Method, device and equipment for predicting service increment of content distribution network and medium

Abstract

The embodiment of the invention provides a service increment prediction method and device for a content distribution network, equipment and a medium. The method comprises the following steps: acquiring bandwidth data of a target service in a plurality of historical periods; selecting the bandwidth data of the latest historical period as reference bandwidth data according to the bandwidth data of the plurality of historical periods; determining the similarity between the reference bandwidth data and the bandwidth data of other historical periods except the reference bandwidth data; determining a preset number of bandwidth data of a target historical period according to the similarity; performing weighted average on the bandwidth data of the target historical period, and determining estimated bandwidth data of the next period; and determining target increment information of the next period according to the estimated bandwidth data of the next period. The bandwidth data of the target historical period is screened out by utilizing the bandwidth data of the plurality of historical periods, and the target increment information is determined according to the bandwidth data of the target historical period, so that the efficiency and accuracy of predicting the target increment information of the next period are improved.

Description

Technical Field

[0001] The present invention relates to the field of network communication technology, and in particular to a method, device, equipment and medium for predicting service increment of a content distribution network. Background Art

[0002] With the booming development of the digital economy and the widespread popularity of the Internet, the content delivery network (CDN) market has ushered in an unprecedented rapid growth. The number of customers accessed by CDN vendors continues to rise, and the scale of business continues to expand. In this context, CDN vendors need to allocate differentiated bandwidth resources to meet the diverse business needs of different customers to ensure a high-quality experience of CDN acceleration services. For large customers, the end of each month is often accompanied by significant incremental business demand. In order to effectively respond to the challenges of limited total resources and internal cost control while ensuring stable business operation, it is particularly critical to accurately assess the resource demand for the next month.

[0003] Traditional methods usually rely on manual collection of customer service increment information in advance, and then allocate resources based on the collected incremental information, and record data through manual inspection of service bandwidth after the increment to compare the actual service increment. However, this method consumes a lot of manpower and time, and has relatively low efficiency and accuracy. Summary of the invention

[0004] In view of the above problems, embodiments of the present invention are proposed to provide a method, apparatus, device and medium for predicting service increment of a content distribution network that overcomes the above problems or at least partially solves the above problems.

[0005] In order to solve the above problem, an embodiment of the present invention discloses a method for predicting service increment of a content distribution network, the method comprising:

[0006] Obtain bandwidth data of multiple historical periods of the target service; wherein the bandwidth data of each historical period includes bandwidth data of multiple sampling moments collected according to a preset sampling interval;

[0007] According to the bandwidth data of the multiple historical periods, the bandwidth data of the most recent historical period is selected as the reference bandwidth data;

[0008] Determine the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data;

[0009] Determining bandwidth data of a preset number of target historical periods according to the similarity;

[0010] Taking a weighted average of the bandwidth data of the target historical period to determine the estimated bandwidth data of the next period;

[0011] The target increment information of the next cycle is determined according to the estimated bandwidth data of the next cycle.

[0012] Optionally, determining the similarity between the reference bandwidth data and bandwidth data of other historical periods other than the reference bandwidth data includes:

[0013] Calculating a first distance between the reference bandwidth data and the bandwidth data of other historical periods according to a dynamic time warping function; the first distance is inversely correlated with the similarity;

[0014] The similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data is determined according to the first distance.

[0015] Optionally, calculating the first distance between the reference bandwidth data and the bandwidth data of other historical periods according to the dynamic time warping function includes:

[0016] Determining sub-bandwidth data of the reference bandwidth data;

[0017] Calculate, according to the dynamic time warping function, a second distance between the sub-bandwidth data of the reference bandwidth data and the bandwidth data of the other historical periods;

[0018] If the second distance is less than or equal to the first preset threshold, the first distance between the reference bandwidth data and bandwidth data of other historical periods corresponding to the second distance is calculated.

[0019] Optionally, determining bandwidth data of a preset number of target historical periods according to the similarity includes:

[0020] Determine bandwidth data of a candidate historical period corresponding to a first distance that is less than or equal to a second preset threshold;

[0021] Sorting the bandwidth data of the candidate historical periods in ascending order of the first distances;

[0022] A preset number of candidate historical period bandwidth data are selected in the order as the target historical period bandwidth data.

[0023] Optionally, determining the target increment information of the next period according to the estimated bandwidth data of the next period includes:

[0024] Determining a first estimated incremental value according to the estimated bandwidth data of the next cycle;

[0025] Obtaining a second estimated increment value;

[0026] The target increment information of the next cycle is determined according to the first estimated increment value and the second estimated increment value.

[0027] Optionally, determining the target increment information of the next cycle according to the first estimated increment value and the second estimated increment value includes:

[0028] Obtain estimated accuracy and business risk values;

[0029] The target increment information of the next cycle is determined according to the estimated accuracy value, the business risk value, the first estimated increment value and the second estimated increment value.

[0030] Optionally, before selecting bandwidth data of the most recent historical period as reference bandwidth data based on the bandwidth data of the multiple historical periods, the method further includes:

[0031] For the bandwidth data of each historical period, if the bandwidth at the sampling moment is less than a third preset threshold, the bandwidth at a sampling moment before the sampling moment is greater than or equal to the third preset threshold, and the bandwidth at a sampling moment after the sampling moment is greater than or equal to the third preset threshold, it is determined that the bandwidth at the sampling moment is abnormal, and the bandwidth at the sampling moment is determined by interpolation based on the bandwidth at the previous sampling moment and the bandwidth at the next sampling moment;

[0032] If the bandwidth at the sampling moment is greater than a fourth preset threshold, and the ratio of the difference between the bandwidth at the sampling moment and the bandwidth at the previous sampling moment to the bandwidth at the previous sampling moment is greater than the fourth preset threshold, the bandwidth at the sampling moment is determined to be a sudden drop value, and the bandwidth at the sampling moment is smoothed.

[0033] Accordingly, an embodiment of the present invention discloses a service increment prediction device for a content distribution network, the device comprising:

[0034] A bandwidth data acquisition module, used to acquire bandwidth data of multiple historical periods of the target service; wherein the bandwidth data of each historical period includes bandwidth data of multiple sampling moments collected according to a preset sampling interval;

[0035] A reference bandwidth data determination module, configured to select bandwidth data of the most recent historical period as reference bandwidth data based on the bandwidth data of the multiple historical periods;

[0036] A similarity determination module, used to determine the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data;

[0037] A target historical period bandwidth data determination module, used to determine a preset number of target historical period bandwidth data according to the similarity;

[0038] An estimated bandwidth data determination module, used to perform weighted averaging on the bandwidth data of the target historical period to determine the estimated bandwidth data of the next period;

[0039] The target increment information determination module is used to determine the target increment information of the next cycle according to the estimated bandwidth data of the next cycle.

[0040] Optionally, the similarity determination module includes:

[0041] A first distance determination submodule, configured to calculate a first distance between the reference bandwidth data and the bandwidth data of other historical periods according to a dynamic time warping function; the first distance is inversely correlated with the similarity;

[0042] The similarity determination submodule is used to determine the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data according to the first distance.

[0043] Optionally, the first distance determination submodule includes:

[0044] a sub-bandwidth data determining unit, configured to determine sub-bandwidth data of the reference bandwidth data;

[0045] A second distance determination unit, configured to calculate, according to the dynamic time warping function, a second distance between the sub-bandwidth data of the reference bandwidth data and the bandwidth data of the other historical periods;

[0046] The first distance determining unit is configured to calculate the first distance between the reference bandwidth data and bandwidth data of other historical periods corresponding to the second distance if the second distance is less than or equal to a first preset threshold.

[0047] Optionally, the bandwidth data determination module of the target historical period includes:

[0048] A bandwidth data determination submodule for a candidate historical period, configured to determine bandwidth data for a candidate historical period corresponding to a first distance that is less than or equal to a second preset threshold;

[0049] A sorting submodule, used to sort the bandwidth data of the candidate historical periods in order of the first distance from small to large;

[0050] The target historical period bandwidth data determination submodule is used to select a preset number of candidate historical period bandwidth data in the order as the target historical period bandwidth data.

[0051] Optionally, the target increment information determination module includes:

[0052] A first estimated incremental value determining submodule, configured to determine a first estimated incremental value according to the estimated bandwidth data of the next period;

[0053] A second estimated incremental value acquisition submodule, used to acquire a second estimated incremental value;

[0054] The target increment information determination submodule is used to determine the target increment information of the next cycle according to the first estimated increment value and the second estimated increment value.

[0055] Optionally, the target increment information determination submodule includes:

[0056] A data acquisition unit, used to obtain an estimated accuracy value and a business risk value;

[0057] A target incremental information determination unit is used to determine the target incremental information of the next cycle according to the estimated accuracy value, the business risk value, the first estimated incremental value and the second estimated incremental value.

[0058] Optionally, the device further comprises:

[0059] an interpolation processing module, for determining, for bandwidth data of each historical period, that the bandwidth at the sampling moment is abnormal if the bandwidth at the sampling moment is less than a third preset threshold, the bandwidth at the previous sampling moment is greater than or equal to the third preset threshold, and the bandwidth at the next sampling moment is greater than or equal to the third preset threshold, and interpolating the bandwidth at the previous sampling moment and the bandwidth at the next sampling moment to determine the bandwidth at the sampling moment;

[0060] A smoothing processing module is used to determine that the bandwidth at the sampling moment is a sudden drop value, and smooth the bandwidth at the sampling moment if the bandwidth at the sampling moment is greater than a fourth preset threshold, and the ratio of the difference between the bandwidth at the sampling moment and the bandwidth at the previous sampling moment to the bandwidth at the previous sampling moment is greater than the fourth preset threshold.

[0061] Accordingly, an embodiment of the present invention discloses an electronic device, comprising: a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, each step of the above-mentioned content distribution network service increment prediction method embodiment is implemented.

[0062] Accordingly, an embodiment of the present invention discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, each step of the above-mentioned content distribution network service increment prediction method embodiment is implemented.

[0063] The embodiments of the present invention include the following advantages:

[0064] The service increment prediction method of the content distribution network of the embodiment of the present invention obtains bandwidth data of multiple historical periods of the target service; wherein the bandwidth data of each historical period includes bandwidth data of multiple sampling moments collected according to a preset sampling interval; based on the bandwidth data of multiple historical periods, the bandwidth data of the most recent historical period is selected as reference bandwidth data; the similarity between the reference bandwidth data and the bandwidth data of other historical periods other than the reference bandwidth data is determined; based on the similarity, the bandwidth data of a preset number of target historical periods is determined; the bandwidth data of the target historical period is weighted averaged to determine the estimated bandwidth data of the next period; based on the estimated bandwidth data of the next period, the target increment information of the next period is determined. By using the bandwidth data of multiple historical periods to filter out the bandwidth data of the target historical period, and determining the target increment information based on the bandwidth data of the target historical period, the efficiency and accuracy of predicting the target increment information of the next period are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a flowchart of the steps of a method for predicting service increment of a content distribution network according to an embodiment of the present invention;

[0066] Figure 2 is a schematic diagram of another method for predicting service increment of a content distribution network according to the present invention;

[0067] Figure 3 The present invention is a structural block diagram of a device for predicting service increment of a content distribution network according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0069] With the booming development of the digital economy and the widespread popularity of the Internet, the content distribution network market has ushered in an unprecedented rapid growth. The number of customers accessed by CDN vendors continues to rise, and the scale of business continues to expand. In this context, CDN vendors need to allocate differentiated bandwidth resources to meet the diverse business needs of different customers to ensure the high-quality experience of CDN acceleration services. For large customers, the end of each month is often accompanied by significant incremental business demand. In order to effectively respond to the challenges of limited total resources and internal cost control while ensuring stable business operation, how to accurately evaluate the resource demand for the next month is particularly critical.

[0070] Traditional methods usually rely on manual collection of customer service increment information in advance, and then allocate resources based on the collected incremental information, and record data through manual inspection of service bandwidth after the increment to compare the actual service increment. However, this method consumes a lot of manpower and time, and has relatively low efficiency and accuracy.

[0071] One of the core concepts of the embodiments of the present invention is to filter out bandwidth data of a target historical period by utilizing bandwidth data of multiple historical periods, and determine target incremental information based on the bandwidth data of the target historical period, thereby improving the efficiency and accuracy of predicting the target incremental information of the next period.

[0072] Reference Figure 1 , shows a flowchart of a method for predicting service increment of a content distribution network according to an embodiment of the present invention, which may specifically include the following steps:

[0073] Step 101: Obtain bandwidth data of a plurality of historical periods of a target service; wherein the bandwidth data of each historical period includes bandwidth data of a plurality of sampling moments collected at a preset sampling interval.

[0074] Specifically, if the customer service has at least one historical period in the content distribution network increment, the historical bandwidth data of multiple historical periods of the target customer service can be obtained according to the customer name or customer domain name. For example, K historical period bandwidth data Y can be obtained. his =(Y1,Y2,…,Y k ), there are N sampling moments in the i-th historical period, and the bandwidth data of a historical period is Y i =(y i1 ,y i2 ,…,y iN ), where 1≤i≤K, and K can be set according to actual needs.

[0075] The length T of a historical cycle can be 1 month, and the sampling interval tp can be 1 minute, 5 minutes, 1 hour, etc. For example, the length T of the historical cycle is 1 month (30 days), and the sampling interval tp is 5 minutes. In theory, historical bandwidth data of 8640 sampling moments can be obtained in one historical cycle. Of course, the length of the historical cycle and the sampling interval can be set according to actual needs in actual applications, and the embodiment of the present invention does not limit this.

[0076] Step 102: Based on the bandwidth data of the multiple historical periods, select the bandwidth data of the most recent historical period as reference bandwidth data.

[0077] Specifically, based on bandwidth data of multiple historical periods, bandwidth data of the historical period closest to the current time is selected as reference bandwidth data, and the reference bandwidth data is the matching sample Y0.

[0078] Step 103: Determine the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data.

[0079] Specifically, after the reference bandwidth data Y0 is selected, the similarity between the reference bandwidth data Y0 and bandwidth data of other historical periods except the reference bandwidth data Y0 is calculated.

[0080] Step 104: Determine bandwidth data of a preset number of target historical periods according to the similarity.

[0081] Specifically, after determining the similarity between the reference bandwidth data Y0 and the bandwidth data of other historical periods, based on the similarity, bandwidth data of a preset number of target historical periods are determined from the bandwidth data of other historical periods. Exemplarily, bandwidth data of w target historical periods are determined. Of course, the preset number can be set according to actual needs in actual applications, and the embodiment of the present invention is not limited to this.

[0082] In actual business, there may be no increment in a historical period, or only increment in some periods, or only a small amount of grayscale in a period. Therefore, it is necessary to first filter out the bandwidth data of the target historical period from the bandwidth data of multiple historical periods, and then perform subsequent estimated increments to improve the accuracy of the estimate.

[0083] Step 105: Perform weighted averaging on the bandwidth data of the target historical period to determine the estimated bandwidth data of the next period.

[0084] Specifically, the bandwidth data of w target historical periods are weighted averaged to obtain the bandwidth data of each sampling time in the next period, which can be calculated according to the following formula:

[0085]

[0086] Among them, Y′ is the estimated bandwidth data, D i is the weight, Y d,i is the bandwidth data of the target historical period. The weight corresponding to each target historical period is determined according to the similarity between the reference bandwidth data Y0 and the bandwidth data of the target historical period. The greater the similarity, the greater the weight corresponding to the bandwidth data of the target historical period.

[0087] Step 106: Determine the target increment information of the next cycle according to the estimated bandwidth data of the next cycle.

[0088] Specifically, after determining the estimated bandwidth data Y′ for the next cycle, the estimated peak value, 95th value and other incremental information for the next cycle can be determined based on the estimated bandwidth data Y′, and the target incremental information for the next cycle can be determined based on the estimated peak value, 95th value and other incremental information for the next cycle.

[0089] In the business analysis of content distribution networks, peak value and 95th value are two important performance indicators used to measure network traffic and service quality. Peak value refers to the highest point of network traffic or service requests within a certain time range, which reflects the load of the system at the busiest time. 95th value refers to the 95th percentile value of network traffic or service requests within a certain time range. It means that 95% of the time, the traffic or request is lower than this value. The 95th value is often used to evaluate the average performance of the system, especially when dealing with large fluctuations in traffic.

[0090] In the embodiment of the present invention, step 103 determines the similarity between the reference bandwidth data and bandwidth data of other historical periods other than the reference bandwidth data, which may specifically include the following sub-steps:

[0091] Sub-step S11, calculating a first distance between the reference bandwidth data and the bandwidth data of other historical periods according to a dynamic time warping function; the first distance is inversely correlated with the similarity.

[0092] Specifically, the length of the historical cycle is 1 month. Since the number of days in each month is not necessarily exactly the same, the number of samples in each historical cycle may be different. In order to solve the problem of unequal lengths of two historical cycle sequences, the dynamic time warping function can be selected as the similarity measurement function D between the historical cycle bandwidth sequences. Dynamic Time Warping (DTW) is an algorithm for measuring the similarity between two time series. It is particularly suitable for time series data that are offset, scaled or distorted on the time axis. DTW finds the best matching path between the two sequences by performing nonlinear alignment on the time axis, thereby calculating the distance or similarity between them.

[0093] According to the dynamic time warping function, the first distance between the reference bandwidth data Y0 and the bandwidth data of other historical periods is calculated. The first distance is D(Y0,Y), where D is the dynamic time warping function and Y is the bandwidth data sequence of other historical periods. The smaller the first distance, the greater the similarity.

[0094] Sub-step S12: determining, based on the first distance, the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data.

[0095] Specifically, after determining the first distance between the reference bandwidth data Y0 and the bandwidth data of other historical periods, since the smaller the first distance, the greater the similarity, the similarity between the reference bandwidth data Y0 and the bandwidth data of other historical periods is determined based on the first distance.

[0096] In the embodiment of the present invention, sub-step S11 calculates the first distance between the reference bandwidth data and the bandwidth data of the other historical periods according to the dynamic time warping function; the first distance is inversely correlated with the similarity, and specifically may include the following sub-steps:

[0097] Sub-step S111, determining sub-bandwidth data of the reference bandwidth data.

[0098] Specifically, the reference bandwidth data Y0=(y 01 ,y 02 ,…,y 0N ), select a time-continuous subset Y0[1:jgap]={y 01 ,y 01 ,…,y 0jgap}, as the sub-bandwidth data sequence of Y0, denoted as P, that is, P is a time series with a length of jgap.

[0099] Sub-step S112, calculating, according to the dynamic time warping function, a second distance between the sub-bandwidth data of the reference bandwidth data and the bandwidth data of other historical periods.

[0100] Specifically, according to the dynamic time warping function, the second distance between the sub-bandwidth data of the reference bandwidth data and the bandwidth data of other historical periods is calculated, and the second distance is D(P,Q).

[0101] Sub-step S113: If the second distance is less than or equal to a first preset threshold, calculating the first distance between the reference bandwidth data and bandwidth data of other historical periods corresponding to the second distance.

[0102] Specifically, the first preset threshold ε′ is a constant greater than 0, which can be set according to actual needs. If D=(P,Q)≤ε′, the corresponding sequence Q in the historical cycle bandwidth data at this time is recorded and the subsequent distance calculation is continued; if D=(P,Q) is greater than ε′, the search for the remaining sequences in the historical cycle bandwidth data is stopped.

[0103] In order to reduce the number of sequences in the historical periodic data that need to be matched with the reference bandwidth data Y0, thereby reducing the amount of calculation of the dynamic time warping distance, fast matching is achieved by extracting subsequences.

[0104] In the embodiment of the present invention, step 104 determines bandwidth data of a preset number of target historical periods according to the similarity, which may specifically include the following sub-steps:

[0105] Sub-step S21, determining bandwidth data of a candidate historical period corresponding to a first distance that is less than or equal to a second preset threshold.

[0106] Specifically, the second preset threshold is ε, and the bandwidth data of the candidate historical period corresponding to the first distance D(Y0,Y) less than or equal to ε is determined, and the set of bandwidth data sequences of the candidate historical period is:

[0107] Sub-step S22, sorting the bandwidth data of the candidate historical periods in order of the first distance from small to large.

[0108] Specifically, the first distances satisfying the condition of being less than or equal to ε are sorted in order from small to large, that is, in order from large to small similarities.

[0109] Sub-step S23, selecting a preset number of bandwidth data of candidate historical periods in the order as bandwidth data of the target historical period.

[0110] Specifically, in the order of the first distance from small to large, the bandwidth data of the candidate historical periods corresponding to the first w first distances are used as the bandwidth data of the target historical period, denoted as Y d,i (1≤i≤w).

[0111] In the embodiment of the present invention, step 106 determines the target increment information of the next cycle according to the estimated bandwidth data of the next cycle, which may specifically include the following sub-steps:

[0112] Sub-step S31, determining a first estimated increment value according to the estimated bandwidth data of the next period.

[0113] Specifically, according to the estimated bandwidth data Y′ of the next cycle, the estimated peak value, 95th value and other incremental information of the next cycle are determined, and the estimated peak value, 95th value and other incremental information of the next cycle are used as the first estimated value of the final output, which is recorded as

[0114] Sub-step S32, obtaining a second estimated increment value.

[0115] Specifically, in the actual business docking process, if the business side or the customer provides the incremental information of the next cycle in advance (such as peak value, 95 value, etc.), it can be used as the second estimated value, denoted as

[0116] Sub-step S33, determining the target increment information of the next cycle according to the first estimated increment value and the second estimated increment value.

[0117] Specifically, the first estimated value is corrected by the second estimated value, and the final estimated increment information Y of the next cycle is output. p .

[0118] In the embodiment of the present invention, sub-step S33 determines the target increment information of the next cycle according to the first estimated increment value and the second estimated increment value, and may specifically include the following sub-steps:

[0119] Sub-step S331, obtaining the estimation accuracy value and the business risk value.

[0120] Specifically, the estimated accuracy is the accuracy of the estimated increment value. You can set a time to count the increment of this business. Suppose the actual increment of the business is Y R , the accuracy of the estimated value a can be evaluated by the following formula:

[0121]

[0122] After each increment, the accuracy of the business increment estimate is updated. Assuming that the increment has been n times, a is updated to:

[0123]

[0124] Where λ is the weight of the first estimated value after the increment, and λ1<λ2<…<λ n The weight value can be set in advance to any number within (0,1]. The closer the time is to the current time, the greater the corresponding weight.

[0125] The business risk value b can be calculated based on the estimated accuracy value, combined with the complexity of business configuration (including the number of domain names involved in the increment, the number of development items, etc.), the cooperation of the business contact person, and the number of historical complaints. A set of proportional coefficients is preset for each evaluation dimension, and the corresponding proportional coefficient is determined from the preset set of proportional coefficients. Finally, the estimated accuracy value and the proportional coefficient of each dimension are added to obtain the final business risk coefficient. Taking the evaluation dimension of the number of domain names as an example, assuming that the proportional coefficient corresponding to the number of domain names in the range of [0,10] is β1, and the proportional coefficient corresponding to the number of domain names in the range of [10,50] is β2, and so on, the proportional coefficients corresponding to different numbers of domain names can be obtained.

[0126] Similarly, after each increment, the risk factor for the business is also updated.

[0127] Sub-step S332, determining the target increment information of the next cycle according to the estimated accuracy value, the business risk value, the first estimated increment value and the second estimated increment value.

[0128] Specifically, the target increment information for the next cycle can be determined according to the following formula:

[0129]

[0130] Among them, a is the estimated accuracy value, and b is the business risk value corresponding to the business.

[0131] The business risk value is initialized to 1. The values ​​of a and b need to be updated after each subsequent business increment. It should be noted that if the business is the first increment, the first estimated increment value cannot be obtained through historical cycle data. If the business or customer side can provide the second estimated value information, it is directly confirmed according to the determination formula of the target increment information of the next cycle. At this time, a=1, and the target increment information of the next cycle is

[0132] When obtaining bandwidth data for multiple historical periods of the target business, we can also obtain information such as the business's historical risk coefficient, business type, business coefficient, and whether the resources are covered by special resources. If the business has special coverage resource requirements, this method is not used for incremental prediction, but resources are evaluated separately. For example, some businesses have a low settlement unit price and the customer's business sensitivity may be very high, so it is necessary to communicate with the customer to adjust the resources. Based on the premise that our cost-benefit is appropriate, some regions will adopt special resource coverage.

[0133] After this increment is completed, the risk coefficients corresponding to different businesses will be sorted in ascending order, and different resource priority levels will be divided according to the risk coefficients of different businesses. The higher the risk coefficient, the lower the resource preparation priority. For example, the risk coefficient sorting information is that the top 30% of the business can be associated with a high priority level tag, the bottom 30% of the business can be associated with a low priority level tag, and the remaining business can be associated with a medium priority level tag. The total amount of platform bandwidth resources is fixed and may not be able to meet the incremental needs of all customers. Therefore, when evaluating resources, they need to be allocated to different customers according to certain priorities.

[0134] In the embodiment of the present invention, before selecting the bandwidth data of the most recent historical period as the reference bandwidth data based on the bandwidth data of the multiple historical periods, the method further includes:

[0135] For the bandwidth data of each historical period, if the bandwidth at the sampling moment is less than a third preset threshold, the bandwidth at the previous sampling moment is greater than or equal to the third preset threshold, and the bandwidth at the next sampling moment is greater than or equal to the third preset threshold, then it is determined that the bandwidth at the sampling moment is abnormal, and the bandwidth at the sampling moment is determined by interpolating the bandwidth at the previous sampling moment and the bandwidth at the next sampling moment.

[0136] Specifically, after obtaining bandwidth data of multiple historical periods of the target service, the bandwidth data of multiple historical periods needs to be preprocessed, mainly including two parts: single point missing value and sudden (drop) value.

[0137] A bandwidth threshold δ2 can be set if the bandwidth y(t) corresponding to the sampling time t satisfies the following relationship:

[0138]

[0139] It is considered that the data at the sampling moment is abnormally missing, which may be caused by platform failure and other reasons. At this time, the bandwidth of the sampling moment can be obtained by interpolating the data of the two previous and next sampling moments. The sampling interval is generally at the minute level, and the bandwidth data of adjacent sampling moments are usually smoother. The common linear interpolation method can be used. Of course, the interpolation method can be set according to actual needs in actual applications, and the embodiment of the present invention does not limit this.

[0140] If the bandwidth at the sampling moment is greater than a fourth preset threshold, and the ratio of the difference between the bandwidth at the sampling moment and the bandwidth at the previous sampling moment to the bandwidth at the previous sampling moment is greater than the fourth preset threshold, the bandwidth at the sampling moment is determined to be a sudden drop value, and the bandwidth at the sampling moment is smoothed.

[0141] Specifically, the burst (decrease) value processing is performed, assuming that the service has been incremented at least once. It can be divided into two situations: expected burst and unexpected burst.

[0142] Expected burst (reduction): After the increment, the service characteristics can be obtained through manual analysis. For example, a certain service will have bandwidth bursts during holidays (such as New Year's Day, Spring Festival Gala, etc.). At this time, the bandwidth burst phenomenon in the corresponding period is within expectations, so the sampled bandwidth data of this period in the historical period can be manually marked and retained.

[0143] Unexpected sudden drop: For example, the bandwidth of a certain service suddenly drops or even hits the bottom within a period of time, but after contacting the customer, it is confirmed that it is caused by a bug in the customer's system. At this time, the sudden drop in bandwidth in the corresponding period is unexpected, so the historical data of the corresponding period needs to be manually marked for subsequent smoothing. If it is not possible to manually mark in advance, you can judge whether it is a sudden drop by the bandwidth change ratio according to the following rules:

[0144]

[0145] Among them, δ3 is the bandwidth mutation ratio threshold, and δ4 is the bandwidth threshold. It should be noted that δ3 is related to the service type, and the bandwidth mutation ratio thresholds of different services are not necessarily the same. When the bandwidth corresponding to the sampling moment is greater than the threshold δ4 and the bandwidth change ratio is greater than the set threshold δ3, the bandwidth data corresponding to the moment is considered to be a sudden (drop) value, and the data needs to be smoothed.

[0146] Reference Figure 2 , showing a schematic diagram of another method for predicting business increment of a content distribution network of the present invention, the method includes a data collection and recording module, a data preprocessing module, a data prediction module, a data analysis module and a data display module.

[0147] The data collection and recording module records historical cycle data and business information, and provides a data basis for the data preprocessing module. It should be noted that in the incremental business scenario of the content distribution network, the cycle length in this article is generally one month. The data preprocessing module preprocesses the collected historical cycle bandwidth data according to certain business characteristic rules, including processing single-point missing values ​​and sudden (drop) values ​​in the collected bandwidth data. The data prediction module considers the actual business scenario, screens the multiple historical cycles obtained after preprocessing to quickly match the target historical cycle, and then constructs a prediction model to obtain the first estimated value of the incremental information of the next cycle. In addition, the incremental information provided by the business or customer side is used as the second estimated value, and finally the incremental information estimated for the next cycle is comprehensively estimated by the first estimated value and the second estimated value. The data analysis module compares the incremental estimated value with the actual incremental situation to obtain the accuracy of the estimated value to analyze whether the business increment meets expectations. The business risk coefficient is calculated based on the accuracy of the estimated value and other dimensions such as the complexity of the business configuration. After the increment is completed, the risk coefficient of each business in this cycle is sorted and graded to associate different resource priority labels. The data display module is used to integrate the analysis results of the data analysis module (including business type, business coefficient, risk coefficient, resource priority and other information) to output relevant reports and send them to relevant operators for subsequent business opportunity evaluation and resource preparation. Finally, the analysis results are uploaded and updated to the business management platform again, and the analysis report is output according to the above module. The report content includes customer account and email, business name, corresponding operator and contact information, front-end personnel name and contact information, domain name, edge and parent node coverage plan name, business configuration complexity (including the number of domain names involved in the increment, the number of development items, etc.), business docking person cooperation, historical complaints, estimated value accuracy, risk coefficient and resource priority label. After the increment is completed, the analysis report will be sent to the corresponding operator and resource colleagues so that the next cycle can better prepare resource redundancy for the business and evaluate the accuracy of the business opportunities of the business.

[0148] The method for predicting service increment of a content distribution network according to an embodiment of the present invention has the following advantages:

[0149] More practical. Considering the actual business scenarios, the collected historical data is preprocessed according to the business characteristics and rules; based on the similar waveform of the content distribution network business cycle, the historical cycle data is fully utilized to filter and match the target historical cycle; in addition to obtaining the estimated value based on the historical cycle, the incremental information provided by the business or customer side is also combined for correction, which improves the accuracy of the estimated incremental information; the business risk coefficient is calculated through multiple dimensions for sorting and associated with different priority tags.

[0150] Stronger adaptability: Make full use of the incremental data of the business to continuously update the incremental estimate, the accuracy of the estimate, and the business risk factor.

[0151] Smaller calculation amount. Fast matching of target historical period is achieved by extracting subsequences of period bandwidth data.

[0152] Incremental automated analysis solves the current inefficiency problem of manual analysis of incremental situations, thereby providing a reference for preparing resource redundancy for different customer businesses, which is conducive to improving the rationality of resource allocation and reducing enterprise costs.

[0153] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0154] Reference Figure 3 , shows a structural block diagram of an embodiment of a device for predicting service increments in a content distribution network according to the present invention, which may specifically include the following modules:

[0155] The bandwidth data acquisition module 201 is used to acquire bandwidth data of multiple historical periods of the target service; wherein the bandwidth data of each historical period includes bandwidth data of multiple sampling moments collected according to a preset sampling interval.

[0156] The reference bandwidth data determining module 202 is configured to select bandwidth data of the most recent historical period as reference bandwidth data based on the bandwidth data of the multiple historical periods.

[0157] The similarity determination module 203 is used to determine the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data.

[0158] The target historical period bandwidth data determination module 204 is used to determine a preset number of target historical period bandwidth data according to the similarity.

[0159] The estimated bandwidth data determination module 205 is used to perform weighted averaging on the bandwidth data of the target historical period to determine the estimated bandwidth data of the next period.

[0160] The target increment information determination module 206 is used to determine the target increment information of the next period according to the estimated bandwidth data of the next period.

[0161] The service increment prediction device of the content distribution network of the embodiment of the present invention obtains bandwidth data of multiple historical periods of the target service; wherein the bandwidth data of each historical period includes bandwidth data of multiple sampling moments collected according to a preset sampling interval; selects bandwidth data of the most recent historical period as reference bandwidth data from the bandwidth data of multiple historical periods; determines the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data; determines bandwidth data of a preset number of target historical periods based on the similarity; performs weighted averaging on the bandwidth data of the target historical period to determine the estimated bandwidth data of the next period; determines the target increment information of the next period based on the estimated bandwidth data of the next period. By using the bandwidth data of multiple historical periods to filter out the bandwidth data of the target historical period, and determining the target increment information based on the bandwidth data of the target historical period, the efficiency and accuracy of predicting the target increment information of the next period are improved.

[0162] In an embodiment of the present invention, the similarity determination module includes:

[0163] A first distance determination submodule, configured to calculate a first distance between the reference bandwidth data and the bandwidth data of other historical periods according to a dynamic time warping function; the first distance is inversely correlated with the similarity;

[0164] The similarity determination submodule is used to determine the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data according to the first distance.

[0165] In this embodiment of the present invention, the first distance determination submodule includes:

[0166] a sub-bandwidth data determining unit, configured to determine sub-bandwidth data of the reference bandwidth data;

[0167] A second distance determination unit, configured to calculate, according to the dynamic time warping function, a second distance between the sub-bandwidth data of the reference bandwidth data and the bandwidth data of the other historical periods;

[0168] The first distance determining unit is configured to calculate the first distance between the reference bandwidth data and bandwidth data of other historical periods corresponding to the second distance if the second distance is less than or equal to a first preset threshold.

[0169] In an embodiment of the present invention, the bandwidth data determination module of the target historical period includes:

[0170] A bandwidth data determination submodule for a candidate historical period, configured to determine bandwidth data for a candidate historical period corresponding to a first distance that is less than or equal to a second preset threshold;

[0171] A sorting submodule, used to sort the bandwidth data of the candidate historical periods in order of the first distance from small to large;

[0172] The target historical period bandwidth data determination submodule is used to select a preset number of candidate historical period bandwidth data in the order as the target historical period bandwidth data.

[0173] In an embodiment of the present invention, the target increment information determination module includes:

[0174] A first estimated incremental value determining submodule, configured to determine a first estimated incremental value according to the estimated bandwidth data of the next period;

[0175] A second estimated incremental value acquisition submodule, used to acquire a second estimated incremental value;

[0176] The target increment information determination submodule is used to determine the target increment information of the next cycle according to the first estimated increment value and the second estimated increment value.

[0177] In an embodiment of the present invention, the target increment information determination submodule includes:

[0178] A data acquisition unit, used to obtain an estimated accuracy value and a business risk value;

[0179] A target incremental information determination unit is used to determine the target incremental information of the next cycle according to the estimated accuracy value, the business risk value, the first estimated incremental value and the second estimated incremental value.

[0180] In an embodiment of the present invention, the device further includes:

[0181] an interpolation processing module, for determining, for bandwidth data of each historical period, that the bandwidth at the sampling moment is abnormal if the bandwidth at the sampling moment is less than a third preset threshold, the bandwidth at the previous sampling moment is greater than or equal to the third preset threshold, and the bandwidth at the next sampling moment is greater than or equal to the third preset threshold, and interpolating the bandwidth at the previous sampling moment and the bandwidth at the next sampling moment to determine the bandwidth at the sampling moment;

[0182] A smoothing processing module is used to determine that the bandwidth at the sampling moment is a sudden drop value, and smooth the bandwidth at the sampling moment if the bandwidth at the sampling moment is greater than a fourth preset threshold, and the ratio of the difference between the bandwidth at the sampling moment and the bandwidth at the previous sampling moment to the bandwidth at the previous sampling moment is greater than the fourth preset threshold.

[0183] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0184] An embodiment of the present invention further provides an electronic device, including:

[0185] The invention comprises a processor, a memory and a computer program stored in the memory and capable of running on the processor. When the computer program is executed by the processor, each process of the method embodiment for predicting service increment of the content distribution network is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0186] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, each process of the method embodiment for predicting service increments in the above-mentioned content distribution network is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0187] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0188] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0189] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0190] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0191] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0192] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional 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 that fall within the scope of the embodiments of the present invention.

[0193] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0194] The above is a detailed introduction to the method, device, equipment and medium for predicting business increments in a content distribution network provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for predicting service increment of a content distribution network, characterized in that: The method comprises: Obtain bandwidth data of multiple historical periods of the target service; wherein the bandwidth data of each historical period includes bandwidth data of multiple sampling moments collected according to a preset sampling interval; According to the bandwidth data of the multiple historical periods, the bandwidth data of the most recent historical period is selected as the reference bandwidth data; Determine the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data; Determining bandwidth data of a preset number of target historical periods according to the similarity; Taking a weighted average of the bandwidth data of the target historical period to determine the estimated bandwidth data of the next period; The target increment information of the next cycle is determined according to the estimated bandwidth data of the next cycle.

2. The method according to claim 1, characterized in that: The determining of the similarity between the reference bandwidth data and bandwidth data of other historical periods other than the reference bandwidth data includes: Calculating a first distance between the reference bandwidth data and the bandwidth data of other historical periods according to a dynamic time warping function; the first distance is inversely correlated with the similarity; The similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data is determined according to the first distance.

3. The method according to claim 2, characterized in that The calculating, according to the dynamic time warping function, a first distance between the reference bandwidth data and the bandwidth data of other historical periods comprises: Determining sub-bandwidth data of the reference bandwidth data; Calculate, according to the dynamic time warping function, a second distance between the sub-bandwidth data of the reference bandwidth data and the bandwidth data of the other historical periods; If the second distance is less than or equal to the first preset threshold, the first distance between the reference bandwidth data and bandwidth data of other historical periods corresponding to the second distance is calculated.

4. The method according to claim 2, characterized in that: Determining bandwidth data of a preset number of target historical periods according to the similarity includes: Determine bandwidth data of a candidate historical period corresponding to a first distance that is less than or equal to a second preset threshold; Sorting the bandwidth data of the candidate historical periods in ascending order of the first distances; A preset number of candidate historical period bandwidth data are selected in the order as the target historical period bandwidth data.

5. The method according to claim 1, characterized in that The determining, according to the estimated bandwidth data of the next period, target increment information of the next period includes: Determining a first estimated incremental value according to the estimated bandwidth data of the next cycle; Obtaining a second estimated increment value; The target increment information of the next cycle is determined according to the first estimated increment value and the second estimated increment value.

6. The method according to claim 5, characterized in that The determining the target increment information of the next cycle according to the first estimated increment value and the second estimated increment value includes: Obtain estimated accuracy and business risk values; The target increment information of the next cycle is determined according to the estimated accuracy value, the business risk value, the first estimated increment value and the second estimated increment value.

7. The method according to claim 1, characterized in that Before selecting bandwidth data of the most recent historical period as reference bandwidth data based on the bandwidth data of the multiple historical periods, the method further includes: For the bandwidth data of each historical period, if the bandwidth at the sampling moment is less than a third preset threshold, the bandwidth at a sampling moment before the sampling moment is greater than or equal to the third preset threshold, and the bandwidth at a sampling moment after the sampling moment is greater than or equal to the third preset threshold, it is determined that the bandwidth at the sampling moment is abnormal, and the bandwidth at the sampling moment is determined by interpolation based on the bandwidth at the previous sampling moment and the bandwidth at the next sampling moment; If the bandwidth at the sampling moment is greater than a fourth preset threshold, and the ratio of the difference between the bandwidth at the sampling moment and the bandwidth at the previous sampling moment to the bandwidth at the previous sampling moment is greater than the fourth preset threshold, the bandwidth at the sampling moment is determined to be a sudden drop value, and the bandwidth at the sampling moment is smoothed.

8. A service increment prediction device for a content distribution network, characterized in that: The device comprises: A bandwidth data acquisition module, used to acquire bandwidth data of multiple historical periods of the target service; wherein the bandwidth data of each historical period includes bandwidth data of multiple sampling moments collected according to a preset sampling interval; A reference bandwidth data determination module, configured to select bandwidth data of the most recent historical period as reference bandwidth data based on the bandwidth data of the multiple historical periods; A similarity determination module, used to determine the similarity between the reference bandwidth data and bandwidth data of other historical periods except the reference bandwidth data; A target historical period bandwidth data determination module, used to determine a preset number of target historical period bandwidth data according to the similarity; An estimated bandwidth data determination module, used to perform weighted averaging on the bandwidth data of the target historical period to determine the estimated bandwidth data of the next period; The target increment information determination module is used to determine the target increment information of the next cycle according to the estimated bandwidth data of the next cycle.

9. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of the method for predicting service increment of a content distribution network as claimed in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for predicting service increment of a content distribution network as claimed in any one of claims 1 to 7 are implemented.

Images