Load balancing method, device, equipment, storage medium and product of heterogeneous network

Abstract

The application relates to the technical field of infrastructure and wireless transmission, and in particular provides a load balancing method, device and equipment of a heterogeneous network, a storage medium and a product. The method comprises the following steps: predicting a traffic prediction result based on user network historical data; predicting a network perception evaluation result based on network attribute information of each network node; obtaining the load capacity of the network node based on the network perception evaluation result; selecting a target network node with a matching load capacity and traffic prediction result from each network node, and controlling the target network node to provide network resources for users. By predicting the traffic demand of the user within a preset time in the future and the advantages and disadvantages of the network quality of the network node, the load capacity of each network node is obtained. When it is predicted that congestion or a large traffic access demand of the user may occur in a certain network node, traffic scheduling can be performed in advance, congestion is avoided or reduced, and the load balancing effect of the heterogeneous network is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of infrastructure and wireless transmission, and in particular to a load balancing method, device and equipment of a heterogeneous network, a storage medium and a product. BACKGROUND

[0002] With the development of wireless communication technology, the popularity of intelligent terminals and the increasing demand of users for high-speed and seamless network access, heterogeneous network integration has become an important technical trend. Heterogeneous network integration can integrate different types of network resources and provide more diverse services, but it also brings many challenges, among which load balancing is a more urgent problem.

[0003] Most of the existing load balancing methods of heterogeneous networks adopt static configuration or simple traffic scheduling strategy. Static configuration manually allocates fixed traffic distribution ratio to each network node based on the processing capacity, hardware resources, bandwidth and other factors of the network node, and the experience and judgment of the network administrator. Simple traffic scheduling strategy uses a polling strategy to allocate requests from the client to each network node in turn to ensure that each node can process a certain number of requests, but this method cannot accurately predict the changes in network traffic and cannot dynamically adjust resource allocation according to the network state. Therefore, in the heterogeneous network environment, the existing methods often fail to achieve the desired load balancing effect, affecting the overall performance of the network and user experience. SUMMARY

[0004] The present application is proposed in view of the above problems, and provides a load balancing method, device and equipment of a heterogeneous network, a storage medium and a product.

[0005] According to one aspect of the present application, a load balancing method of a heterogeneous network is provided, comprising:

[0006] Based on the obtained user network history data of the user in the heterogeneous network cell, a traffic prediction result representing the traffic demand degree of the user in a future preset time is predicted; and based on the obtained network attribute information of each network node of the heterogeneous network cell, a network perception evaluation result representing the network quality of each network node is predicted;

[0007] Based on the network perception evaluation result, the load capacity of the network node is obtained;

[0008] From each of the network nodes, a target network node with a load capacity matching the traffic prediction result is selected, and the target network node is controlled to provide network resources for the user.

[0009] Further, the load balancing method of the heterogeneous network according to an aspect of the present application further comprises: predicting a traffic prediction result representing a traffic demand degree of the user in a future preset time based on the obtained user network history data of the user in the heterogeneous network cell, including:

[0010] extracting a first group of feature data, a second group of feature data and a third group of feature data from the user network history data; the first group of feature data is used to represent the traffic demand change rule of the user in a preset historical time range, the second group of feature data is used to represent the statistical characteristics of the traffic used by the user, and the third group of feature data is used to represent external factors affecting the traffic demand of the user;

[0011] Based on the first group of feature data, the second group of feature data and the third group of feature data, the traffic trend is predicted, and the traffic prediction result is predicted.

[0012] Further, the load balancing method of the heterogeneous network according to an aspect of the present application further comprises: selecting a target network node matching the load capacity and the traffic prediction result from each of the network nodes, including: in the case that there are multiple matching network nodes matching the load capacity and the traffic prediction result in each of the network nodes, obtaining the location of the user;

[0013] From the multiple matching network nodes, multiple nearest network nodes with a location difference from the location of the user within a distance threshold are screened;

[0014] The target network node is determined from the multiple nearest network nodes.

[0015] Further, the load balancing method of the heterogeneous network according to an aspect of the present application further comprises: determining the target network node from the multiple nearest network nodes, including:

[0016] Obtaining the target nearest network node closest to the user from the multiple nearest network nodes;

[0017] In the case that the idle resource amount of the target nearest network node is greater than or equal to the resource demand amount represented by the traffic prediction result, the target nearest network node is determined as the target network node;

[0018] In the case that the idle resource amount of the target nearest network node is less than the resource demand amount represented by the traffic prediction result, the difference idle resource amount is calculated based on the idle resource amount and the resource demand amount, the remaining nearest network node with an idle resource amount greater than or equal to the difference idle resource amount is obtained from the multiple nearest network nodes, and the target nearest network node and the remaining nearest network node are taken as the target network node.

[0019] Further, the load balancing method of the heterogeneous network according to an aspect of the present application further comprises: in the case that there are a plurality of users in the heterogeneous network cell, determining a priority of each user based on a user type and a user attribute included in the traffic prediction result of each user; the priority is used to represent an order of selecting a user of a target network node;

[0020] Correspondingly, the target network node is selected from the network nodes based on the load capacity and the traffic prediction result, comprising:

[0021] The target network node is selected from the network nodes based on the priority and the load capacity and the traffic prediction result of each user.

[0022] Further, the load balancing method of the heterogeneous network according to an aspect of the present application further comprises, after the target network node is controlled to provide network resources for the user:

[0023] Obtaining a current carrying value of the network node; the current carrying value is used to represent a proportion of currently occupied network resources of the network node;

[0024] Determining an expected carrying value of each network node of the heterogeneous network cell based on the current carrying value and a current carrying value of a remaining network node of the heterogeneous network cell;

[0025] Migrating services on each network node of the heterogeneous network cell according to the expected carrying value.

[0026] Further, the load balancing method of the heterogeneous network according to an aspect of the present application further comprises: determining an expected carrying value of each network node of the heterogeneous network cell based on the current carrying value and a current carrying value of a remaining network node of the heterogeneous network cell, comprising:

[0027] Based on the current carrying value and the current carrying value of the remaining network node of the heterogeneous network cell, obtaining a first network node with a maximum carrying value and a second network node with a minimum carrying value from each network node of the heterogeneous network cell;

[0028] Determining the expected carrying value of the first network node and the second network node as an average of the maximum carrying value and the minimum carrying value, and determining the expected carrying value of a network node outside the first network node and the second network node as a current carrying value.

[0029] According to another aspect of the present application, a load balancing device of a heterogeneous network is provided, comprising:

[0030] a prediction module configured to predict, based on the obtained user network history data of the user in the heterogeneous network cell, a traffic prediction result representing a degree of traffic demand of the user in a future preset time, and predict, based on the obtained network attribute information of each network node of the heterogeneous network cell, a network perception evaluation result representing a quality of each network node;

[0031] an acquisition module configured to acquire, based on the network perception evaluation result, a load capacity of the network node;

[0032] a selection module configured to select, from each network node, a target network node whose load capacity matches the traffic prediction result;

[0033] a control module configured to control the target network node to provide network resources for the user.

[0034] According to yet another aspect of the present application, a computer device is provided, which includes a memory, a processor, and a computer program stored in the memory, and the processor executes the computer program to implement the method of the above aspect.

[0035] According to yet another aspect of the present application, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the method of the above aspect.

[0036] According to yet another aspect of the present application, a computer program product is provided, which includes a computer program, and the computer program is executed by a processor to implement the method of the above aspect.

[0037] As will be described in detail below, according to the load balancing method, device, equipment, storage medium and product of the heterogeneous network of the embodiments of the present application, by predicting the traffic demand of the user in the future preset time and the quality of the network node, the load capacity of each network node is obtained, when it is predicted that a certain network node may be congested (i.e. the load capacity is poor) or the traffic access demand of the user is large, the traffic scheduling can be performed in advance, the user is allocated to the network node with suitable load capacity, the congestion situation is avoided or alleviated, and the load balancing effect in the heterogeneous network environment is improved.

[0038] It is to be understood that both the foregoing general description and the following detailed description are exemplary, and are intended to provide further explanation of the subject technology. BRIEF DESCRIPTION OF DRAWINGS

[0039] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which:

[0040] Figure 1 is a flow chart illustrating a load balancing method of a heterogeneous network according to an embodiment of the present application.

[0041] Figure 2 is yet another flow chart illustrating a load balancing method of a heterogeneous network according to an embodiment of the present application.

[0042] Figure 3 is a flow chart illustrating a dynamic resource scheduling according to an embodiment of the present application.

[0043] Figure 4 is a structural schematic diagram of a load balancing apparatus of a heterogeneous network according to an embodiment of the present application.

[0044] Figure 5 is a structural schematic diagram of a computer device according to an embodiment of the present application.

[0045] Figure 6 is a schematic diagram of a computer program product according to an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to make the objects, technical solutions and advantages of the present application more apparent, the following will describe the example embodiments according to the present application in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.

[0047] Load balancing is an important network optimization technology, which realizes the balanced distribution of load by distributing different types and demands of service traffic to different network nodes, and improves network performance and resource utilization.

[0048] Existing load balancing methods for heterogeneous networks mostly employ static configuration or simple traffic scheduling strategies. Static configuration manually allocates a fixed traffic distribution ratio to each network node based on factors such as the network node's processing capacity, hardware resources, and bandwidth, relying on the network administrator's experience and judgment. Simple traffic scheduling strategies use a round-robin approach, distributing requests from clients sequentially to each network node, ensuring that each node can handle a certain number of requests. However, this method cannot accurately predict changes in network traffic, nor can it dynamically adjust resource allocation according to network conditions. Therefore, in heterogeneous network environments, these existing methods often fail to achieve ideal load balancing results, impacting overall network performance and user experience.

[0049] The above description, with reference to the accompanying drawings, illustrates a load balancing method, apparatus, device, storage medium, and product for heterogeneous networks according to embodiments of this application. By predicting user traffic demand within a preset timeframe and the network quality of network nodes, the load capacity of each network node can be obtained. When it is predicted that a network node may experience congestion (i.e., poor load capacity) or that a user's traffic access demand is high, traffic scheduling can be performed in advance to allocate users to network nodes with matching load capacities, thus avoiding or mitigating congestion. By setting a load value for each network node, the load status of the network node is obtained. When the difference between the maximum and minimum load values ​​is significant, specifically, when the maximum load value is three times or more than the minimum load value, the services on the loads of the two network nodes can be dynamically adjusted, distributing the services on the loads of the two network nodes evenly, so that the entire heterogeneous network environment is in a dynamic equilibrium state, which is beneficial to improving the load balancing effect in a heterogeneous network environment.

[0050] To facilitate understanding of this embodiment, a load balancing method for heterogeneous networks disclosed in this application will first be described in detail. The execution entity of the load balancing method for heterogeneous networks provided in this application is generally a computer device with certain computing capabilities. This computer device may include, for example, a terminal device, a server, or other processing devices. The terminal device may be a user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, vehicle-mounted device, wearable device, etc. In some possible implementations, this load balancing method for heterogeneous networks can be implemented by the processor calling computer-readable instructions stored in memory.

[0051] like Figure 1 The diagram shows a flowchart of a load balancing method for heterogeneous networks provided in an embodiment of this application. The method includes steps S1-S3:

[0052] S1: predicting a traffic prediction result representing a degree of traffic demand of the user in a future preset time based on the obtained user network history data of the user in the heterogeneous network cell, and predicting a network perception evaluation result representing a quality of the network of each network node based on the obtained network attribute information of each network node of the heterogeneous network cell.

[0053] Specifically, the prediction process of the traffic prediction result includes steps S1.1-S1.3:

[0054] S1.1 extracts a first group of feature data, a second group of feature data and a third group of feature data from the user network history data.

[0055] The first group of feature data is used to represent the traffic demand change rule of the user in a preset historical time range, such as online duration, access time, access frequency, traffic trend, inbound and outbound traffic record; the second group of feature data is used to represent the statistical characteristics of the traffic used by the user, such as packet size, transmission speed, total traffic, peak traffic, average traffic; and the third group of feature data is used to represent external factors affecting the traffic demand of the user, such as traffic type, terminal type, source address, destination address.

[0056] S1.2 trains a first preset classification model based on the first group of feature data, the second group of feature data and the third group of feature data, and obtains a traffic prediction model, for example, an LSTM model;

[0057] S1.3 predicts a traffic trend prediction result by the traffic prediction model.

[0058] Specifically, the prediction process of the network perception evaluation result includes steps S1-1-S1-3:

[0059] S1-1 obtains the network attribute information of each network node of the heterogeneous network cell.

[0060] The network attribute information includes CPU usage, memory occupancy, rate, disk I / O, interface type, transmission quality, packet loss rate, throughput, time delay, network jitter, node relationship, node path, node delay, node bandwidth and other data information.

[0061] S1-2 trains a second preset classification model based on the network attribute information, and obtains a network perception classification model, for example, an SVM model;

[0062] S1-3 predicts the quality of the network of each network node by the network perception classification model, and obtains the network perception evaluation result.

[0063] S2: obtaining the load capacity of the network node based on the network perception evaluation result.

[0064] Specifically, the network node load capacity can be obtained by the network perception evaluation result including the occupied resource rate and the average blocking rate and the like, and the network nodes are divided into high load nodes, low load nodes and network bottleneck load nodes according to the load capacity.

[0065] The high load node indicates that the network node has occupied more network resources, i.e., the load bottleneck is reached, and the load capacity is poor; the low load node indicates that the network node has occupied less network resources, and the load capacity is good; and the network bottleneck load node indicates that the network node has reached the load bottleneck, and the load capacity is the worst, and cannot load new network resources.

[0066] S3: Select a target network node whose load capacity matches the traffic prediction result from the network nodes, and control the target network node to provide network resources for the user.

[0067] It needs to be additionally explained by the embodiments of the application that in the case that multiple users exist in the heterogeneous network cell, the priority of the multiple users can be determined based on the user type and the user attribute included in the traffic prediction result of each user, and the priority is used to represent the order of the user selected as the target network node. The user type is, for example, a VIP user with a private network demand such as a confidential agency, a government unit, an enterprise and the like, and the user attribute is, for example, user gender, age and the like. Based on the priority order, the target network node is selected for the user in turn. Specifically, the decision tree model can be trained based on the traffic prediction result of each user, and the priority of each user is predicted by the decision tree model. It also needs to be explained that there can be multiple users in the same priority, and then all the users in the same priority (i.e., in the same batch) can select the target network node at the same time.

[0068] If there is only one network node whose load capacity matches the traffic prediction result in the network nodes, the network node can be directly used as the target network node.

[0069] In the case that multiple matching network nodes whose load capacity matches the traffic prediction result exist in the network nodes, the steps S3.1-S3.3 are specifically included:

[0070] S3.1: Obtain the location of the user. Specifically, the cell where the user is located can be obtained.

[0071] S3.2: From the multiple matching network nodes, multiple nearest network nodes whose location is different from the location of the user by a distance threshold are screened.

[0072] Specifically, the distance threshold can be adjusted according to actual needs, for example, 200 meters, 500 meters, 1000 meters, etc. In this embodiment, 1000 meters is taken as an example, and the matching network nodes with a distance less than or equal to 1000 meters from the user are selected as the nearest network nodes.

[0073] S3.3. Determine the target network node from the plurality of nearest network nodes.

[0074] Specifically, the target nearest network node closest to the user is obtained from the plurality of nearest network nodes; in a case where the idle resource amount of the target nearest network node is greater than or equal to the resource demand amount represented by the traffic prediction result, the target nearest network node is determined as the target network node; in a case where the idle resource amount of the target nearest network node is less than the resource demand amount represented by the traffic prediction result, the difference idle resource amount is calculated based on the idle resource amount and the resource demand amount, the remaining nearest network nodes with an idle resource amount greater than or equal to the difference idle resource amount are obtained from the plurality of nearest network nodes, and the target nearest network node and the remaining nearest network nodes are taken as the target network nodes.

[0075] For example, the resource demand amount represented by the traffic prediction result is 50T, the traffic demand is large, and 50T is the total resource demand amount of the users in the same batch. At this time, the network node matched therewith needs to be a low-load node. When the target nearest network node is a low-load node, the target nearest network node is the target network node. Further, it is also necessary to determine whether the target nearest network node can be the only target network node, that is, whether the target nearest network node can carry all the users in the same batch. The judgment conditions include:

[0076] Condition one: the idle resource amount of the target nearest network node is greater than or equal to the resource demand amount represented by the traffic prediction result (for example, 50T);

[0077] Condition two: the number of users represented by the traffic prediction result (that is, the users in the same batch) is less than or equal to a preset threshold (for example, 1000 people);

[0078] When conditions one and two are met at the same time, the target nearest network node is determined as the only target network node. By setting conditions one and two, it is possible to avoid excessive load of the target nearest network node, and to guarantee signal strength and user service experience.

[0079] If the idle resource amount of the target nearest network node is less than 50T (taking 40T as an example), the 40T resource demand amount represented by the traffic prediction result is allocated to the target nearest network node, and the difference idle resource amount is 10T. From the remaining nearest network nodes, a remaining nearest network node with an idle resource amount greater than or equal to 10T is obtained, and the 10T resource demand amount is allocated to the remaining nearest network node. The target nearest network node and the remaining nearest network node together serve as the target network node. It should be noted that when allocating resources to all users of the same priority, the resource demand amount represented by the traffic prediction result can represent the total resource demand amount of all users in the same batch.

[0080] Suppose there are 10 users in the same batch, and the user numbers are 1-10. When the 40T resource demand amount represented by the traffic prediction result is allocated to the target nearest network node, it can be explained that the resource demand amounts of users 1-7 are allocated to the target nearest network node, and the resource demand amounts of users 8-10 are allocated to the remaining nearest network nodes.

[0081] It should be noted that the 50T resource demand amount represented by the traffic prediction result is allocated to the target nearest network node and the remaining nearest network node, which is to allocate the resource demand amounts of 10 users represented by the traffic prediction result. One user can only be allocated to one network node.

[0082] After the target network node is controlled to provide network resources for the user, the embodiment further includes the following steps:

[0083] S4.1 Obtain the current carrying value of the network node; the current carrying value is used to represent the proportion of the currently occupied network resources of the network node.

[0084] The current carrying value of the network node is updated after each time the target network node is controlled to provide network resources for the user. For example, the specific setting rule of the carrying value can be:

[0085] Rule 1: The total resource amount that the network node can carry is 100T, and the currently occupied network resource amount is 20T. Therefore, the current carrying value of the network node is 20%.

[0086] Rule 2: The number of users that the network node can carry is 2000, and the number of currently carried users is 200. Therefore, the current carrying value of the network node is 10%.

[0087] It should be noted that the carrying value can be set according to rule 1 or rule 2 according to actual needs, or the carrying value can be set in combination with rule 1 and rule 2, and the embodiment is not limited.

[0088] S4.2 Based on the current bearer value and the current bearer value of the remaining network nodes in the heterogeneous network cell, determine the expected bearer value of each network node in the heterogeneous network cell. The specific steps are as follows:

[0089] Based on the current bearer value and the current bearer value of the remaining network nodes in the heterogeneous network cell, obtain the first network node with the maximum bearer value and the second network node with the minimum bearer value from each network node in the heterogeneous network cell.

[0090] The expected carrying capacity of the first network node and the second network node is determined to be the average of the maximum carrying capacity and the minimum carrying capacity. The expected carrying capacity of the network nodes other than the first network node and the second network node is determined to be the current carrying capacity.

[0091] For example, if the current carrying capacity of the first network node is 90% (maximum carrying capacity) and the current carrying capacity of the second network node is 20% (minimum carrying capacity), then the formula for calculating the expected carrying capacity of the first and second network nodes is: (90% + 20%) / 2 = 55%. The current carrying capacity of the network nodes other than the first and second network nodes is the expected carrying capacity.

[0092] S4.3 Migrate services on each network node of the heterogeneous network cell according to the expected carrying capacity.

[0093] In conjunction with S4.2, some services of the first network node are migrated to the second network node, so that the amount of network resources occupied (i.e., the current carrying capacity) of the first network node and the second network node is close to or reaches 55%.

[0094] In addition, this application embodiment also provides another implementation of S2, as follows:

[0095] Based on the resource utilization rate included in the network awareness assessment results, the current load value of the network node is obtained. The load capacity corresponding to the network node is obtained through the current load value of the network node. For example, nodes with a current load value between 0% and 40% can be manually defined as low load nodes, nodes with a current load value between 40% and 90% can be defined as high load nodes, and nodes with a current load value between 90% and 100% can be defined as network bottleneck load nodes.

[0096] like Figure 2 The diagram shows a flowchart of a load balancing method for heterogeneous networks provided in an embodiment of this application. The method includes steps S1-S5:

[0097] S1: Heterogeneous network data acquisition.

[0098] Specifically, the heterogeneous network data includes heterogeneous environment traffic monitoring data and heterogeneous network historical data, wherein the heterogeneous environment traffic monitoring data includes user behavior data and network quality data, and the heterogeneous network historical data includes traffic history data, network topology data and network device performance data.

[0099] S2: constructing a traffic prediction model to obtain a traffic prediction result. Specifically, the following steps are included:

[0100] S2.1 extracting a first group of feature data, a second group of feature data and a third group of feature data from user network historical data.

[0101] The user network historical data includes user behavior data and traffic history data, specifically,

[0102] The first group of feature data is obtained by extracting the user network historical data based on time characteristics, and is related to network traffic time series, and is used to capture the traffic change rule of the user in different time periods, such as online duration, access time, access frequency, traffic trend, inbound and outbound traffic record.

[0103] The second group of feature data is obtained by extracting the user network historical data based on statistical characteristics of the traffic data itself, and can be used to predict the basic attributes and change trend of the traffic in a preset future time, such as packet size, transmission speed, total traffic, peak traffic and average traffic.

[0104] The third group of feature data is obtained by extracting the user network historical data based on external factor characteristics that affect the size of the user's traffic demand, such as traffic type, terminal type, source address, destination address, and also holidays, weather, seasons, etc.

[0105] S2.2 training an LSTM model based on the first group of feature data, the second group of feature data and the third group of feature data to construct a traffic prediction model;

[0106] Specifically, the Keras library is used to set the dimension of the input layer, the number of LSTM layers, and the number of neurons units of each LSTM layer, use the dropout layer to prevent overfitting, configure the activation function and dimension of the Dense output layer, use the mean_squared_error loss function to measure the difference between the predicted value and the actual value, and the adam optimizer updates the weights of the model according to the gradient of the loss function. By configuring parameters such as learning rate, batch size, epoch number, loss function, number of hidden layers and number of units, the traffic prediction model that can accurately predict the network traffic change trend is constructed.

[0107] S2.3, predicting traffic trend by the traffic prediction model, and obtaining a traffic prediction result.

[0108] Specifically, the traffic prediction model can be input with user data such as a user name and a password, and the user name can be a user name, a nickname, or a mobile phone number, etc. The traffic prediction model outputs a traffic prediction result according to the user data, which can be used not only to predict the traffic demand trend of each user in a future preset time, but also to determine whether the traffic demand trend belongs to a periodic pattern or a sudden increase pattern. If it is a periodic pattern, the network resource allocation can be adjusted in advance according to the traffic prediction result combined with the load condition of the cell or network node to avoid overload. If it is a sudden increase pattern, it is determined whether there is malicious access and the possible sudden increase time, sudden increase traffic size, voice demand, etc. are predicted. If there is malicious access, flow limiting or access termination decisions are taken in time. If there is no malicious access, the network node suitable for the user is selected according to the traffic prediction result, and the load state of each network node is dynamically adjusted, which is beneficial to reduce the load of the network node and realize load balancing.

[0109] S3: constructing a network perception classification model to obtain a network perception evaluation result. Specifically, the following steps are included:

[0110] S3.1, obtaining network attribute information of each network node of the heterogeneous network cell.

[0111] The network attribute information includes data information in three dimensions of network quality data, network topology data, and network device performance data.

[0112] S3.2, training the SVM model (Support Vector Machine) based on the network attribute information to obtain a network perception classification model.

[0113] The SVM model is used to extract features from the data in the three dimensions, reduce the data dimension, eliminate outliers, and rely on the strong classification ability of SVM in processing high-dimensional data and nonlinear relationships. The extracted features are combined into a feature vector, which specifically includes CPU usage, memory occupancy, rate, disk I / O, interface type, transmission quality, packet loss rate, throughput, latency, network jitter, node relationship, node path, node delay, node bandwidth, etc. The feature vector is input into the SVM model, which is converted into a linear classification problem in a certain dimension feature space through nonlinear transformation. In the dual problem of high-dimensional linear support vector machine learning, the objective function and the classification decision function involve the inner product of instances and between instances, and the inner product between two instances is converted through nonlinear transformation.

[0114] S3.3, predicting the pros and cons of the network quality of the network node by the network perception classification model to obtain a network perception evaluation result.

[0115] Specifically, based on the network quality, the network perception evaluation result sets three latitudes of good, medium and poor, which correspond to different weights in the network perception classification model, for example, the weights are 30%, 50% and 20% in turn, and the specific size of the weight can be adjusted according to the actual situation. The network perception classification model finds the best hyperplane by solving the optimization problem, so that the sample nodes of different categories are separated as much as possible, and judges whether the node network state is good or poor, thereby realizing the classification of the node network quality. According to the classification result and the node performance index, the utility function is set, the performance index of the current network node is substituted into the utility function, the utility value is calculated, the advantages and disadvantages of the network node are evaluated according to the utility value, various factors are comprehensively considered, the decision is made based on fuzzy logic, the uncertainty and incomplete information are processed, the feasibility of the current network node as the target network node is evaluated, and more accurate basis is provided for resource scheduling decision.

[0116] S4: Dynamic resource scheduling, as shown in Figure 3 Specifically, steps S4.1-S4.3 are included:

[0117] S4.1 Based on the network perception evaluation result, the load capacity of the network node is obtained.

[0118] Specifically, the load capacity of the network node can be obtained according to the following two ways:

[0119] Method one: The load capacity of the network node is obtained through the parameters such as the occupied resource rate and the average blocking rate included in the network perception evaluation result, and the network node is divided into high-load node, low-load node and network bottleneck load node according to the load capacity.

[0120] Method two: Based on the occupied resource rate included in the network perception evaluation result, the current carrying value of the network node is obtained, which is used to represent the proportion of the currently occupied network resources of the network node. For example, the total resource amount that the network node can load is 100T, and the currently occupied network resource amount is 20T, so the current carrying value of the network node is 20%. After the target network node provides network resources for users each time, the current carrying value of the network node is updated. The current carrying value of the network node is obtained to obtain the corresponding load capacity of the network node, for example, the current carrying value between 0-40% is artificially set as a low-load node, the current carrying value between 40%-90% is artificially set as a high-load node, and the current carrying value between 90%-100% is artificially set as a network bottleneck load node.

[0121] Among them, the high-load node means that the network node has occupied more network resources, that is, it will reach the load bottleneck, and the load capacity is poor; the low-load node means that the network node has occupied less network resources, and the load capacity is good; the network bottleneck load node means that the network node has reached the load bottleneck, and the load capacity is the worst, and it cannot load new network resources.

[0122] S4.2 determine priorities of the plurality of users based on the user types and user attributes included in the traffic prediction results of the users, the priorities being used to represent the order of the users in selecting the target network node.

[0123] Specifically, a decision tree model can be trained based on the traffic prediction results of the users. The decision tree supervised learning method is used to input the traffic prediction results into the decision tree model, select the best partition attribute through information gain, Gini impurity and other indicators, set the corresponding resource allocation weight and priority, construct a tree structure for decision making, and prune the decision tree to reduce the training time of the decision number, prevent overfitting, constantly optimize the model performance through the feedback mechanism, and output the output results used to represent the priorities of the plurality of users. Based on the output results, the target network node is selected for the users in turn, which helps to realize the dynamic resource scheduling strategy.

[0124] According to the traffic prediction results, the user types and the network types of the network nodes are combined, the user types are for example confidential agencies, government units, enterprises and institutions, whether there is a demand for a private network, etc., and the network types are for example QOS guarantee, uplink and downlink rate, PRB, 5QI, etc. The dynamic resource scheduling strategy is preferentially executed, the resource allocation of each network node is dynamically adjusted, and the state of each network node is guaranteed to be good. In addition, for users with high call requirements, the voice service situation of the user terminal can be set according to the user terminal, the VoLTE, EPSFallback, VoNR and other systems can select the corresponding frequency band for carrying; for high demand users, the frequency band 700M, 2.6T, 4.9T function can be used to preferentially guarantee the voice service quality.

[0125] S4.3 select a target network node with a load capacity matching the traffic prediction result from the plurality of network nodes, and control the target network node to provide network resources for the user.

[0126] In the case where there are a plurality of matching network nodes with a load capacity matching the traffic prediction result in the plurality of network nodes, the following steps are specifically included:

[0127] S4.3.1 obtain the location of the user. Specifically, the cell where the user is located is obtained.

[0128] S4.3.2 select a plurality of nearest network nodes with a location that differs from the location of the user by a distance threshold from the plurality of matching network nodes.

[0129] Specifically, the distance threshold can be adjusted according to actual needs, for example, 200 meters, 500 meters, 1000 meters, etc. In this embodiment, 1000 meters is taken as an example, and the matching network nodes with a distance less than or equal to 1000 meters from the user are selected as the nearest network nodes.

[0130] S4.3.3 determining the target network node from the plurality of recent network nodes.

[0131] Specifically, the target recent network node closest to the user is obtained from the plurality of recent network nodes; in the case that the idle resource amount of the target recent network node is greater than or equal to the resource demand amount represented by the traffic prediction result, the target recent network node is determined as the target network node; in the case that the idle resource amount of the target recent network node is less than the resource demand amount represented by the traffic prediction result, the difference idle resource amount is calculated based on the idle resource amount and the resource demand amount, the remaining recent network nodes with the idle resource amount greater than or equal to the difference idle resource amount are obtained from the plurality of recent network nodes, and the target recent network node and the remaining recent network nodes are taken as the target network nodes.

[0132] For example, the resource demand amount represented by the traffic prediction result is 50T, and the traffic demand is large, so the network node matched therewith needs to be a low-load node. When the target recent network node is a low-load node, the target recent network node is the target network node. Further, it is also necessary to determine whether the idle resource amount of the target recent network node is greater than 50T. If the idle resource amount of the target recent network node is 50T or above, the target recent network node is determined as the only target network node. If the idle resource amount of the target recent network node is less than 50T (for example, 40T), the resource demand amount of 40T represented by the traffic prediction result is allocated to the target recent network node, the difference idle resource amount is 10T, the remaining recent network nodes with the idle resource amount greater than or equal to 10T are obtained from the remaining plurality of recent network nodes, the resource demand amount of 10T is allocated to one of the remaining recent network nodes, and the target recent network node and the remaining recent network nodes are taken as the target network nodes.

[0133] S5: load balancing optimization. Specifically, the following steps are included:

[0134] S5.1 obtaining the current load value of the network node; the current load value is used to represent the proportion of the network resources currently occupied by the network node. For example, the total resource amount that can be loaded by the network node is 100T, and the network resource amount currently occupied is 20T, so the current load value of the network node is 20%. After the target network node provides network resources for the user each time, the current load value of the network node is updated.

[0135] S5.2 determining the expected load value of each network node of the heterogeneous network cell based on the current load value and the current load value of the remaining network nodes of the heterogeneous network cell. The specific step process is as follows:

[0136] acquiring, from each network node of the heterogeneous network cell, a first network node with a maximum load value and a second network node with a minimum load value based on the current load value of the current bearer and the current load value of the remaining network nodes of the heterogeneous network cell;

[0137] determining an expected load value of the first network node and the second network node as an average of the maximum load value and the minimum load value, and determining an expected load value of the network nodes other than the first network node and the second network node as the current load value.

[0138] For example, the current load value of the first network node is 90% (maximum load value), and the current load value of the second network node is 20% (minimum load value). Then, the calculation formula of the expected load value of the first network node and the second network node is (90%+20%) / 2=55%, and the current load value of the network nodes other than the first network node and the second network node is the expected load value.

[0139] S5.3 Migrating services on each network node of the heterogeneous network cell according to the expected load value.

[0140] In combination with S5.2, part of the services of the first network node are migrated to the second network node, so that the occupied network resource amount (i.e. the current load value) of the first network node and the second network node approaches or reaches 55%.

[0141] According to another aspect of the embodiments of the present application, a load balancing device of a heterogeneous network is provided, as shown in the figure, the device comprises: Figure 4

[0142] a prediction module 101 configured to predict a traffic prediction result representing a traffic demand degree of a user in a future preset time based on obtained user network history data of the user in a heterogeneous network cell, and predict a network perception evaluation result representing a network quality of each network node of the heterogeneous network cell based on obtained network attribute information of each network node.

[0143] an acquisition module 102 configured to obtain a load capacity of the network node based on the network perception evaluation result.

[0144] a selection module 103 configured to select a target network node with a load capacity matching the traffic prediction result from each network node.

[0145] a control module 104 configured to control the target network node to provide network resources for the user.

[0146] In one or more embodiments, the prediction module 101 is configured to:

[0147] ​extracting a first group of feature data, a second group of feature data and a third group of feature data from the user network history data; the first group of feature data is used to represent the traffic demand change law of the user in a preset history time range, the second group of feature data is used to represent the statistical characteristics of the traffic used by the user, and the third group of feature data is used to represent external factors affecting the traffic demand size of the user;

[0148] performing traffic trend prediction based on the first group of feature data, the second group of feature data and the third group of feature data, and predicting the traffic prediction result.

[0149] In one or more embodiments, the selection module 103 is configured to:

[0150] In the case that there are multiple matching network nodes in each of the network nodes, the position of the user is obtained;

[0151] From the multiple matching network nodes, multiple nearest network nodes whose positions differ from the position of the user by a distance threshold are screened out;

[0152] The target network node is determined from the multiple nearest network nodes.

[0153] In one or more embodiments, the selection module 103 is further configured to:

[0154] From the multiple nearest network nodes, the target nearest network node closest to the user is obtained;

[0155] In the case that the idle resource amount of the target nearest network node is greater than or equal to the resource demand amount represented by the traffic prediction result, the target nearest network node is determined as the target network node;

[0156] In the case that the idle resource amount of the target nearest network node is less than the resource demand amount represented by the traffic prediction result, the difference idle resource amount is calculated based on the idle resource amount and the resource demand amount, the remaining nearest network node whose idle resource amount is greater than or equal to the difference idle resource amount is obtained from the multiple nearest network nodes, and the target nearest network node and the remaining nearest network node are taken as the target network node.

[0157] The load balancing device is further configured to: in the case that there are multiple users in the heterogeneous network cell, determine the priority of each user based on the user type and user attribute included in the traffic prediction result of each user; and the priority is used to represent the order of selecting the target network node for the user;

[0158] Correspondingly, the selection module 103 is further configured to:

[0159] select, from the network nodes, a target network node whose load capacity matches the traffic prediction result of each user based on the priority.

[0160] The load balancing apparatus is further configured to:

[0161] obtain a current carrying value of the network node, the current carrying value being used to represent a proportion of currently occupied network resources of the network node;

[0162] determine an expected carrying value of each network node of the heterogeneous network cell based on the current carrying value and current carrying values of the remaining network nodes of the heterogeneous network cell;

[0163] migrate services on each network node of the heterogeneous network cell according to the expected carrying value.

[0164] In one or more embodiments, determining the expected carrying value of each network node of the heterogeneous network cell based on the current carrying value and current carrying values of the remaining network nodes of the heterogeneous network cell comprises:

[0165] obtaining, from each network node of the heterogeneous network cell, a first network node with a maximum carrying value and a second network node with a minimum carrying value based on the current carrying value and current carrying values of the remaining network nodes of the heterogeneous network cell;

[0166] determining the expected carrying value of the first network node and the second network node as an average of the maximum carrying value and the minimum carrying value, and determining the expected carrying value of the network nodes other than the first network node and the second network node as the current carrying value.

[0167] The load balancing apparatus of the heterogeneous network provided by the embodiments of the present application and the load balancing method of the heterogeneous network provided by the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the method adopted, run or implemented by them.

[0168] The embodiments of the present application further provide a computer device for executing the load balancing method of the heterogeneous network. Please refer to Figure 5 which shows a schematic diagram of a computer device provided by some embodiments of the present application. As shown in Figure 5As shown, the computer device 8 comprises a processor 800, a memory 801, a bus 802 and a communication interface 803, the processor 800, the communication interface 803 and the memory 801 are connected through the bus 802; the memory 801 stores a computer program which can run on the processor 800, and the processor 800 runs the computer program to execute the load balancing method of the heterogeneous network provided by any one of the preceding embodiments of the present application.

[0169] The memory 801 can include a high-speed random access memory (RAM: Random Access Memory) and can also include a non-volatile memory such as at least one disk memory. The communication connection between the device network element and at least one other network element is realized through at least one communication interface 803 (which can be wired or wireless), and the Internet, a wide area network, a local network, a metropolitan area network, etc. can be used.

[0170] The bus 802 can be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. The memory 801 is used to store programs, and the processor 800 executes the programs after receiving execution instructions. The load balancing method of the heterogeneous network disclosed in any one of the preceding embodiments of the present application can be applied to the processor 800 or implemented by the processor 800.

[0171] The processor 800 can be an integrated circuit chip with a processing capability of signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of hardware in the processor 800 or the instruction in the form of software. The processor 800 described above can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a ready programmable gate array (FPTA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. Each method, step and logic block disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as a hardware coding processor for execution, or a combination of hardware and software modules in the coding processor for execution. The software module can be located in a random access memory, a flash memory, a read only memory, a programmable read only memory or an electrically erasable programmable memory, a register, etc. The storage medium in the art. The storage medium is located in the memory 801, and the processor 800 reads the information in the memory 801, and combines the hardware to complete the steps of the above method.

[0172] The computer device provided by the embodiments of the present application and the load balancing method of the heterogeneous network provided by the embodiments of the present application have the same beneficial effects as the method adopted, run or implemented by the computer device.

[0173] The computer readable storage medium provided by the embodiments of the present application also provides a computer readable storage medium corresponding to the load balancing method of the heterogeneous network provided by the preceding embodiments. The computer readable storage medium is an optical disc, and a computer program (i.e. a computer program product) is stored on the optical disc. When the computer program is run by a processor, the load balancing method of the heterogeneous network provided by any of the preceding embodiments is executed.

[0174] It should be noted that examples of the computer readable storage medium can also include, but are not limited to, a phase change memory (PRAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), other types of random access memory (RAM), a read only memory (ROM), an electrically erasable programmable read only memory (EEPROM), a flash memory or other optical, magnetic storage medium, which will not be described one by one here.

[0175] The computer readable storage medium provided by the above embodiments of the present application and the load balancing method of the heterogeneous network provided by the embodiments of the present application have the same beneficial effects as the method adopted, run or implemented by the computer readable storage medium.

[0176] The embodiment of the present application further provides a computer program product, please refer to Figure 6 The computer program product 600 carries a program code, i.e. a computer program 601, which comprises instructions for executing the steps of the load balancing method of the heterogeneous network described in the above method embodiments, and details can be referred to the above method embodiments, which will not be repeated here.

[0177] The above computer program product can be specifically implemented by means of hardware, software or combination thereof. In an optional embodiment, the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (SDK) and the like.

[0178] The basic principles of the present application are described above in combination with specific embodiments, but it should be pointed out that the advantages, advantages, effects and the like mentioned in the present application are only examples and not limitations, and these advantages, advantages, effects and the like cannot be considered as the necessary possession of each embodiment of the present application. In addition, the above specific details are only for the purpose of example and for the purpose of understanding, and the above details do not limit the present application to the above specific details.

[0179] The block diagram of the device, apparatus, equipment, system involved in the present application is only an illustrative example and is not intended to require or imply the connection, arrangement and configuration shown in the block diagram. As those skilled in the art will recognize, these devices, apparatus, equipment, system can be connected, arranged and configured in any way. Words such as "include", "contain", "have" and the like are open-ended words, which mean "including but not limited to", and can be used interchangeably. The words "or" and "and" used herein mean the word "and / or", and can be used interchangeably unless the context clearly indicates otherwise. The word "such as" used herein means the phrase "such as but not limited to", and can be used interchangeably.

[0180] In addition, as used herein, "or" used in the list of items starting with "at least one" indicates separate listing, so that for example, the list of "at least one of A, B or C" means A or B or C, or AB or AC or BC, or ABC (i.e. A and B and C). In addition, the word "exemplary" does not mean that the described example is preferred or better than other examples.

[0181] It should also be noted that in the system and method of the present application, each component or step can be decomposed and / or recombined. These decompositions and / or recombination should be considered as equivalent solutions of the present application.

[0182] Various changes, modifications, and improvements in the technologies described herein can be made without departing from the teachings of the technology defined by the appended claims. Moreover, the scope of the claims of this application is not limited to the specific aspects described above. Rather, the aspects of the application include all alternatives, modifications, and equivalents falling within the scope of the claims below.

[0183] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0184] The above description has been presented for the purpose of illustration and description. Furthermore, this description is not intended to limit the embodiments of the application to the forms disclosed herein. Although several example aspects and embodiments have been discussed above, those of ordinary skill in the art will appreciate a variety of modifications, alternatives, permutations, additions, and sub-combinations of the described aspects and features.

Claims

1. A load balancing method of a heterogeneous network, characterized by, The method comprises: predicting, based on obtained user network history data of a user in a heterogeneous network cell, a traffic prediction result representing a traffic demand degree of the user in a future preset time; and predicting, based on obtained network attribute information of each network node in the heterogeneous network cell, a network perception evaluation result representing the quality of each network node; obtaining the load capacity of the network node based on the network perception evaluation result; selecting a target network node from each network node whose load capacity matches the traffic prediction result, and controlling the target network node to provide network resources for the user.

2. The load balancing method of a heterogeneous network according to claim 1, wherein, The method comprises: extracting a first group of feature data, a second group of feature data, and a third group of feature data from the user network history data; the first group of feature data is used to represent the traffic demand change rule of the user in a preset historical time range, the second group of feature data is used to represent the statistical characteristics of the traffic used by the user, and the third group of feature data is used to represent external factors affecting the traffic demand of the user; performing traffic trend prediction based on the first group of feature data, the second group of feature data, and the third group of feature data to predict the traffic prediction result.

3. The load balancing method of a heterogeneous network according to claim 1, wherein, The method comprises: in the case where there are multiple matching network nodes in each network node whose load capacity matches the traffic prediction result, obtaining the location of the user; from the multiple matching network nodes, filtering multiple nearest network nodes whose location differs from the location of the user by a distance threshold; determining the target network node from the multiple nearest network nodes.

4. The load balancing method of a heterogeneous network according to claim 3, wherein, The method comprises: obtaining a target nearest network node closest to the user from the multiple nearest network nodes; in the case where the idle resource amount of the target nearest network node is greater than or equal to the resource demand amount represented by the traffic prediction result, determining the target nearest network node as the target network node; in the case where the idle resource amount of the target nearest network node is less than the resource demand amount represented by the traffic prediction result, calculating a difference idle resource amount based on the idle resource amount and the resource demand amount, obtaining a remaining nearest network node whose idle resource amount is greater than or equal to the difference idle resource amount from the multiple nearest network nodes, and taking the target nearest network node and the remaining nearest network node as the target network node.

5. The load balancing method of a heterogeneous network according to claim 1, wherein, The method further comprises: in the case where there are multiple users in the heterogeneous network cell, determining the priority of each user based on the user type and user attribute included in the traffic prediction result of each user; the priority is used to represent the order of selecting a target network node for a user. Correspondingly, selecting a target network node from the network nodes, which has a load capacity matching the traffic prediction result, comprises: Selecting a target network node from the network nodes, which has a load capacity matching the traffic prediction result of each user based on the priority.

6. The load balancing method of a heterogeneous network according to claim 1, wherein, After controlling the target network node to provide network resources for the user, further comprising: Obtaining a current carrying value of the network node; the current carrying value is used to represent a proportion of currently occupied network resources of the network node; Determining an expected carrying value of each network node of the heterogeneous network cell based on the current carrying value and current carrying values of the remaining network nodes of the heterogeneous network cell; Migrating services on each network node of the heterogeneous network cell according to the expected carrying value.

7. The load balancing method of a heterogeneous network according to claim 6, wherein, Determining an expected carrying value of each network node of the heterogeneous network cell based on the current carrying value and current carrying values of the remaining network nodes of the heterogeneous network cell comprises: Based on the current carrying value and current carrying values of the remaining network nodes of the heterogeneous network cell, obtaining a first network node with a maximum carrying value and a second network node with a minimum carrying value from each network node of the heterogeneous network cell; Determining the expected carrying value of the first network node and the second network node as a mean value of the maximum carrying value and the minimum carrying value, and determining the expected carrying value of the network nodes other than the first network node and the second network node as the current carrying value.

8. A load balancing apparatus for a heterogeneous network, characterized by Comprise: A prediction module configured to predict a traffic prediction result representing a degree of traffic demand of a user in a future preset time based on obtained user network history data of the user in a heterogeneous network cell, and predict a network perception evaluation result representing advantages and disadvantages of network quality of each network node of the heterogeneous network cell based on obtained network attribute information of each network node of the heterogeneous network cell; An obtaining module configured to obtain a load capacity of the network node based on the network perception evaluation result; A selection module configured to select a target network node from the network nodes, which has a load capacity matching the traffic prediction result; A control module configured to control the target network node to provide network resources for the user.

9. A computer device comprising a memory, a processor, and a computer program stored on the memory, wherein the computer program comprises instructions that, when executed by the processor, cause the processor to perform the method of any one of claims 1-8. The processor executes the computer program to implement the method of any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the method of any one of claims 1-7.

11. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the method of any one of claims 1-7.

Images