CN117354072B - Automatic charging adjustment method and device for data center network bandwidth flow - Google Patents

Abstract

The application provides a method and a device for automatically charging and adjusting network bandwidth flow of a data center, which are characterized in that historical flow charging data of a user are collected and preprocessed to obtain a plurality of flow data subsequences, flow data in the flow data subsequences are used for determining accumulation residual errors, further accumulation likelihood coefficients are determined, the flow data subsequences are subjected to characteristic reconstruction, a flow characteristic space matrix is determined by the reconstructed flow data subsequences, further a flow characteristic covariance matrix is obtained, a flow characteristic value is determined by the flow characteristic covariance matrix, a flow charging variation index of the flow data subsequences is determined according to the flow characteristic value, a flow charging prediction function value of the flow data subsequences is determined according to the accumulation likelihood coefficients and the flow charging variation index, and the network bandwidth time-consuming by a bandwidth flowmeter of a flow transmission link is adjusted according to the flow charging prediction function value, so that the network stability under different network environments is improved, and the accuracy of charging results is ensured.

Description

A data center network bandwidth flow automatic billing adjustment method and device

Technical Field

The present application relates to the technical field of data centers, and more specifically, to a method and device for automatically billing and adjusting network bandwidth flow in a data center.

Background technique

A data center is a specially designed and configured facility used to centrally store, process and manage a large number of computer servers, network equipment, storage systems and related infrastructure to provide computing, storage and network services. Traffic billing for bandwidth used in a data center is a method of billing based on peak bandwidth and traffic usage.

However, in the existing traffic billing methods, since users have different network usage habits and their traffic usage patterns are also different, the billing results are uncertain. In addition, the peak value of network traffic in traditional billing methods only considers the traffic changes in a short period of time, and does not consider long-term usage. This will cause the billing results to be quite different from the actual network usage. Therefore, how to improve the accuracy of billing results is an issue that the industry urgently needs to solve.

Summary of the invention

The present application provides a method and device for automatically adjusting the billing of data center network bandwidth traffic to improve the accuracy of billing results.

In order to solve the above technical problems, this application adopts the following technical solutions:

In a first aspect, the present application provides a method for automatically charging and adjusting bandwidth flow in a data center network, comprising the following steps:

Collect historical user traffic billing data from the data center and pre-process it to obtain multiple traffic data subsequences;

Determine an aggregated residual from the flow data in each flow data subsequence, obtain a re-aggregation factor for each flow data subsequence according to the aggregated residual, and then determine an aggregated likelihood coefficient corresponding to each flow data subsequence respectively through the re-aggregation factor;

For each of the multiple traffic data subsequences, feature reconstruction is performed on the traffic data subsequence, and the traffic data in the reconstructed traffic data subsequence is mapped to a spatial matrix to obtain a traffic feature spatial matrix;

Determine a traffic feature covariance matrix according to the traffic feature space matrix, determine the traffic feature value of the traffic data subsequence according to the traffic feature covariance matrix, determine the traffic billing iterative index of the traffic data subsequence according to the traffic feature value, and then determine the traffic billing iterative index corresponding to each traffic data subsequence;

According to the accumulated likelihood coefficient and flow billing iteration index corresponding to each flow data subsequence, the flow billing prediction function value corresponding to each flow data subsequence is determined, and the flow billing prediction function value of each flow data subsequence is superimposed to obtain the flow billing prediction function value of the flow data sequence;

The network bandwidth during bandwidth flow billing of the traffic transmission link is adjusted according to the flow billing prediction function value.

In some embodiments, collecting historical traffic billing data of users in a data center and preprocessing the data to obtain multiple traffic data subsequences specifically includes:

Collecting historical traffic billing data of users in the data center, and dividing the historical traffic billing data of users into blocks to obtain a traffic data sequence;

Smoothing the flow data sequence to obtain a flow data smoothing sequence;

The flow data smoothing sequence is decomposed according to the user flow usage frequency to obtain a plurality of flow data subsequences.

In some embodiments, determining the accumulated residual from the flow data in each flow data subsequence specifically includes:

Determine the mean value of the traffic data corresponding to each traffic data subsequence ;

The accumulated residual of each flow data subsequence is determined according to the flow data mean, wherein the accumulated residual is determined according to the following formula:

in, Indicates the first/> The accumulated residual of the subsequences of traffic data, /> Indicates the first/> The first/> in the flow data subsequence flow data values, /> Represents the number of flow data values in a single flow data subsequence.

In some embodiments, obtaining the re-aggregation factor of each flow data subsequence according to the aggregated residual specifically includes:

Get the accumulated residual of each flow data subsequence ;

Determine the accumulation heterogeneity value corresponding to each flow data subsequence ;

The re-aggregation factor of each flow data subsequence is determined by the aggregation residual and the aggregation heterogeneity value, wherein the re-aggregation factor is determined according to the following formula:

Indicates the first/> The re-aggregation factor of the traffic data subsequences, /> Indicates the first/> The first flow data values, /> Represents the number of flow data values in a single flow data subsequence.

In some embodiments, determining the accumulation likelihood coefficient of each traffic data subsequence by the re-aggregation factor specifically includes:

Obtain the re-aggregation factor of each flow data subsequence;

determining an average reaggregation factor for each subsequence of flow data;

determining a re-aggregation attenuation factor for each flow data subsequence;

An aggregation likelihood coefficient is determined according to the re-aggregation factor, the average re-aggregation factor and the re-aggregation attenuation factor.

In some embodiments, determining the traffic feature covariance matrix according to the traffic feature space matrix specifically includes:

Centralize the traffic feature space matrix;

Determine the covariance of the flow data values in the flow feature space matrix after the centralization process;

The covariance of all traffic data values in the traffic feature space matrix is filled into the corresponding positions of the matrix in chronological order to obtain the traffic feature covariance matrix.

In some embodiments, determining the traffic billing iteration index of the traffic data subsequence according to the traffic characteristic value specifically includes:

Obtaining the flow characteristic value of the flow data subsequence;

Determine the mean value of the flow data in the flow data subsequence;

The flow billing iterative index of the flow data subsequence is determined according to the flow characteristic value and the flow data mean.

In a second aspect, the present application provides a data center network bandwidth flow automatic billing adjustment device, which includes a bandwidth control unit, and the bandwidth control unit includes:

A flow data subsequence determination module is used to collect historical flow billing data of users in a data center, divide the historical flow billing data of users into blocks to obtain a flow data sequence, and then smooth the flow data sequence to obtain a flow data smoothed sequence, and decompose the flow data smoothed sequence according to the user flow usage frequency to obtain multiple flow data subsequences;

An accumulation likelihood coefficient determination module, used to determine an accumulation residual from the flow data in each flow data subsequence, obtain a re-aggregation factor of each flow data subsequence according to the accumulation residual, and then determine the accumulation likelihood coefficient of each flow data subsequence through the re-aggregation factor;

A traffic feature space matrix determination module is used to reconstruct each traffic data subsequence in a plurality of traffic data subsequences according to the traffic feature, and map the traffic data in the reconstructed traffic data subsequence to the space matrix to obtain the traffic feature space matrix;

A flow billing iterative index determination module is used to determine a flow feature covariance matrix according to the flow feature space matrix, determine the flow feature value of the flow data subsequence by the flow feature covariance matrix, determine the flow billing iterative index of the flow data subsequence according to the flow feature value, and further determine the flow billing iterative index corresponding to each flow data subsequence;

The flow billing prediction function value determination module is used to determine the flow billing prediction function value corresponding to each flow data subsequence according to the accumulation likelihood coefficient and flow billing iteration index corresponding to each flow data subsequence, and superimpose the flow billing prediction function values of each flow data subsequence to obtain the flow billing prediction function value of the flow data sequence;

The bandwidth control module adjusts the network bandwidth of the traffic transmission link during bandwidth flow billing according to the traffic billing prediction function value.

In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned data center network bandwidth traffic automatic billing adjustment method.

In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned data center network bandwidth traffic automatic billing and adjustment method is implemented.

The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

In a data center network bandwidth flow automatic billing and adjustment method and device provided by the present application, historical flow billing data of users in the data center is first collected and preprocessed to obtain multiple flow data subsequences, and the accumulation residual is determined from the flow data in each flow data subsequence. The re-aggregation factor of each flow data subsequence is obtained according to the accumulation residual, and then the accumulation likelihood coefficient corresponding to each flow data subsequence is determined according to the re-aggregation factor. For each flow data subsequence in the multiple flow data subsequences, the flow data subsequence is feature reconstructed, and the flow data in the reconstructed flow data subsequence is mapped to a space matrix to obtain a flow feature space matrix, and the flow feature covariance moment is determined according to the flow feature space matrix. The flow characteristic value of each flow data subsequence is determined by the flow characteristic covariance matrix, and the flow billing iteration index of the flow data subsequence is determined according to the flow characteristic value, and then the flow billing iteration index corresponding to each flow data subsequence is determined, and the flow billing prediction function value corresponding to each flow data subsequence is determined according to the accumulation likelihood coefficient and the flow billing iteration index corresponding to each flow data subsequence, and the flow billing prediction function value of each flow data subsequence is superimposed to obtain the flow billing prediction function value of the flow data sequence, and the network bandwidth of the flow transmission link is adjusted according to the flow billing prediction function value during bandwidth flow billing, thereby improving the network stability under different network environments and ensuring the accuracy of the billing results.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is an exemplary flow chart of a method for automatically charging and adjusting bandwidth flow in a data center network according to some embodiments of the present application;

FIG2 is a schematic diagram of exemplary hardware and/or software of a bandwidth control unit according to some embodiments of the present application;

FIG3 is a schematic diagram of the structure of a computer device that applies a method for automatically charging and adjusting bandwidth flow in a data center network as shown in some embodiments of the present application.

Detailed ways

The core of the present application is to collect historical user traffic billing data of a data center and perform preprocessing to obtain multiple traffic data subsequences; determine the aggregation residual from the traffic data in each traffic data subsequence, obtain the re-aggregation factor of each traffic data subsequence according to the aggregation residual, and then determine the aggregation likelihood coefficient corresponding to each traffic data subsequence through the re-aggregation factor; for each traffic data subsequence in the multiple traffic data subsequences, perform feature reconstruction on the traffic data subsequence, map the traffic data in the reconstructed traffic data subsequence to a space matrix, and obtain a traffic feature space matrix; determine the traffic feature covariance matrix according to the traffic feature space matrix, and obtain the traffic feature covariance matrix from the traffic feature covariance matrix. The method comprises the following steps: determining the traffic characteristic value of the traffic data subsequence by an array, determining the traffic billing iteration index of the traffic data subsequence according to the traffic characteristic value, and then determining the traffic billing iteration index corresponding to each traffic data subsequence; determining the traffic billing prediction function value corresponding to each traffic data subsequence according to the accumulation likelihood coefficient and the traffic billing iteration index corresponding to each traffic data subsequence, and superimposing the traffic billing prediction function values of each traffic data subsequence to obtain the traffic billing prediction function value of the traffic data sequence; adjusting the network bandwidth of the traffic transmission link during bandwidth flow billing according to the traffic billing prediction function value, thereby improving the network stability under different network environments and ensuring the accuracy of the billing results.

In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification. Referring to Figure 1, this figure is an exemplary flow chart of a data center network bandwidth flow automatic billing adjustment method according to some embodiments of the present application. The data center network bandwidth flow automatic billing adjustment method 100 mainly includes the following steps:

In step S101, historical traffic billing data of users in a data center is collected and preprocessed to obtain a plurality of traffic data subsequences.

In some embodiments, historical traffic billing data of users in a data center is collected, and the historical traffic billing data of users is divided into blocks to obtain a traffic data sequence;

The flow data sequence is smoothed to obtain a flow data smoothed sequence, and the flow data smoothed sequence is decomposed according to the user flow usage frequency to obtain a plurality of flow data subsequences.

In specific implementation, the data collection tool can be deployed on the servers and devices in the data center to collect the historical traffic billing data of users in the data center. For example, Wireshark in the existing technology can be used to obtain traffic billing data. Wireshark is an open source tool for network protocol analysis and data packet capture, which allows users to capture network data packets, analyze network communications, and detect and diagnose network problems.

In some embodiments, a flow quantity threshold may be set, and when the flow data volume reaches the threshold, the data is divided into a new block to obtain a flow data sequence.

It should be noted that the smoothing process in the present application can be processed using the moving average method in the prior art. The moving average method calculates the average value of the flow data in the flow data sequence and replaces the original data points with the average value. Smoothing the flow data sequence can reduce noise and fluctuations in the data and make the data tend to be stable.

In specific implementation, the traffic data sequence is decomposed according to the user's traffic usage frequency. For example, according to the user's traffic usage demand, the traffic data sequence is decomposed according to the hourly traffic usage frequency and decomposed into 24 sub-blocks corresponding to 24 hours, and the decomposed sub-blocks are used as traffic data sub-sequences.

In step S102, an accumulation residual is determined from the flow data in each flow data subsequence, a re-accumulation factor of each flow data subsequence is obtained according to the accumulation residual, and then the accumulation likelihood coefficient corresponding to each flow data subsequence is determined by the re-accumulation factor.

In some embodiments, determining the aggregate residual from the flow data in each flow data subsequence may be implemented by the following steps:

Determine the mean value of the traffic data corresponding to each traffic data subsequence ;

The accumulated residual of each flow data subsequence is determined according to the flow data mean, wherein the accumulated residual is determined according to the following formula:

in, Indicates the first/> The accumulated residual of the subsequences of traffic data, /> Indicates the first/> The first/> in the flow data subsequence flow data values, /> Represents the number of flow data values in a single flow data subsequence.

In some embodiments, obtaining the re-aggregation factor of each flow data subsequence according to the aggregated residual may be implemented by the following steps:

Get the accumulated residual of each flow data subsequence ;

Determine the accumulation heterogeneity value corresponding to each flow data subsequence ;

The re-aggregation factor of each flow data subsequence is determined by the aggregation residual and the aggregation heterogeneity value, wherein the re-aggregation factor is determined according to the following formula:

Indicates the first/> The re-aggregation factor of the traffic data subsequences, /> Indicates the first/> The first flow data values, /> Represents the number of flow data values in a single flow data subsequence.

In some embodiments, determining the accumulation likelihood coefficients corresponding to each traffic data subsequence respectively by the re-aggregation factor can be implemented by the following steps:

Get the re-aggregation factor of each flow data subsequence ;

Determine the average reaggregation factor of each flow data subsequence based on the reaggregation factor ;

Determine the re-aggregation attenuation factor for each flow data subsequence ;

The aggregation likelihood coefficient is determined according to the re-aggregation factor, the average re-aggregation factor and the re-aggregation attenuation factor, wherein the aggregation likelihood coefficient is determined according to the following formula:

Indicates the first/> The cumulative likelihood coefficient of a traffic data subsequence.

In step S103, for each of the multiple traffic data subsequences, feature reconstruction is performed on the traffic data subsequence, and the traffic data in the reconstructed traffic data subsequence is mapped into a space matrix to obtain a traffic feature space matrix.

In some embodiments, for each of the plurality of traffic data subsequences, feature reconstruction of the traffic data subsequence may be implemented by the following steps:

Determine the delay time according to the flow data value in the flow data subsequence;

Determine the embedding dimension according to the flow data value in the flow data subsequence;

The characteristics of the traffic subsequence are reconstructed according to the delay time value and the embedding dimension.

It should be noted that the delay time represents the time interval between flow data values in the flow data subsequence.

In the specific implementation, the delay time is obtained, and the data in the traffic data subsequence is mapped into a one-dimensional space, and the dimension of the space is gradually increased to obtain a traffic embedding vector. For example, the dimension of the space is increased to 2, 3, 4, etc., so as to better represent the dynamic characteristics of the traffic time subsequence. The distance between each traffic data point and its nearest neighboring point in each dimensional space is calculated, and these nearest neighboring points are recorded as neighboring points.

A distance threshold is preset. If the distance between neighboring points is within the threshold, they are judged as real neighboring points. If the distance between them exceeds the threshold, they are judged as false neighboring points. The proportion of false neighboring points in each dimension is calculated, that is, the ratio of the number of false neighboring points to the total number of neighboring points. The embedding dimension is selected according to the ratio of false neighboring points, and the minimum value of the ratio of false neighboring points is rounded as the embedding dimension.

The embedding dimension in the present application is determined by using the false neighbor point method in the prior art. Other methods may also be used in other embodiments, which are not limited here.

In the specific implementation, the traffic data subsequence is feature reconstructed according to the delay time and the embedding dimension. The delay time determines the time interval between the traffic data points in the reconstructed subsequence, and the embedding dimension determines the dimension of the reconstructed embedding vector. Then, the normalized features of the traffic data subsequence are extracted and converted into a traffic embedding vector subsequence. The traffic embedding vector subsequence is a high-dimensional representation of the original traffic data subsequence, which can better obtain the features of the traffic data for further analysis and processing.

In some embodiments, mapping the traffic data in the reconstructed traffic data subsequence into a spatial matrix to obtain a traffic feature spatial matrix may be implemented by the following steps:

Obtain a spatial matrix, where each row of the matrix represents a different time period, and each column of the matrix represents a flow data feature;

The traffic data in the traffic data subsequence are mapped one by one to the space matrix in chronological order from small to large, and the traffic feature space matrix is obtained.

It should be noted that the mapping method used in the present application is a direct mapping method in the prior art. Other mapping methods may also be used in other embodiments, which are not limited here.

In step S104, the traffic feature covariance matrix is determined according to the traffic feature space matrix, the traffic feature value of the traffic data subsequence is determined by the traffic feature covariance matrix, the traffic billing iterative index of the traffic data subsequence is determined according to the traffic feature value, and then the traffic billing iterative index corresponding to each traffic data subsequence is determined.

In some embodiments, determining the traffic feature covariance matrix according to the traffic feature space matrix may be implemented by the following steps:

Centralize the traffic feature space matrix;

Determine the covariance of the flow data values in the flow feature space matrix after the centralization process;

The covariance of all traffic data values in the traffic feature space matrix is filled into the corresponding positions of the matrix in chronological order to obtain the traffic feature covariance matrix.

It should be noted that the centralization processing in the present application is a method of data preprocessing for the traffic feature space matrix. The centralization processing is to move the mean of the traffic data in the traffic feature space matrix to near zero, eliminate the translation difference of the traffic data values in the traffic feature space matrix, reduce collinearity, and because the mean of the traffic data after centering is zero, the calculation of the covariance matrix is simplified.

In the specific implementation, an empty matrix is created. Each row of the matrix represents a different time period, and each column of the matrix represents the traffic data feature. The covariance of all traffic data values in the traffic feature space matrix is filled into the corresponding position of the matrix in chronological order from small to large to obtain the traffic feature covariance matrix.

In some embodiments, the flow characteristic value of the flow data subsequence is determined by the flow characteristic covariance matrix, and the flow billing iterative index of the flow data subsequence is determined according to the flow characteristic value, and then the flow billing iterative index corresponding to each flow data subsequence is determined, which can be implemented by the following steps:

Calculate the eigenvalues of the traffic characteristics covariance matrix ;

Get the mean value of the traffic data corresponding to each traffic data subsequence ;

Determine the flow attribute value of each flow data value in the flow characteristic covariance matrix ;

The flow billing iterative index is determined by the eigenvalue of the flow characteristic covariance matrix and the flow change value of each flow data value, wherein the flow billing iterative index can be determined according to the following formula:

Indicates the first/> The traffic billing iteration index of a traffic data subsequence.

In step S105, the flow billing prediction function value corresponding to each flow data subsequence is determined according to the accumulated likelihood coefficient and flow billing iteration index corresponding to each flow data subsequence, and the flow billing prediction function value of each flow data subsequence is superimposed to obtain the flow billing prediction function value of the flow data sequence.

In specific implementation, the traffic billing prediction function value can be realized by the following steps:

Get the cumulative likelihood coefficient corresponding to each traffic data subsequence ;

Get the traffic billing iteration index corresponding to each traffic data subsequence ;

The flow billing prediction function value is determined by the accumulation likelihood coefficient and the flow billing iteration index, wherein the flow billing prediction function value is determined according to the following formula:

Indicates the first/> The traffic billing prediction function value of a traffic data subsequence,/> Indicates the first/> The first/> in the flow data subsequence flow data value.

In some embodiments, the process of superimposing the traffic billing prediction function values of each traffic data subsequence to obtain the traffic billing prediction function value of the traffic data sequence can be obtained by adding the traffic billing prediction function values of all traffic data subsequences. It should be noted that the traffic billing prediction function value is a prediction result obtained by function prediction based on the historical traffic data of the data center, reflecting the next traffic transmission demand based on the behavior habits of the data center. Therefore, the bandwidth of the traffic transmission link can be adjusted by the traffic billing prediction function value, thereby improving the network stability under different network environments during network billing, and ensuring the accuracy of the billing results when the network fluctuations are large.

In step S106, the network bandwidth of the traffic transmission link during bandwidth flow billing is adjusted according to the traffic billing prediction function value.

In some embodiments, a function value proportional coefficient is obtained by comparing the traffic billing prediction function value with a preset function value threshold. In specific implementation, the ratio between the traffic billing prediction function value and the preset function value threshold can be used as the function value proportional coefficient, and then the bandwidth of the traffic transmission link is proportionally adjusted by the function value proportional coefficient. For example, a proportional-integral-derivative (PID) controller is added to the forward path of the router bandwidth control system of the data center, and the function value proportional coefficient is used as a proportional parameter in the PID controller, so that the bandwidth occupied by the data center is adjusted in advance according to the function value proportional coefficient, thereby improving the network stability under different network environments during network billing and ensuring the accuracy of the billing results when the network fluctuation is large.

In addition, in another aspect of the present application, in some embodiments, the present application provides a data center network bandwidth flow automatic billing adjustment device, the device includes a bandwidth control unit, refer to Figure 2, which is a schematic diagram of exemplary hardware and/or software of the bandwidth control unit shown in some embodiments of the present application, the bandwidth control unit 200 includes: a flow data subsequence determination module 201, an accumulation likelihood coefficient determination module 202, a flow feature space matrix determination module 203, a flow billing iterative index determination module 204, a flow billing prediction function value determination module 205 and a bandwidth control module 206, which are respectively described as follows:

Traffic data subsequence determination module 201, in the present application, the traffic data subsequence determination module 201 is mainly used to collect historical traffic billing data of users in the data center and perform preprocessing to obtain multiple traffic data subsequences;

An accumulation likelihood coefficient determination module 202, in the present application, is mainly used to determine an accumulation residual from the flow data in each flow data subsequence, obtain a re-aggregation factor of each flow data subsequence according to the accumulation residual, and then determine the accumulation likelihood coefficient of each flow data subsequence through the re-aggregation factor;

The traffic feature space matrix determination module 203 in the present application is mainly used to reconstruct each traffic data subsequence in a plurality of traffic data subsequences according to the traffic feature, and map the traffic data in the reconstructed traffic data subsequence to the space matrix to obtain the traffic feature space matrix;

The flow billing iterative index determination module 204 in the present application is mainly used to determine the flow feature covariance matrix according to the flow feature space matrix, determine the flow feature value of the flow data subsequence by the flow feature covariance matrix, determine the flow billing iterative index of the flow data subsequence according to the flow feature value, and then determine the flow billing iterative index corresponding to each flow data subsequence;

The flow billing prediction function value determination module 205 in the present application is mainly used to determine the flow billing prediction function value corresponding to each flow data subsequence according to the accumulation likelihood coefficient and flow billing iterative index corresponding to each flow data subsequence, and superimpose the flow billing prediction function values of each flow data subsequence to obtain the flow billing prediction function value of the flow data sequence;

The bandwidth control module 206 in the present application is mainly used to adjust the network bandwidth of the traffic transmission link during bandwidth flow billing according to the traffic billing prediction function value.

In addition, the present application also provides a computer device, which includes a memory and a processor, the memory stores code, and the processor is configured to obtain the code and execute the above-mentioned data center network bandwidth traffic automatic billing adjustment method.

In some embodiments, referring to FIG3, this figure is a schematic diagram of the structure of a computer device for applying a data center network bandwidth flow automatic billing adjustment method according to some embodiments of the present application. The data center network bandwidth flow automatic billing adjustment method in the above embodiment can be implemented by a computer device as shown in FIG3, which includes at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

Processor 301 can be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC) or one or more processors for controlling the execution of the automatic billing and adjustment method for data center network bandwidth traffic in the present application.

The communication bus 302 may include a pathway for transmitting information between the above-mentioned components.

The memory 303 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via the communication bus 302. The memory 303 may also be integrated with the processor 301.

The memory 303 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 301. The processor 301 is used to execute the program code stored in the memory 303. The program code may include one or more software modules. The automatic billing and adjustment method for data center network bandwidth flow in the above embodiment can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

The communication interface 304 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-CPU processor or a multi-CPU processor. The processor here may refer to one or more devices, circuits, and/or processing cores for processing data (e.g., computer program instructions).

The above-mentioned computer device may be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device may be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

In addition, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned data center network bandwidth flow automatic billing adjustment method is implemented.

In summary, in the method and device for automatic billing and adjustment of data center network bandwidth traffic disclosed in the embodiment of the present application, first, historical traffic billing data of users in the data center is collected and preprocessed to obtain multiple traffic data subsequences, the accumulation residual is determined from the traffic data in the traffic data subsequence, and then the accumulation likelihood coefficient of the traffic data subsequence is determined, the traffic data subsequence is feature reconstructed, the traffic feature space matrix is determined from the reconstructed traffic data subsequence, and then the traffic feature covariance matrix is obtained, the traffic feature value is determined from the traffic feature covariance matrix, the traffic billing iteration index of the traffic data subsequence is determined according to the traffic feature value, the traffic billing prediction function of the traffic data subsequence is determined according to the accumulation likelihood coefficient and the traffic billing iteration index, and the network bandwidth of the traffic transmission link is adjusted according to the traffic billing prediction function during bandwidth traffic billing, thereby improving the network stability under different network environments and ensuring the accuracy of the billing results.

Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims (8)

1. An automatic charging adjustment method for data center network bandwidth flow is characterized by comprising the following steps:

Collecting historical flow charging data of a user of a data center and preprocessing the data to obtain a plurality of flow data subsequences;

Determining accumulation residual errors according to flow data in each flow data subsequence, obtaining a re-accumulation factor of each flow data subsequence according to the accumulation residual errors, and further determining accumulation likelihood coefficients corresponding to each flow data subsequence respectively through the re-accumulation factors;

for each flow data subsequence in the plurality of flow data subsequences, performing feature reconstruction on the flow data subsequence, and mapping flow data in the reconstructed flow data subsequence into a space matrix to obtain a flow feature space matrix;

determining a flow characteristic covariance matrix according to the flow characteristic space matrix, determining a flow characteristic value of the flow data subsequence by the flow characteristic covariance matrix, determining a flow charging iteration index of the flow data subsequence according to the flow characteristic value, and further determining flow charging iteration indexes respectively corresponding to the flow data subsequences;

Determining flow charging prediction function values corresponding to the flow data subsequences according to the accumulation likelihood coefficients and the flow charging variation indexes corresponding to the flow data subsequences, and superposing the flow charging prediction function values of the flow data subsequences to obtain flow charging prediction function values of the flow data sequences;

and adjusting the time-consuming network bandwidth of the bandwidth flowmeter of the flow transmission link according to the flow charging prediction function value.

2. The method of claim 1, wherein collecting and preprocessing historical traffic billing data for a user of the data center to obtain a plurality of traffic data subsequences comprises:

collecting user historical flow charging data of a data center, and dividing the user historical flow charging data into blocks to obtain a flow data sequence;

performing smoothing treatment on the flow data sequence to obtain a flow data smoothing sequence;

and decomposing the flow data smoothing sequence according to the use frequency of the user flow to obtain a plurality of flow data subsequences.

3. The method of claim 1, wherein determining respective accumulation likelihood coefficients for each flow data sub-sequence by the re-accumulation factor comprises:

acquiring a reaggregation factor of each flow data subsequence;

determining an average re-accumulation factor of each flow data sub-sequence;

determining a refocusing attenuation factor of each flow data sub-sequence;

and determining an accumulation likelihood coefficient according to the re-accumulation factor, the average re-accumulation factor and the re-accumulation attenuation factor.

4. The method of claim 1, wherein determining a flow feature covariance matrix from the flow feature space matrix comprises:

Carrying out centering treatment on the flow characteristic space matrix;

Determining covariance of flow data values in the flow characteristic space matrix after the centralization treatment;

And filling covariance of all flow data values in the flow characteristic space matrix to corresponding positions of the matrix according to time sequence to obtain a flow characteristic covariance matrix.

5. The method of claim 1 wherein determining a flow charging iteration index for the flow data subsequence based on the flow characteristic value comprises:

Acquiring a flow characteristic value of the flow data subsequence;

determining a flow data mean value in the flow data subsequence;

And determining the flow charging iteration index of the flow data subsequence according to the flow characteristic value and the flow data average value.

6. The utility model provides a data center network bandwidth flow automatic charging adjusting device which characterized in that, including bandwidth control unit, bandwidth control unit includes:

The flow data subsequence determining module is used for collecting user historical flow charging data of the data center, dividing the user historical flow charging data into blocks to obtain a flow data sequence, further smoothing the flow data sequence to obtain a flow data smoothing sequence, and decomposing the flow data smoothing sequence according to the use frequency of the user flow to obtain a plurality of flow data subsequences;

The accumulation likelihood coefficient determining module is used for determining an accumulation residual error from the flow data in each flow data subsequence, obtaining a re-accumulation factor of each flow data subsequence according to the accumulation residual error, and further determining the accumulation likelihood coefficient of each flow data subsequence through the re-accumulation factor;

The flow characteristic space matrix determining module is used for reconstructing each flow data subsequence in the plurality of flow data subsequences according to the flow characteristics, and mapping flow data in the reconstructed flow data subsequence into a space matrix to obtain a flow characteristic space matrix;

The flow charging iteration index determining module is used for determining a flow characteristic covariance matrix according to the flow characteristic space matrix, determining a flow characteristic value of the flow data subsequence according to the flow characteristic covariance matrix, determining a flow charging iteration index of the flow data subsequence according to the flow characteristic value, and further determining flow charging iteration indexes corresponding to the flow data subsequences respectively;

the flow charging prediction function value determining module is used for determining flow charging prediction function values respectively corresponding to the flow data subsequences according to the accumulation likelihood coefficients and the flow charging variation indexes respectively corresponding to the flow data subsequences, and superposing the flow charging prediction function values of the flow data subsequences to obtain flow charging prediction function values of the flow data sequences;

and the bandwidth control module is used for adjusting the time-consuming network bandwidth of the bandwidth flowmeter of the flow transmission link according to the flow charging prediction function value.

7. A computer device comprising a memory storing code and a processor configured to obtain the code and to perform the data center network bandwidth flow automatic billing adjustment method of any of claims 1 to 5.

8. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the data center network bandwidth flow automatic billing adjustment method of any of claims 1 to 5.