, on the one hand, it can ensure the stable operation of the network device, and thus ensure the normal operation of network services. On the other hand, it can flexibly respond to the dynamically changing network environment and user needs, and can perform network optimization in a timely and accurate manner without manual intervention to ensure the normal operation of network services, reduce response latency, and reduce the misjudgment rate.

[0155] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments. In particular, for a system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The systems and system embodiments described above are only illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0156] Those skilled in the art may further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of the present invention.

[0157] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for adjusting a network strategy, characterized in that: The method comprises: Obtain device status data and user behavior data of network devices; Calling a preset user behavior model to process the device status data and the user behavior data, and adjusting the network strategy according to the processing result, wherein the user behavior model is obtained by training a specified model based on sample data; The adjusted network policy is sent to the network device.

2. The method according to claim 1, characterized in that The device status data is real-time device status data or historical device status data, and the user behavior data is real-time user behavior data or historical user behavior data; Obtain device status data and user behavior data of network devices, including: When the real-time resource adjustment mechanism is triggered, the real-time device status data and real-time user behavior data of the network devices uploaded by the wireless controller at regular intervals are obtained; When the prediction mechanism is triggered, historical device status data and historical user behavior data of the network device collected within a historical time period are obtained.

3. The method according to claim 2, characterized in that The device status data is real-time device status data, and the user behavior data is real-time user behavior data; Calling a preset user behavior model to process the device status data and the user behavior data, and adjusting the network strategy according to the processing result, including: Inputting the real-time device status data and the real-time user behavior data into a preset user behavior model for fitting and judgment to obtain a model analysis result; Based on the policy database and the model analysis results, the network policy is adjusted.

4. The method according to claim 2, characterized in that: The device status data is historical device status data, and the user behavior data is historical user behavior data; Calling a preset user behavior model to process the device status data and the user behavior data, and adjusting the network strategy according to the processing result, including: Inputting the historical device status data and the historical user behavior data into a preset user behavior model for prediction to obtain a model prediction result; Adjust the network strategy according to the prediction results of the model.

5. The method according to claim 4, characterized in that Adjust the network strategy according to the prediction results of the model, including: The model prediction results are converted into adjusted network strategies using a decision tree algorithm.

6. The method according to any one of claims 1 to 5, characterized in that: The process of training a specified model based on sample data to obtain a user behavior model includes: Load a pre-trained source model; The source model is adjusted or retrained using sample data and a transfer learning algorithm to obtain a user behavior model.

7. The method according to any one of claims 1 to 5, characterized in that: Also includes: The user behavior model is updated regularly.

8. A device, characterized in that: The device comprises: An acquisition unit, used to acquire device status data and user behavior data of network devices; A processing unit, configured to call a preset user behavior model to process the device status data and the user behavior data, and adjust the network strategy according to the processing result, wherein the user behavior model is obtained by training a specified model based on sample data; The sending unit is used to send the adjusted network policy to the network device.

9. An electronic device, characterized in that: include: A processor, a memory, and a communication interface, wherein the processor is used to call and execute a program stored in the memory; The memory is used to store a program, and the program is used to implement the network policy adjustment method as described in any one of claims 1 to 7; the communication interface is used to implement communication with a network device.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for adjusting a network policy as described in any one of claims 1 to 7 is implemented.

Images