Network resource adjustment method, apparatus and device, and readable storage medium
By obtaining client log information in multi-source consumption scenarios, network condition prediction and adaptive adjustments are performed, the problem that traditional evaluation methods cannot accurately evaluate network quality is solved, and resource optimization and user experience improvement are achieved.
Patent Information
- Application Number
- CN202510151870.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-11
- Publication Date
- 2025-05-06
AI Technical Summary
In multi-source consumption scenarios, traditional network transmission quality evaluation methods cannot accurately evaluate changes in network transmission quality, resulting in frequent service switching and excessive resource utilization, affecting user experience.
By obtaining the log information of the target client, building time series data, using the trained network condition prediction model to predict network conditions, adaptively adjusting the maximum network bandwidth of a single connection, and then adjusting resources based on the network parameters to be adjusted.
In the case of multi-source consumption, network parameters are accurately predicted based on logs, resource adjustments are optimized, user experience is improved, business data services are reduced, and network quality assessment and business data services are decoupled.
Smart Images
Figure CN119945919A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of computer application technology, and in particular to a network resource adjustment method, device, equipment and readable storage medium. Background Art
[0002] In traditional data transmission scenarios, in order to improve the user's QoE (Quality of Experience), a network detection component is deployed on the client or server to evaluate the quality of the current network transmission based on throughput and adjust service usage resources when the network quality changes.
[0003] Traditional network transmission quality assessment methods are only applicable to single-source consumption. When the client supports multi-source consumption, the data demand for a source in the multi-source changes dynamically, and the throughput of a single source cannot accurately assess the changes in the true network transmission quality. In addition, the throughput is lagging, and accurate throughput prediction is difficult to achieve, resulting in frequent business switching; the network environment is unclear, and in order not to affect data transmission, low-throughput connections still occupy more system resources.
[0004] In summary, how to improve user experience in data transmission scenarios and other issues are technical problems that technical personnel in this field urgently need to solve. Summary of the invention
[0005] The purpose of this application is to provide a network resource adjustment method, device, equipment and readable storage medium, so as to achieve the ability to accurately predict network parameters based on logs in the case of multi-source consumption, optimize the evaluation parameters of resource adjustment, and then improve user experience by effectively adjusting resources.
[0006] In order to solve the above technical problems, this application provides the following technical solutions:
[0007] A network resource adjustment method, applied to a peer-to-peer content distribution network architecture, comprising:
[0008] Obtaining log information corresponding to the target client, and using the log information to construct time series data; wherein the log information includes network transmission performance parameters and data request demand;
[0009] Using the trained network condition prediction model to perform network condition prediction processing on the time series data to obtain the maximum network bandwidth of a single connection of the target client;
[0010] Adaptively adjusting the single connection maximum network bandwidth using an adaptive controller to obtain a network parameter to be adjusted;
[0011] The network connection of the target client is resource adjusted based on the network parameter to be adjusted.
[0012] Preferably, obtaining log information corresponding to the target client and constructing time series data using the log information includes:
[0013] Obtaining the network transmission log and application layer request log of the target client;
[0014] Extracting the network transmission performance parameter from the network transmission log;
[0015] Extracting data request demand from the application layer request log;
[0016] Constructing a multi-dimensional vector feature based on network transmission performance parameters and data request requirements corresponding to the same network request identifier;
[0017] Performing hierarchical analysis on the multi-dimensional vector features to obtain a feature vector at each moment;
[0018] The feature vector at each moment and the corresponding network condition constitute the time series data.
[0019] Preferably, the multi-dimensional vector features are subjected to hierarchical analysis to obtain the feature vector at each moment, including:
[0020] Layering the multi-dimensional vector features;
[0021] Compare the features of the same layer pairwise and determine the weight values according to the scale table;
[0022] A pairwise comparison matrix is established based on the weight values, and the eigenvector is obtained using the pairwise comparison matrix.
[0023] Preferably, based on the network transmission performance parameters and data request demand corresponding to the same network request identifier, a multi-dimensional vector feature is constructed, including:
[0024] Find out the network transmission performance parameters and data request demand corresponding to the same network request identifier; wherein the network transmission performance parameters include: retransmission rate, packet loss rate, network delay, network jitter, and network congestion level;
[0025] The multi-dimensional vector feature is constructed based on the retransmission rate, packet loss rate, network delay, network jitter, network congestion level and data request demand corresponding to the same network request identifier.
[0026] Preferably, the adaptive controller is used to adaptively adjust the single connection maximum network bandwidth, and before obtaining the network parameters to be adjusted, the method includes:
[0027] According to the historical TCP connection data, a mapping relationship between the maximum network bandwidth of a single connection and the parameter to be adjusted is established;
[0028] Accordingly, the single connection maximum network bandwidth is adaptively adjusted using an adaptive controller to obtain network parameters to be adjusted, including:
[0029] Adaptively adjusting the single connection maximum network bandwidth by using an adaptive controller in combination with the mapping relationship to obtain a network parameter to be adjusted;
[0030] The network parameters to be adjusted include: TCP send buffer size, receive buffer size and network transmission bandwidth.
[0031] Preferably, performing resource adjustment on the network connection of the target client based on the network parameter to be adjusted includes:
[0032] generating a parameter change event using the network parameter to be adjusted;
[0033] The parameter change event is sent outwardly by way of subscription notification so as to adjust the corresponding network parameters and adjust the system resource occupancy of the target client.
[0034] Preferably, after adjusting the resources of the network connection of the target client based on the network parameter to be adjusted, the method further includes:
[0035] Monitor the network parameters of a single TCP connection to obtain the actual value of the network conditions;
[0036] Calculate the error between the single connection maximum network bandwidth and the actual value of the network condition;
[0037] The corresponding network parameters are adjusted according to the error.
[0038] A network resource adjustment device, comprising:
[0039] A log processing module, used to obtain log information corresponding to the target client and construct time series data using the log information; wherein the log information includes network transmission performance parameters and data request demand;
[0040] A network prediction module, used to perform network condition prediction processing on the time series data using a trained network condition prediction model to obtain a single-connection maximum network bandwidth of the target client;
[0041] A parameter determination module, used to adaptively adjust the single connection maximum network bandwidth using an adaptive controller to obtain a network parameter to be adjusted;
[0042] The resource adjustment module is used to adjust the resources of the network connection of the target client based on the network parameters to be adjusted.
[0043] An electronic device, comprising:
[0044] Memory for storing computer programs;
[0045] A processor is used to implement the steps of the above-mentioned network resource adjustment method when executing the computer program.
[0046] A readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned network resource adjustment method are implemented.
[0047] The method provided by the embodiment of the present application is applied in a peer-to-peer content distribution network architecture, including: obtaining log information corresponding to a target client, and constructing time series data using the log information; wherein the log information includes network transmission performance parameters and data request demand; performing network condition prediction processing on the time series data using a trained network condition prediction model to obtain a single-connection maximum network bandwidth of the target client; adaptively adjusting the single-connection maximum network bandwidth using an adaptive controller to obtain a network parameter to be adjusted; and adjusting resources on the network connection of the target client based on the network parameter to be adjusted.
[0048] In the application, the log information corresponding to the target client is first obtained in the point-to-point content distribution network architecture, and the time series data is constructed based on the log information, wherein the log information includes network transmission performance parameters and data request requirements. Then, the time series data can be processed for network condition prediction based on the trained network condition prediction model to obtain the maximum network bandwidth of the single connection of the target client. The maximum network bandwidth of the single connection is adaptively adjusted using adaptive control to obtain the network parameters to be adjusted, and finally the network connection of the target client can be adjusted based on the network parameters to be adjusted. It can be seen that in this application, it is only necessary to obtain the log information corresponding to the target customer and combine it with the network condition prediction model to realize network quality assessment. In other words, the business data service only needs to output the log, which can greatly reduce the participation of the business data service and realize the decoupling of network quality assessment from the business data service. Based on the log and combined with the network condition prediction model, the network condition prediction is completed, and the adaptive controller is used to adaptively adjust the maximum network bandwidth of the single connection to obtain the network parameters to be adjusted, thereby realizing resource adjustment. The complexity of the resource adjustment can also be greatly reduced. Since this application does not provide throughput to evaluate network quality, but determines whether the service is affected by the network transmission quality through log analysis, this application implements resource adjustment based on network quality in the case of multi-source consumption, thereby improving user experience.
[0049] Correspondingly, the embodiments of the present application also provide a network resource adjustment device, equipment and readable storage medium corresponding to the above-mentioned network resource adjustment method, which have the above-mentioned technical effects and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0051] Figure 1 This is a flowchart of an implementation method of a network resource adjustment method in an embodiment of the present application;
[0052] Figure 2 This is a schematic diagram of a peer-to-peer content distribution network architecture in an embodiment of the present application;
[0053] Figure 3 This is a schematic diagram of the structure of a network resource adjustment device in an embodiment of the present application;
[0054] Figure 4 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application;
[0055] Figure 5 This is a schematic diagram of the specific structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION
[0056] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.
[0057] Please refer to Figure 1 , Figure 1 This is a flow chart of a network resource adjustment method in an embodiment of the present application, which is applied to a point-to-point content distribution network architecture (PCDN). The method includes the following steps:
[0058] S101, obtaining log information corresponding to a target client, and constructing time series data using the log information.
[0059] The log information includes network transmission performance parameters and data request requirements.
[0060] PCDN (Peer-to-Peer Content Distribution Network) incorporates user devices (such as personal computers, mobile phones, etc.) into the content distribution network through P2P technology. Each user is not only a consumer of content, but also a provider of content. Therefore, there are multi-source consumption situations in the peer-to-peer content distribution network architecture. It should be noted that for multi-source consumption, traffic from different sources has different costs for service providers. By reasonably allocating data, the cost of service providers can be reduced without reducing service quality.
[0061] In this application, it is only necessary to obtain the log information corresponding to the target client, which includes network transmission performance parameters and data request requirements, and then use these log information to construct time series data.
[0062] In a specific implementation of the present application, obtaining log information corresponding to a target client and constructing time series data using the log information includes:
[0063] Step 1: Obtain the target client's network transmission log and application layer request log;
[0064] Step 2: extract network transmission performance parameters from the network transmission log;
[0065] Step 3: Extract data request demand from the application layer request log;
[0066] Step 4: construct a multi-dimensional vector feature based on the network transmission performance parameters and data request requirements corresponding to the same network request identifier;
[0067] Step 5: Perform hierarchical analysis on the multi-dimensional vector features to obtain the feature vector at each moment;
[0068] Step 6: Combine the feature vector at each moment and the corresponding network conditions into time series data.
[0069] For ease of description, the above steps are combined for explanation below.
[0070] Please refer to Figure 2 , the application layer request log can be obtained from the business data service end in the business server, and the network transmission log can be obtained from the network indicator monitoring component. Then, the network transmission performance parameters are extracted from the network transmission log, and the data request demand is extracted from the application layer request log. Since both the network transmission performance parameters and the data request demand have corresponding network request identifiers, a multi-dimensional vector feature can be constructed based on the network request identifier for the corresponding network transmission performance parameters and data request demand. That is, the multi-dimensional vector feature can represent the corresponding data request demand under a specific network transmission performance parameter.
[0071] In a specific implementation of the present application, the above step 4, based on the network transmission performance parameters and data request requirements corresponding to the same network request identifier, constructs a multi-dimensional vector feature, including:
[0072] Step 1: Find out the network transmission performance parameters and data request demand corresponding to the same network request identifier; wherein the network transmission performance parameters include: retransmission rate, packet loss rate, network delay, network jitter, and network congestion level;
[0073] Step 2: Construct a multi-dimensional vector feature based on the retransmission rate, packet loss rate, network delay, network jitter, network congestion level and data request demand corresponding to the same network request identifier.
[0074] When constructing multidimensional vector features, we first find out the retransmission rate, packet loss rate, network delay, network jitter, network congestion level and data request demand corresponding to the same network request identifier, and then construct a multidimensional vector feature based on the retransmission rate, packet loss rate, network delay, network jitter, network congestion level and data request demand.
[0075] After constructing the multi-dimensional vector features, the hierarchical analysis method can be used to perform hierarchical analysis on the multi-dimensional vector features to obtain the feature vector at each moment. That is, the feature vector can represent the corresponding retransmission rate, packet loss rate, network delay, network jitter, network congestion level and data request demand at each moment.
[0076] In a specific implementation of the present application, the above step 5 performs hierarchical analysis on the multi-dimensional vector features to obtain the feature vector at each moment, including:
[0077] Step 1: Layer the multi-dimensional vector features;
[0078] Step 2: Compare the features of the same layer in pairs and determine the weight values according to the scale table;
[0079] Step 3: Establish a pairwise comparison matrix based on the weight values, and use the pairwise comparison matrix to find the eigenvector.
[0080] Among them, the scale table can refer to Table 1:
[0081] Table 1. A ratio scale
[0082]
[0083] Specifically, the hierarchical analysis method is used to stratify the multi-dimensional vector features, compare the features of the same layer in pairs, and give weight values to the relative importance of each part according to Table 1. Then, a pairwise comparison matrix is established, and the eigenvector and eigenvalue are calculated. The eigenvector represents the priority of each part in each level. In practical applications, how to stratify, how to compare, how to establish a pairwise comparison matrix and the solution process can be specifically referred to the specific implementation process of the hierarchical analysis method, which will not be repeated here.
[0084] Then, the feature vector at each moment is combined with the corresponding network conditions to form time series data.
[0085] S102: Perform network condition prediction processing on the time series data using the trained network condition prediction model to obtain the maximum network bandwidth of a single connection of the target client.
[0086] Among them, the network condition prediction model can be a vector autoregression model (VAR), a vector error correction model (VECM) or a long short-term memory network model (LSTM).
[0087] In this embodiment, the network condition prediction model can be trained in advance using training samples, which are network conditions and their corresponding single-path maximum network bandwidths. The trained network condition prediction model can perform network condition prediction processing on the input network conditions, i.e., time series data, to obtain the single-connection maximum network bandwidth of the target client.
[0088] The maximum network bandwidth of a single connection may specifically be the maximum network bandwidth of a TCP connection.
[0089] S103: Use an adaptive controller to adaptively adjust the maximum network bandwidth of a single connection to obtain network parameters to be adjusted.
[0090] The adaptive controller may be a proportional-integral-derivative controller (PID) or a fuzzy controller.
[0091] TCP connection is the basis of network communication. Each TCP connection will occupy system resources, and the resources occupied by different application layer protocols are different. Due to client network conditions, client resource reading methods and other reasons, the resources occupied by TCP connection are not fully utilized, resulting in resource waste. Changes in client network conditions cause the maximum bandwidth that can be transmitted between the client and the server to change dynamically. In the scenario of audio and video transmission, the bit rate of the transmitted data needs to be adjusted according to the change in bandwidth to improve the user's playback experience.
[0092] The predicted value output by the client network condition prediction model is converted into TCP send buffer size, receive buffer size, network transmission bandwidth, etc. through the adaptive controller, and then the corresponding network parameters are adjusted based on the TCP send buffer size, receive buffer size and network transmission bandwidth to adjust the system resource occupancy.
[0093] In a specific implementation of the present application, the adaptive controller is used to perform adaptive adjustment processing on the maximum network bandwidth of a single connection, and before obtaining the network parameter to be adjusted, it includes: establishing a mapping relationship between the maximum network bandwidth of a single connection and the parameter to be adjusted according to the data of the historical TCP connection;
[0094] Accordingly, the adaptive controller is used to adaptively adjust the maximum network bandwidth of a single connection to obtain the network parameters to be adjusted, including:
[0095] The adaptive controller is used in combination with the mapping relationship to adaptively adjust the maximum network bandwidth of a single connection to obtain the network parameters to be adjusted; wherein the network parameters to be adjusted include: TCP sending buffer size, receiving buffer size and network transmission bandwidth.
[0096] Based on the historical TCP connection data, a mapping relationship is established between the predicted value output by the network condition prediction model (i.e., the maximum network bandwidth of a single connection) and the parameters to be adjusted (e.g., the relationship between the bandwidth-delay product (BDP) and the send and receive buffer sizes). Then, after obtaining the maximum network bandwidth of a single connection, the adaptive controller is used in combination with the mapping relationship to determine the TCP send buffer size, receive buffer size, and network transmission bandwidth.
[0097] S104: Adjust resources of the network connection of the target client based on the network parameters to be adjusted.
[0098] In a specific implementation of the present application, adjusting resources of a network connection of a target client based on a network parameter to be adjusted includes:
[0099] Generate a parameter change event using the network parameter to be adjusted;
[0100] Parameter change events are sent out through subscription notifications to adjust the corresponding network parameters and adjust the system resource usage of the target client.
[0101] That is to say, the predicted value output by the network condition prediction model is converted into TCP send buffer size, receive buffer size, network transmission bandwidth, etc. through the adaptive controller, and the parameter change event is sent out through the subscription notification method, so as to adjust the corresponding network parameters and adjust the system resource occupancy. The feedback mechanism is used to compare whether the actual network conditions before and after the adjustment have been improved, so as to realize the adaptive optimization adjustment strategy. For example, after obtaining the TCP send buffer size, receive buffer size, and network transmission bandwidth for resource adjustment, the network bandwidth of the largest single TCP connection measured after the adjustment is compared with the network bandwidth of the largest single TCP connection before the adjustment. If the value increases, it means that there is no problem with the adjustment strategy and no adjustment is required. If the value does not increase, it means that there is a problem with the adjustment strategy and it can be further optimized.
[0102] In a specific implementation of the present application, after adjusting the resources of the network connection of the target client based on the network parameters to be adjusted, the method further includes:
[0103] Monitor the network parameters of a single TCP connection to obtain the actual value of the network conditions;
[0104] Calculate the error between the maximum network bandwidth of a single connection and the actual value of the network conditions;
[0105] Adjust the corresponding network parameters according to the error.
[0106] That is to say, the error between the predicted value and the actual value can be calculated by monitoring the changes in the network indicators of a single TCP connection in real time, and the corresponding network parameters can be adjusted according to the error until the optimal state is reached.
[0107] The method provided by the embodiment of the present application is applied in a peer-to-peer content distribution network architecture, including: obtaining log information corresponding to a target client, and constructing time series data using the log information; wherein the log information includes network transmission performance parameters and data request demand; performing network condition prediction processing on the time series data using a trained network condition prediction model to obtain a single-connection maximum network bandwidth of the target client; adaptively adjusting the single-connection maximum network bandwidth using an adaptive controller to obtain a network parameter to be adjusted; and adjusting resources on the network connection of the target client based on the network parameter to be adjusted.
[0108] In the application, the log information corresponding to the target client is first obtained in the point-to-point content distribution network architecture, and the time series data is constructed based on the log information, wherein the log information includes network transmission performance parameters and data request requirements. Then, the time series data can be processed for network condition prediction based on the trained network condition prediction model to obtain the maximum network bandwidth of the single connection of the target client. The maximum network bandwidth of the single connection is adaptively adjusted using adaptive control to obtain the network parameters to be adjusted, and finally the network connection of the target client can be adjusted based on the network parameters to be adjusted. It can be seen that in this application, it is only necessary to obtain the log information corresponding to the target customer and combine it with the network condition prediction model to realize network quality assessment. In other words, the business data service only needs to output the log, which can greatly reduce the participation of the business data service and realize the decoupling of network quality assessment from the business data service. Based on the log and combined with the network condition prediction model, the network condition prediction is completed, and the adaptive controller is used to adaptively adjust the maximum network bandwidth of the single connection to obtain the network parameters to be adjusted, thereby realizing resource adjustment. The complexity of the resource adjustment can also be greatly reduced. Since this application does not provide throughput to evaluate network quality, but determines whether the service is affected by the network transmission quality through log analysis, this application implements resource adjustment based on network quality in the case of multi-source consumption, thereby improving user experience.
[0109] To facilitate those skilled in the art to better understand and implement the network resource adjustment method provided in the embodiment of the present application, the network resource adjustment method is described in detail below with reference to a specific scenario as an example.
[0110] Traditional network transmission quality assessment methods are only applicable to single-source consumption. When the client supports multi-source consumption, the data demand for a certain source changes dynamically, and the throughput of a single source cannot accurately assess the actual changes in network transmission quality. The network resource adjustment method provided in the embodiment of the present application determines whether the service is affected by the network transmission quality by assessing the changes in the client network conditions and combining the service demand.
[0111] The network resource adjustment method provided in the embodiment of the present application supports network transmission quality evaluation in P2SP (peer-to-peer and peer-to-server technology) scenarios.
[0112] like Figure 2 As shown, the implementation system of the network resource adjustment method provided in the embodiment of the present application can be deployed together with the business data service on the server side as a bypass service. The business server only needs to output the request log, and other work is completed independently by the system, which is decoupled from the business and has low complexity.
[0113] The implementation of the network resource adjustment method provided in the embodiment of the present application includes two main components: a client network condition prediction component and a network transmission parameter adaptive component. The client network condition prediction component outputs a predicted value of network condition changes by monitoring various indicators of network transmission and combining the service request log; the network transmission parameter adaptive component automatically adjusts network transmission related parameters after the predicted value of the client network condition changes, reclaims system resources of idle connections, and allocates system resources to connections with increasing demand.
[0114] Among them, regarding the prediction of client network conditions: through the joint analysis of network transmission logs and application layer request logs, the client's request behavior can be clarified and the changing trend of its network conditions can be predicted. Collect network transmission logs and application layer request logs to obtain the network transmission performance indicators of client requests and the magnitude of client data requests, process data anomalies and missing values and standardize the data indicators, and associate network transmission data and application layer requests through network request identifiers, such as TCP four-tuples. Extract the features reflecting the network status from the associated log information, such as retransmission rate, packet loss rate, network delay, network jitter, network congestion level and client request demand, and construct a multi-dimensional vector feature. For example, the client request demand corresponds to the TCP four-tuple, and the retransmission rate, packet loss rate, network delay, network jitter, and network congestion level extracted from the log information also correspond to the TCP four-tuple. When establishing the association, the retransmission rate, packet loss rate, network delay, network jitter, network congestion level and client request demand corresponding to the same group of source TCP four-tuples can be summarized together to construct a multi-dimensional vector feature.
[0115] Then, the hierarchical analysis method is used to stratify the multidimensional vector features, compare the features of the same layer in pairs, and give weight values to the relative importance of each part according to Table 1. Then, a pairwise comparison matrix is established, and the eigenvector and eigenvalue are obtained. The eigenvector represents the priority of each part (element) in each level. The hierarchical analysis method is to decompose the decision-making problem into different hierarchical structures according to the order of the overall goal, the sub-goals of each layer, the evaluation criteria, and the specific alternative investment plan. Then, the method of solving the eigenvector of the judgment matrix is used to obtain the priority weight of each element of each level to an element of the previous level. Finally, the weighted sum method is used to recursively merge the final weight of each alternative plan to the overall goal. The one with the largest final weight is the optimal plan.
[0116] The characteristic index data collected is used to calculate the characteristic vector of each moment according to the hierarchical analysis method, and the characteristic vector and the corresponding target variable (client network condition) are combined into a time series data set. According to the characteristics of the data and the prediction target, a suitable model is selected, and the model parameters are obtained by training the model using historical data. Time series data such as real-time retransmission rate, packet loss rate, network delay, network jitter, network congestion level, and client request demand are collected, and the trained model is used to predict the client's future network conditions.
[0117] Regarding network transmission parameter adaptation: Changes in client network conditions cause the maximum bandwidth that can be transmitted between the client and the server to change dynamically. In the scenario of audio and video transmission, it is necessary to adjust the bit rate of the transmitted data according to the change in bandwidth to improve the user's playback experience. The predicted value output by the client network condition prediction model is converted into TCP send buffer size, receive buffer size, network transmission bandwidth, etc. through the adaptive controller, and parameter change events are sent out through subscription notifications. Then, by adjusting the corresponding network parameters, the system resource occupancy allocation is adjusted. The adjusted parameters are compared with the actual network conditions through the feedback mechanism to achieve adaptive optimization and adjustment strategies.
[0118] First, based on the data of historical TCP connections, a mapping relationship between the predicted value output by the network condition prediction model and the parameters to be adjusted is established (for example, the relationship between the bandwidth delay product (BDP) and the send and receive buffer sizes). Second, the error between the predicted value and the actual value is calculated by real-time monitoring of the changes in the network indicators of a single TCP connection, and the corresponding network parameters are adjusted according to the error until the optimal state is achieved.
[0119] That is to say, in this application, by constructing a client network condition prediction model and an adaptive controller to adjust the optimal sending buffer size, receiving buffer size and connection network bandwidth corresponding to different network conditions, it is possible to avoid the situation where data packet loss affects the sending efficiency due to the sending and receiving buffer being too small or the system memory being wasted due to being too large, thereby improving the utilization of system resources and reducing service costs.
[0120] Specifically, the present application controls the sending and receiving buffers through the client network condition prediction model, reduces the connection's occupation of system resources, and reduces service costs; outputs the maximum network bandwidth of a single connection through the client network condition prediction model, and provides signals for adjusting parameters based on bandwidth for other scenarios; the core components of the system are independently deployed as a bypass system, and the application layer only needs to output log access and is transparent to the client, with low implementation complexity.
[0121] Corresponding to the above method embodiment, the embodiment of the present application further provides a network resource adjustment device. The network resource adjustment device described below and the network resource adjustment method described above can refer to each other.
[0122] See also Figure 3 As shown, the device includes the following modules:
[0123] The log processing module 101 is used to obtain log information corresponding to the target client and construct time series data using the log information; wherein the log information includes network transmission performance parameters and data request demand;
[0124] The network prediction module 102 is used to perform network condition prediction processing on the time series data using the trained network condition prediction model to obtain the maximum network bandwidth of a single connection of the target client;
[0125] The parameter determination module 103 is used to adaptively adjust the maximum network bandwidth of a single connection by using an adaptive controller to obtain a network parameter to be adjusted;
[0126] The resource adjustment module 104 is used to adjust the resources of the network connection of the target client based on the network parameters to be adjusted.
[0127] The device provided by the embodiment of the present application is applied in a peer-to-peer content distribution network architecture, including: obtaining log information corresponding to a target client, and constructing time series data using the log information; wherein the log information includes network transmission performance parameters and data request demand; performing network condition prediction processing on the time series data using a trained network condition prediction model to obtain a single-connection maximum network bandwidth of the target client; using an adaptive controller to adaptively adjust the single-connection maximum network bandwidth to obtain a network parameter to be adjusted; and adjusting resources of the network connection of the target client based on the network parameter to be adjusted.
[0128] In the application, the log information corresponding to the target client is first obtained in the point-to-point content distribution network architecture, and the time series data is constructed based on the log information, wherein the log information includes network transmission performance parameters and data request requirements. Then, the time series data can be processed for network condition prediction based on the trained network condition prediction model to obtain the maximum network bandwidth of the single connection of the target client. The maximum network bandwidth of the single connection is adaptively adjusted using adaptive control to obtain the network parameters to be adjusted, and finally the network connection of the target client can be adjusted based on the network parameters to be adjusted. It can be seen that in this application, it is only necessary to obtain the log information corresponding to the target customer and combine it with the network condition prediction model to realize network quality assessment. In other words, the business data service only needs to output the log, which can greatly reduce the participation of the business data service and realize the decoupling of network quality assessment from the business data service. Based on the log and combined with the network condition prediction model, the network condition prediction is completed, and the adaptive controller is used to adaptively adjust the maximum network bandwidth of the single connection to obtain the network parameters to be adjusted, thereby realizing resource adjustment. The complexity of the resource adjustment can also be greatly reduced. Since this application does not provide throughput to evaluate network quality, but determines whether the service is affected by the network transmission quality through log analysis, this application implements resource adjustment based on network quality in the case of multi-source consumption, thereby improving user experience.
[0129] In a specific implementation of the present application, the network prediction module is specifically used to obtain the network transmission log and application layer request log of the target client;
[0130] Extract network transmission performance parameters from network transmission logs;
[0131] Extract data request demand from application layer request logs;
[0132] Constructing a multi-dimensional vector feature based on network transmission performance parameters and data request requirements corresponding to the same network request identifier;
[0133] Perform hierarchical analysis on multi-dimensional vector features to obtain the feature vector at each moment;
[0134] The feature vector at each moment and the corresponding network conditions are combined into time series data.
[0135] In a specific implementation of the present application, the network prediction module is specifically used to stratify the multi-dimensional vector features;
[0136] Compare the features of the same layer pairwise and determine the weight values according to the scale table;
[0137] A pairwise comparison matrix is established based on the weight values, and the eigenvector is obtained using the pairwise comparison matrix.
[0138] In a specific implementation of the present application, the network prediction module is specifically used to find out the network transmission performance parameters and data request demand corresponding to the same network request identifier; wherein the network transmission performance parameters include: retransmission rate, packet loss rate, network delay, network jitter, and network congestion level;
[0139] A multi-dimensional vector feature is constructed based on the retransmission rate, packet loss rate, network delay, network jitter, network congestion level and data request demand corresponding to the same network request identifier.
[0140] In a specific implementation of the present application, it also includes:
[0141] A mapping establishment module is used to use an adaptive controller to perform adaptive adjustment processing on the maximum network bandwidth of a single connection, and before obtaining the network parameters to be adjusted, establish a mapping relationship between the maximum network bandwidth of a single connection and the parameters to be adjusted based on historical TCP connection data;
[0142] Accordingly, the parameter determination module is specifically used to adaptively adjust the maximum network bandwidth of a single connection by using an adaptive controller in combination with a mapping relationship to obtain a network parameter to be adjusted;
[0143] The network parameters to be adjusted include: TCP send buffer size, receive buffer size and network transmission bandwidth.
[0144] In a specific implementation of the present application, the resource adjustment module is specifically used to generate a parameter change event using the network parameters to be adjusted;
[0145] Parameter change events are sent out through subscription notifications to adjust the corresponding network parameters and adjust the system resource usage of the target client.
[0146] In a specific implementation of the present application, it also includes:
[0147] A feedback adjustment module is used to monitor the network parameters of a single TCP connection to obtain the actual value of the network condition after adjusting the resources of the network connection of the target client based on the network parameters to be adjusted;
[0148] Calculate the error between the maximum network bandwidth of a single connection and the actual value of the network conditions;
[0149] Adjust the corresponding network parameters according to the error.
[0150] Corresponding to the above method embodiment, an embodiment of the present application further provides an electronic device, and the electronic device described below and the network resource adjustment method described above can refer to each other.
[0151] See also Figure 4 As shown, the electronic device includes:
[0152] A memory 332, for storing computer programs;
[0153] The processor 322 is configured to implement the steps of the network resource adjustment method of the above method embodiment when executing a computer program.
[0154] For details, please refer to Figure 5 , Figure 5 A schematic diagram of the specific structure of an electronic device provided for this embodiment, which may have relatively large differences due to different configurations or performances, may include one or more processors (central processing units, CPU) (for example, one or more processors) and a memory 332, and the memory 332 stores one or more computer programs 342 or data 344. Among them, the memory 332 can be a temporary storage or a permanent storage. The program stored in the memory 332 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the data processing device. Furthermore, the processor 322 can be configured to communicate with the memory 332 to execute a series of instruction operations in the memory 332 on the electronic device 301.
[0155] The electronic device 301 may further include one or more power supplies 326 , one or more wired or wireless network interfaces 350 , one or more input and output interfaces 358 , and / or one or more operating systems 341 .
[0156] The steps in the network resource adjustment method described above can be implemented by the structure of an electronic device.
[0157] Corresponding to the above method embodiment, the embodiment of the present application further provides a readable storage medium. The readable storage medium described below and the network resource adjustment method described above can refer to each other.
[0158] A readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the network resource adjustment method of the above method embodiment are implemented.
[0159] The readable storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other readable storage medium that can store program codes.
[0160] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0161] Those skilled in the art may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0162] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
[0163] Finally, it should be noted that, in this article, relationships such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms include, include or any other variations are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0164] Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of the present application. At the same time, for those skilled in the art, according to the idea of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
Claims
1. A network resource adjustment method, characterized in that: Applied to peer-to-peer content distribution network architecture, including: Obtaining log information corresponding to the target client, and using the log information to construct time series data; wherein the log information includes network transmission performance parameters and data request demand; Using the trained network condition prediction model to perform network condition prediction processing on the time series data to obtain the maximum network bandwidth of a single connection of the target client; Adaptively adjusting the single connection maximum network bandwidth using an adaptive controller to obtain a network parameter to be adjusted; The network connection of the target client is resource adjusted based on the network parameter to be adjusted.
2. The method according to claim 1, characterized in that Obtaining log information corresponding to the target client and using the log information to construct time series data, including: Obtaining the network transmission log and application layer request log of the target client; Extracting the network transmission performance parameter from the network transmission log; Extracting data request demand from the application layer request log; Constructing a multi-dimensional vector feature based on network transmission performance parameters and data request requirements corresponding to the same network request identifier; Performing hierarchical analysis on the multi-dimensional vector features to obtain a feature vector at each moment; The feature vector at each moment and the corresponding network condition constitute the time series data.
3. The method according to claim 2, characterized in that The multi-dimensional vector features are analyzed hierarchically to obtain the feature vector at each moment, including: Layering the multi-dimensional vector features; Compare the features of the same layer pairwise and determine the weight values according to the scale table; A pairwise comparison matrix is established based on the weight values, and the eigenvector is obtained using the pairwise comparison matrix.
4. The method according to claim 2, characterized in that: Based on the network transmission performance parameters and data request requirements corresponding to the same network request identifier, a multi-dimensional vector feature is constructed, including: Find out the network transmission performance parameters and data request demand corresponding to the same network request identifier; wherein the network transmission performance parameters include: retransmission rate, packet loss rate, network delay, network jitter, and network congestion level; The multi-dimensional vector feature is constructed based on the retransmission rate, packet loss rate, network delay, network jitter, network congestion level and data request demand corresponding to the same network request identifier.
5. The method according to claim 1, characterized in that: Before the adaptive controller is used to adaptively adjust the single connection maximum network bandwidth to obtain the network parameters to be adjusted, the method includes: Based on the historical TCP connection data, a mapping relationship between the maximum network bandwidth of a single connection and the parameter to be adjusted is established; Accordingly, the single connection maximum network bandwidth is adaptively adjusted using an adaptive controller to obtain network parameters to be adjusted, including: Adaptively adjusting the single connection maximum network bandwidth by using an adaptive controller in combination with the mapping relationship to obtain a network parameter to be adjusted; The network parameters to be adjusted include: TCP send buffer size, receive buffer size and network transmission bandwidth.
6. The method according to claim 1, characterized in that The method further comprises: adjusting resources of the network connection of the target client based on the network parameter to be adjusted, comprising: Generating a parameter change event using the network parameter to be adjusted; The parameter change event is sent outwardly by way of subscription notification so as to adjust the corresponding network parameters and adjust the system resource occupancy of the target client.
7. The method according to any one of claims 1 to 6, characterized in that: After adjusting the resources of the network connection of the target client based on the network parameter to be adjusted, the method further includes: Monitor the network parameters of a single TCP connection to obtain the actual value of the network conditions; Calculate the error between the single connection maximum network bandwidth and the actual value of the network condition; The corresponding network parameters are adjusted according to the error.
8. A network resource adjustment device, characterized in that: include: A log processing module, used to obtain log information corresponding to the target client and construct time series data using the log information; wherein the log information includes network transmission performance parameters and data request demand; A network prediction module, used to perform network condition prediction processing on the time series data using a trained network condition prediction model to obtain a single-connection maximum network bandwidth of the target client; A parameter determination module, used to adaptively adjust the single connection maximum network bandwidth using an adaptive controller to obtain a network parameter to be adjusted; The resource adjustment module is used to adjust the resources of the network connection of the target client based on the network parameters to be adjusted.
9. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the network resource adjustment method according to any one of claims 1 to 7 when executing the computer program.
10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the network resource adjustment method according to any one of claims 1 to 7 are implemented.