Network bandwidth prediction method and device, computer device, readable storage medium and program product

Abstract

The application relates to a network bandwidth prediction method and device, computer equipment, a readable storage medium and a program product. The method comprises the following steps: obtaining a plurality of historical data of a target network, the plurality of historical data being related data corresponding to different sampling time points of the target network in a preset period, each historical data comprising total bandwidth information, bandwidth information corresponding to different application programs and bandwidth information corresponding to different users; performing data processing on the plurality of historical data to obtain feature data; inputting the feature data into a bandwidth prediction model which has been pre-trained to obtain a bandwidth prediction result output by the bandwidth prediction model. The method can improve the network bandwidth prediction accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of network security, and in particular to a network bandwidth prediction method and device, a computer device, a readable storage medium and a program product. BACKGROUND

[0002] With the rapid development of the Internet and communication technology, the rapid growth of broadband users and applications has driven the rapid growth of bandwidth demand. Users' demand for the network is increasing, and it has become the focus of broadband operators to fully utilize existing network resources and ensure high-quality online experience for most broadband users.

[0003] The bandwidth planning during bandwidth operation is basically a fixed available bandwidth determined after manual assessment, and most of the time, the bandwidth demand is ensured at the cost of a high redundant bandwidth. When the available bandwidth is used up due to sudden surge in bandwidth demand in special circumstances, network congestion will occur, resulting in poor user experience. If the bandwidth value and usage details at a future time can be predicted in advance, the available bandwidth can be flexibly configured according to specific needs to avoid high redundancy, and the prediction value can also be used to adjust specific management strategies to significantly improve user experience quality.

[0004] In the traditional technology, bandwidth prediction mainly relies on historical network overall bandwidth to predict a specific bandwidth at a future time. However, the network has various complex and variable applications, and the bandwidth changes over time, so the accuracy of the bandwidth predicted by only relying on the historical network overall bandwidth is low. SUMMARY

[0005] Therefore, it is necessary to provide a network bandwidth prediction method and device, a computer device, a readable storage medium and a program product capable of improving the accuracy of network bandwidth prediction.

[0006] In a first aspect, the present application provides a network bandwidth prediction method, which comprises: obtaining a plurality of historical data of a target network, the plurality of historical data being related data corresponding to different sampling time points of the target network in a preset period, each historical data comprising total bandwidth information, bandwidth information corresponding to different application programs and bandwidth information corresponding to different users; performing data processing on the plurality of historical data to obtain feature data; inputting the feature data into a bandwidth prediction model pre-trained to obtain a bandwidth prediction result output by the bandwidth prediction model.

[0007] In one of the embodiments, the historical data further comprises at least one of the total number of different types of terminals, the total number of concurrent users, the total number of concurrent connections, the number of connections corresponding to different application programs, the number of terminals used by different users using different types of terminals and state data corresponding to the sampling time point.

[0008] In one of the embodiments, the data processing is performed on the plurality of historical data to obtain the feature data, including: performing a corresponding type of data processing manner on each type of data in the plurality of historical data to obtain first feature data; the data processing manner includes logarithm, linear normalization, standard deviation normalization and one-hot encoding; and the feature data is obtained according to the first feature data.

[0009] In one of the embodiments, the feature data is obtained according to the first feature data, including: performing default value filling on the first feature data to obtain second feature data; and the second feature data is taken as the feature data.

[0010] In one of the embodiments, the default value filling is performed on the first feature data to obtain the second feature data, including: performing a corresponding type of default value filling manner on each type of data in the first feature data to obtain the second feature data; wherein the default value filling manner is determined according to the data processing manner corresponding to each type of data; and the default value filling manner includes observation value filling method, mean value filling method, median value filling method and mode value filling method.

[0011] In one of the embodiments, the method further includes: obtaining a plurality of sample historical data of a target network; each sample historical data includes sample total bandwidth information, sample bandwidth information corresponding to different application programs and sample bandwidth information corresponding to different users; performing data processing on the plurality of sample historical data to obtain sample feature data; and training a to-be-trained model according to the sample feature data to obtain a bandwidth prediction model after the training is completed.

[0012] In a second aspect, the application further provides a network bandwidth prediction device, including: an acquisition module, configured to acquire a plurality of historical data of a target network, the plurality of historical data being related data corresponding to different sampling time points of the target network in a preset period, and each historical data including total bandwidth information, bandwidth information corresponding to different application programs and bandwidth information corresponding to different users; a first determination module, configured to perform data processing on the plurality of historical data to obtain feature data; and a second determination module, configured to input the feature data into a bandwidth prediction model that is pre-trained to obtain a bandwidth prediction result output by the bandwidth prediction model.

[0013] In one of the embodiments, the historical data further includes at least one of total terminal quantity of different types of terminals, total concurrent user quantity, total concurrent connection quantity, connection quantity corresponding to different application programs, terminal quantity of different users using different types of terminals and state data corresponding to the sampling time point.

[0014] In one of the embodiments, the first determining module is specifically configured to process each type of data in the plurality of historical data by a corresponding type of data processing manner to obtain first feature data; the data processing manner comprises logarithm taking, linear normalization, standard deviation normalization and one-hot encoding; and the feature data is obtained according to the first feature data.

[0015] In one of the embodiments, the first determining module is specifically configured to perform default value filling on the first feature data to obtain second feature data; and the second feature data is taken as the feature data.

[0016] In one of the embodiments, the first determining module is specifically configured to perform default value filling on each type of data in the first feature data by a corresponding type of default value filling manner to obtain second feature data; wherein the default value filling manner is determined according to the data processing manner corresponding to each type of data; and the default value filling manner comprises observation value filling method, mean value filling method, median value filling method and mode value filling method.

[0017] In one of the embodiments, the apparatus further comprises a training module configured to obtain a plurality of sample historical data of a target network; each sample historical data comprises sample total bandwidth information, sample bandwidth information corresponding to different application programs and sample bandwidth information corresponding to different users; perform data processing on the plurality of sample historical data to obtain sample feature data; and train the to-be-trained model according to the sample feature data to obtain a bandwidth prediction model after the training is completed.

[0018] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method in any one of the first aspect when executing the computer program.

[0019] In a fourth aspect, the present application further provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of the method in any one of the first aspect.

[0020] In a fifth aspect, the present application further provides a computer program product comprising a computer program, wherein the computer program is executed by a processor to implement the steps of the method in any one of the first aspect.

[0021] The network bandwidth prediction method, device, computer device, readable storage medium and program product can obtain a plurality of historical data of a target network, the plurality of historical data are related data corresponding to different sampling time points of the target network in a preset period, each historical data includes total bandwidth information, bandwidth information corresponding to different application programs and bandwidth information corresponding to different users. Then, the plurality of historical data are processed to obtain feature data, and the feature data is input into a bandwidth prediction model that is pre-trained to obtain a bandwidth prediction result output by the bandwidth prediction model. Since the total bandwidth information, the bandwidth information corresponding to different application programs and the bandwidth information corresponding to different users are considered when the network bandwidth is predicted, the bandwidth usage in the target network can be more comprehensively, diversely and finely displayed, and therefore the accuracy of network bandwidth prediction can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the description of the embodiments of the present application or the related art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other related drawings can be obtained by those skilled in the art without creative labor.

[0023] Figure 1 The application environment diagram of the network bandwidth prediction method in an embodiment;

[0024] Figure 2 The flowchart of the network bandwidth prediction method in an embodiment;

[0025] Figure 3 The flowchart of the network bandwidth prediction method in another embodiment;

[0026] Figure 4 The flowchart of the network bandwidth prediction method in another embodiment;

[0027] Figure 5 The structural block diagram of the network bandwidth prediction device in an embodiment;

[0028] Figure 6 The internal structure diagram of the computer device in an embodiment. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0030] With the rapid development of the Internet and communication technology, the rapid growth of bandwidth demand is driven by the growth of broadband users and applications. Users' demand for the network is increasing, and it is the focus of broadband operators to fully utilize existing network resources and ensure high-quality online experience for most broadband users.

[0031] The bandwidth planning in bandwidth operation is basically a fixed available bandwidth determined after artificial evaluation, and most of the time, the cost of high redundancy bandwidth is sacrificed to ensure bandwidth demand. When the available bandwidth is used up due to sudden surge in bandwidth demand in special circumstances, network congestion will occur, resulting in poor user experience. If the bandwidth value and usage details at future time can be predicted in advance, the available bandwidth can be flexibly configured according to specific needs to avoid high redundancy, and the prediction value can also be used to adjust the specific management strategy to significantly improve the quality of user experience.

[0032] In the traditional technology, bandwidth prediction mainly relies on historical network overall bandwidth to predict a specific bandwidth at future time. However, various applications in the network are complex and changeable, and bandwidth changes over time. Therefore, the bandwidth obtained by prediction according to the historical network overall bandwidth is difficult to reflect the specific bandwidth change trend, and the accuracy of the specific bandwidth obtained by prediction is also low.

[0033] The network bandwidth prediction method provided by the embodiment of the present application can be applied to the application environment as shown in the Figure 1 The terminal 102 communicates with the server 104 through the network. The data storage system can store the data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on the cloud or other network servers. The server 104 obtains a plurality of historical data of a target network, the plurality of historical data are related data corresponding to different sampling time points of the target network in a preset period, each historical data includes total bandwidth information, bandwidth information corresponding to different application programs and bandwidth information corresponding to different users; the plurality of historical data are processed to obtain feature data; and the feature data is input into a bandwidth prediction model pre-trained to obtain a bandwidth prediction result output by the bandwidth prediction model.

[0034] The terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things device can be a smart speaker, a smart television, a smart air conditioner, a smart vehicle device, a projection device, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc. The head-mounted device can be a virtual reality (VR) device, an augmented reality (AR) device, smart glasses, etc.

[0035] The server 104 can be a stand-alone physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0036] In an exemplary embodiment, as shown in Figure 2 , a network bandwidth prediction method is provided, which is applied to the server in Figure 1 for example, and includes the following steps 201 to 203.

[0037] Among them:

[0038] Step 201, obtaining a plurality of historical data of a target network. The plurality of historical data are related data corresponding to different sampling time points of the target network in a preset period, and each historical data includes total bandwidth information, bandwidth information corresponding to different application programs, and bandwidth information corresponding to different users.

[0039] The target network can refer to a network in a cell area, or a network in a company area. The historical data of the target network can be obtained by deploying an application recognition function in the target network, such as identifying the category of traffic in the target network in real time by at least one of DPI (Deep Packet Inspection), DFI (Deep Flow Inspection), and intelligent recognition technology, i.e., identifying the application category generating the traffic, which includes at least one of video conference, video call, voice call, video, live broadcast, game, download, and e-commerce shopping.

[0040] The preset period can be a week or a month, etc., and adjacent sampling time points are separated by a preset time length, which can be one day or 12 hours, etc.

[0041] The bandwidth information corresponding to the application program refers to the bandwidth consumed by the application program running when used by different users in the target network; the bandwidth information corresponding to the user includes the bandwidth consumed by the user when using each application program, and the sum of the bandwidths consumed by the user when using all application programs; the total bandwidth information refers to the total bandwidth of the target network, which is the sum of the bandwidths in the bandwidth information corresponding to each application program, or the sum of the bandwidths in the bandwidth information corresponding to each user.

[0042] The bandwidth can be divided into uplink bandwidth and downlink bandwidth, and therefore, the total bandwidth information includes total uplink bandwidth and total downlink bandwidth; the bandwidth information corresponding to the application program includes uplink bandwidth corresponding to the application program and downlink bandwidth corresponding to the application program; and the bandwidth information corresponding to the user includes uplink bandwidth corresponding to the user and downlink bandwidth corresponding to the user. The total uplink bandwidth in the total bandwidth information can be the sum of the uplink bandwidths corresponding to all the application programs or the sum of the uplink bandwidths corresponding to all the users; and the total downlink bandwidth in the total bandwidth information can be the sum of the downlink bandwidths corresponding to all the application programs or the sum of the downlink bandwidths corresponding to all the users.

[0043] Optionally, a plurality of initial historical data (i.e., real-time historical data in the target network) of the target network in a preset period are acquired, the plurality of initial historical data are recorded and sorted according to an interval preset time length as one sampling time point, and a plurality of historical data are obtained.

[0044] In step 202, the plurality of historical data are processed to obtain feature data.

[0045] Optionally, the plurality of historical data are processed by using at least one processing mode of normalization processing, logarithm processing, one-hot encoding, and default value filling to obtain the feature data.

[0046] In step 203, the feature data are input into a bandwidth prediction model that is pre-trained to obtain a bandwidth prediction result output by the bandwidth prediction model.

[0047] The bandwidth prediction model is obtained by pre-training according to a large amount of sample historical data in the target network. The bandwidth prediction model can be a time series prediction model, which can be one of an ARIMA (Auto-Regressive Moving Average) model, a CNN (Convolutional Neural Networks) model, a GRU (Gate Recurrent Unit) model, a VAR (Value at Risk) model, an LSTM (Long Short-Term Memory) model, a Prophet model, a random forest, multivariate deep learning, a Transformer, and an MTS-Mixers (Multi-Time-Series Mixers) model, or a mixed model obtained by combining multiple models.

[0048] The bandwidth prediction result can include at least one of a total uplink bandwidth in the target network, a total downlink bandwidth in the target network, an uplink bandwidth corresponding to different application programs, a downlink bandwidth corresponding to different application programs, a total uplink bandwidth corresponding to different users, an uplink bandwidth corresponding to different users using different application programs, a downlink bandwidth corresponding to different users using different application programs, and the like at different time points in a future period of time. In other words, the bandwidth prediction result includes different types of bandwidth sequences predicted.

[0049] In a possible implementation, the bandwidth prediction model includes a plurality of input ports and output ports corresponding to the plurality of input ports, each input port is configured to input feature data of a corresponding type, and each output port is configured to output a bandwidth prediction result of a corresponding type. In other words, the bandwidth prediction model includes a plurality of trained sub-models, each trained sub-model is configured to make a prediction based on feature data of a corresponding type. For example, the bandwidth consumed by user A using a video call in a preset period of time is input to input port A, and output port A outputs the predicted bandwidth that will be consumed by user A using a video call in a future period of time; the bandwidth consumed by a video conference in a preset period of time is input to input port B, and output port B outputs the predicted bandwidth that will be consumed by a video conference in a future period of time.

[0050] In another possible implementation, the bandwidth prediction model includes one input port and one output port. In other words, the bandwidth prediction model includes one trained model, which can directly make a prediction based on feature data of different types.

[0051] The above bandwidth prediction method obtains a plurality of historical data of the target network, the plurality of historical data are related data corresponding to different sampling time points of the target network in a preset period of time, and each historical data includes total bandwidth information, bandwidth information corresponding to different application programs, and bandwidth information corresponding to different users. Then, the plurality of historical data is processed to obtain feature data, and the feature data is input into a pre-trained bandwidth prediction model to obtain a bandwidth prediction result output by the bandwidth prediction model. Since the total bandwidth information, the bandwidth information corresponding to different application programs, and the bandwidth information corresponding to different users are considered when predicting the network bandwidth, the bandwidth usage in the target network can be more comprehensively, diversely, and finely displayed, thereby improving the accuracy of network bandwidth prediction. In addition, the bandwidth prediction model is a time series prediction model, and the bandwidth prediction result includes different types of bandwidth sequences predicted, which can reflect the change trend of different types of bandwidth.

[0052] In one embodiment, the historical data further includes at least one of a total number of different types of terminals, a total number of concurrent users, a total number of concurrent connections, a number of connections corresponding to different application programs, a number of terminals corresponding to different users using different types of terminals, and state data corresponding to the sampling time point.

[0053] wherein the total number of different types of terminals is the sum of the number of different types of terminals used by all users in the target network. For example, the total number of laptops corresponds to the sum of the number of laptops used by all users in the target network.

[0054] The total number of concurrent users is the total number of people in the target network who are simultaneously generating network traffic. The total number of concurrent connections is the sum of the number of connections simultaneously established by the applications used in the target network. The number of connections refers to the number of connections established by the applications with other devices or software for transmitting data.

[0055] The application identification function can not only identify the category of network traffic, but also identify the type of terminal generating the network traffic, such as PC (desktop computer), laptop, tablet, smartphone, and smart appliance. Therefore, the type of terminal used by the user can also be obtained through the application identification function, so that the number of terminals of different types used by the user can be counted.

[0056] The state data corresponding to the sampling time point includes at least one of state data whether it is a weekday, state data whether it is a holiday, and state data whether it is a special date. The state data whether it is a weekday can further include at least one of state data whether it is a working time and state data whether it is a rest time; the special date is, for example, a stage of a sports event. The state data can be yes or no.

[0057] Therefore, based on the plurality of initial historical data, the plurality of historical data can be counted according to the following table 1, table 2, table 3 and table 4. Table 1 is counted from the overall level of the target network; table 2 is counted from the application level of the target network; table 3 is counted from the user level of the target network; and table 4 is counted from the sampling time point level.

[0058] Table 1

[0059]

[0060] Table 2

[0061]

[0062]

[0063] Table 3

[0064]

[0065]

[0066] Table 4

[0067]

[0068] In the embodiment, the total terminal quantity of different types of terminals, the total concurrent user quantity, the total concurrent connection quantity, the connection quantity corresponding to different application programs, the terminal quantity of different users using different types of terminals, and the state data corresponding to the sampling time point are further considered in the network bandwidth prediction, that is, the application group features, the user behavior features, the terminal features, and the sampling time point features are considered, so that the bandwidth prediction result obtained by prediction is more accurate, and more diversified bandwidth prediction is realized. The state data corresponding to the sampling time point is considered, that is, the feature that the network usage (that is, the bandwidth consumption) of users on weekdays, holidays, and special dates is different is considered, which can greatly improve the accuracy of bandwidth prediction.

[0069] In one of the embodiments, as shown in Figure 3 Another network bandwidth prediction method is provided. The plurality of historical data is processed to obtain feature data, including:

[0070] In step 301, each type of data in the plurality of historical data is processed by a corresponding type of data processing mode to obtain first feature data; the data processing mode includes logarithm, linear normalization, standard deviation normalization, and one-hot encoding.

[0071] The type of data refers to bandwidth type data, connection quantity type data, terminal quantity type data, user quantity type data, and state data type data. The bandwidth type data includes the total uplink bandwidth and the total downlink bandwidth in the total bandwidth information, the uplink bandwidth and the downlink bandwidth in the bandwidth information corresponding to each application program, and the uplink bandwidth and the downlink bandwidth in the bandwidth information corresponding to each user. The connection quantity type data includes the total concurrent connection quantity and the connection quantity corresponding to each application program. The terminal quantity type data includes the total terminal quantity of each type of terminal and the terminal quantity of each type of terminal used by each user. The user quantity type data includes the total concurrent user quantity. The state data type includes the state data of whether it is a weekday, the state data of whether it is a holiday, the state data of whether it is a special date, the state data of whether it is a working time, and the state data of whether it is a rest time.

[0072] For the bandwidth type data, the logarithmic data processing mode can be used to process it because the values of the crest and the trough can be extremely different.

[0073] For the connection number type data, the terminal number type data, and the user number type data, the data processing manner can be determined according to the scale of the target network. In an example, if the number of users in the target network is less than a preset number (for example, 1000), the connection number type data, the terminal number type data, and the user number type data can have a large difference between the peak value and the trough value. At this time, the logarithmic data processing manner can be used to process the connection number type data, the terminal number type data, and the user number type data respectively. If the number of users in the target network is less than the preset number, the linear normalization or the standard deviation normalization data processing manner can be used to process the connection number type data, the terminal number type data, and the user number type data respectively.

[0074] The formula of the linear normalization is as follows:

[0075]

[0076] wherein, x in is an original data value; x out is a normalized data value; and x0 is the maximum value supported by the dimension.

[0077] The formula of the standard deviation normalization is as follows:

[0078]

[0079] wherein, x in is an original data value; x out is a normalized data value; x mean is the mean value of the data of the same dimension; and x std is the standard deviation of the data of the same dimension.

[0080] For the state data type data, there are only two cases of yes and no, and therefore the one-hot encoding can be used to process the state data type data.

[0081] In step 302, the feature data is obtained according to the first feature data.

[0082] In a possible implementation, the first feature data is directly used as the feature data.

[0083] In another possible implementation, the first feature data is filled with default values to obtain second feature data, and the second feature data is used as the feature data. The first feature data can be filled with default values in a random manner or in a manner corresponding to the type of the data.

[0084] In this embodiment, the feature data obtained by processing each type of data in the plurality of historical data in a corresponding type of data processing manner can reduce the processing time of the bandwidth prediction model on the feature data, thereby improving the bandwidth prediction efficiency.

[0085] In one of the embodiments, as shown in Figure 4 Another bandwidth prediction method is provided. The first feature data is filled with default values to obtain second feature data; the second feature data is taken as feature data, including:

[0086] In step 401, each type of data in the first feature data is filled with a corresponding type of default value filling manner to obtain second feature data; wherein the default value filling manner is determined according to the data processing manner corresponding to each type of data; the default value filling manner includes observation value filling method, mean value filling method, median value filling method and mode value filling method.

[0087] In step 402, the second feature data is taken as feature data.

[0088] The observation value filling method refers to filling according to the mean value of the data of the time points before and after the default time point.

[0089] Optionally, each type of data in the first feature data is filled with a corresponding type of default value filling manner, which can be specifically that for each type of data, the default value filling manner corresponding to the type of data is determined according to the data processing manner of the type of data. For example, for data processed by a data processing manner of taking logarithm or linear normalization, the observation value filling method can be selected for filling; for data processed by a data processing manner of standard deviation normalization, any one of the mean value filling method, the median value filling method and the mode value filling method can be selected for filling. Preferably, for data processed by a data processing manner of standard deviation normalization, the mean value filling method is selected for filling.

[0090] In this embodiment, the feature data obtained by filling each type of data in the first feature data with a corresponding type of default value filling manner is more in line with the actual situation, and the bandwidth prediction result predicted by the bandwidth prediction model based on the feature data is more accurate.

[0091] In one of the embodiments, the method further includes:

[0092] In step 501, a plurality of sample historical data of a target network are obtained; each sample historical data includes sample total bandwidth information, sample bandwidth information corresponding to different application programs and sample bandwidth information corresponding to different users.

[0093] Optionally, the sample historical data can further include at least one of a sample total terminal quantity of different types of terminals, a sample total concurrent user quantity, a sample total concurrent connection quantity, a sample connection quantity corresponding to different application programs, a sample terminal quantity of different users using different types of terminals, and sample state data corresponding to a sampling time point.

[0094] The sample historical data can be large sample data, for example, initial historical data of the target network for several months, one year, or multiple years is obtained, and then the initial historical data is recorded and sorted according to an interval preset time length as a sampling time point, to obtain multiple sample historical data. The specific recording and sorting manner can refer to Tables 1-4.

[0095] At step 502, the multiple sample historical data is processed to obtain sample feature data.

[0096] The data processing process of the multiple sample historical data can refer to the data processing process of the multiple historical data, which will not be described here.

[0097] At step 503, the trained model is trained according to the sample feature data, and a bandwidth prediction model is obtained after the training is completed.

[0098] The trained model can be a time series prediction model, which can be one of an ARIMA model, a CNN, a GRU, a VAR model, an LSTM, a Prophet model, a random forest, multivariate deep learning, a Transformer, and an MTS-Mixers, or a hybrid model obtained by combining multiple models.

[0099] In one possible embodiment, the trained model includes multiple trained sub-models, and each trained sub-model is trained by using corresponding type of sample feature data. For example, a trained sub-model A is trained according to the bandwidth consumed by user A using a video phone at different time points; and a trained sub-model B is trained according to the bandwidth consumed by a video conference at different time points.

[0100] In another possible embodiment, the trained model is a complete model, which can be directly trained by using the sample feature data. During the training process, the results output by the model can be cross-validated to obtain an optimal bandwidth prediction model.

[0101] In addition, during the training process, the sample feature data can be divided into training data and test data. The training data is used to train the trained model, and the test data is used to test the trained bandwidth prediction model. When the bandwidth prediction model passes the test, it can be deployed in the target network to realize the network bandwidth prediction function of the target network.

[0102] It should be understood that although each step in the flowchart involved in the embodiments described above is shown in sequence according to the arrow, the steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, the execution of the steps is not strictly limited in sequence, and the steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in the embodiments described above can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of the steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0103] Based on the same inventive concept, the embodiments of the present application also provide a network bandwidth prediction device for implementing the network bandwidth prediction method described above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more network bandwidth prediction device embodiments provided below can refer to the limitations of the network bandwidth prediction method described above, which will not be repeated here.

[0104] In one exemplary embodiment, as shown in Figure 5 A network bandwidth prediction device is provided, and the network bandwidth prediction device 500 includes an acquisition module 501, a first determination module 502, and a second determination module 503, wherein:

[0105] The acquisition module 501 is configured to acquire a plurality of historical data of a target network, the plurality of historical data being related data corresponding to different sampling time points of the target network within a preset period, each historical data including total bandwidth information, bandwidth information corresponding to different application programs, and bandwidth information corresponding to different users.

[0106] The first determination module 502 is configured to perform data processing on the plurality of historical data to obtain feature data.

[0107] The second determination module 503 is configured to input the feature data into a bandwidth prediction model that has been pre-trained to obtain a bandwidth prediction result output by the bandwidth prediction model.

[0108] In one embodiment, the historical data further includes at least one of a total number of different types of terminals, a total number of concurrent users, a total number of concurrent connections, a number of connections corresponding to different application programs, a number of terminals of different types of terminals used by different users, and state data corresponding to the sampling time point.

[0109] In one of the embodiments, the first determining module 502 is specifically configured to process each type of data in the plurality of historical data by a corresponding type of data processing manner to obtain first feature data; the data processing manner includes logarithm, linear normalization, standard deviation normalization, and one-hot encoding; and the feature data is obtained according to the first feature data.

[0110] In one of the embodiments, the first determining module 502 is specifically configured to perform default value filling on the first feature data to obtain second feature data; and the second feature data is taken as the feature data.

[0111] In one of the embodiments, the first determining module 502 is specifically configured to perform a corresponding type of default value filling manner on each type of data in the first feature data to obtain second feature data; wherein the default value filling manner is determined according to the data processing manner corresponding to each type of data; and the default value filling manner includes an observed value filling method, a mean value filling method, a median value filling method, and a mode value filling method.

[0112] In one of the embodiments, the apparatus further includes a training module configured to obtain a plurality of sample historical data of a target network; each sample historical data includes sample total bandwidth information, sample bandwidth information corresponding to different application programs, and sample bandwidth information corresponding to different users; perform data processing on the plurality of sample historical data to obtain sample feature data; and train the to-be-trained model according to the sample feature data to obtain a bandwidth prediction model after the training is completed.

[0113] Each of the modules in the network bandwidth prediction apparatus can be realized by software, hardware, or a combination thereof, in whole or in part. Each of the modules can be embedded in or independent of a processor in a computer device in a hardware form, or stored in a memory in a computer device in a software form, so as to be called and executed by a processor to perform operations corresponding to each of the modules.

[0114] In one of the exemplary embodiments, a computer device is provided, which can be a server, and an internal structure diagram of the computer device can be as shown in FIG. 1. Figure 6As shown in the figure. The computer device includes a processor, a memory, an input / output interface (I / O for short) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store historical data, feature data, etc. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through network connection. The computer program is executed by the processor to implement a network bandwidth prediction method.

[0115] Those skilled in the art can understand that, Figure 6 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0116] In one exemplary embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of any of the method embodiments described above.

[0117] In one embodiment, a computer readable storage medium is provided, having a computer program stored thereon, and the computer program is executed by the processor to implement the steps of any of the method embodiments described above.

[0118] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by the processor to implement the steps of any of the method embodiments described above.

[0119] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0120] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. The non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. The volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, the RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., without being limited thereto.

[0121] The technical features of the above embodiments can be combined arbitrarily. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combinations of the technical features do not exist contradictory, they should be considered as the scope of the present application.

[0122] The above-described embodiments are merely illustrative of several embodiments of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the scope of the patent. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the scope of protection of the present application should be subject to the appended claims.

Claims

1. A network bandwidth prediction method, characterized in that, The method includes: Acquire multiple historical data of the target network. The multiple historical data are relevant data corresponding to different sampling time points of the target network within a preset period. Each of the historical data includes total bandwidth information, bandwidth information corresponding to different applications, and bandwidth information corresponding to different users. Data processing is performed on the multiple historical data sets to obtain feature data; The feature data is input into a pre-trained bandwidth prediction model to obtain the bandwidth prediction result output by the bandwidth prediction model.

2. The method according to claim 1, characterized in that, The historical data also includes at least one of the following: the total number of terminals of different types, the total number of concurrent users, the total number of concurrent connections, the number of connections corresponding to different applications, the number of terminals used by different users for different types of terminals, and the status data corresponding to the sampling time point.

3. The method according to claim 1 or 2, characterized in that, The process of processing the multiple historical data to obtain feature data includes: The data of each type in the multiple historical data are processed according to the corresponding data processing method to obtain the first feature data; the data processing method includes logarithmic transformation, linear normalization, standard deviation normalization and one-hot encoding; The feature data is obtained based on the first feature data.

4. The method according to claim 3, characterized in that, The step of obtaining the feature data based on the first feature data includes: The first feature data is filled with default values ​​to obtain the second feature data; The second feature data is used as the feature data.

5. The method according to claim 4, characterized in that, The step of filling the first feature data with default values ​​to obtain the second feature data includes: The second feature data is obtained by filling each type of data in the first feature data with the corresponding default value filling method; wherein, the default value filling method is determined according to the data processing method corresponding to each type of data; the default value filling method includes observation value filling method, mean filling method, median filling method and mode filling method.

6. The method according to claim 1, characterized in that, The method further includes: Obtain multiple historical sample data of the target network; each historical sample data includes total sample bandwidth information, sample bandwidth information corresponding to different applications, and sample bandwidth information corresponding to different users; Data processing is performed on the historical data of the multiple samples to obtain sample feature data; The model to be trained is trained based on the sample feature data, and the bandwidth prediction model is obtained after training is completed.

7. A network bandwidth prediction device, characterized in that, The device includes: The acquisition module is used to acquire multiple historical data of the target network. The multiple historical data are relevant data corresponding to different sampling time points of the target network within a preset period. Each of the historical data includes total bandwidth information, bandwidth information corresponding to different applications, and bandwidth information corresponding to different users. The first determining module is used to process the multiple historical data to obtain feature data; The second determining module is used to input the feature data into a pre-trained bandwidth prediction model to obtain the bandwidth prediction result output by the bandwidth prediction model.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images