A refined management method, system, device and medium for shared bandwidth packages

Abstract

The present invention relates to the field of Internet technology, and in particular to a refined management method, system, device and medium for a shared bandwidth package, wherein the method specifically comprises: constructing an IP node and a business association map based on a multi-dimensional resource usage data set, analyzing the IP node usage based on the IP node and business association map using a graph traversal algorithm, identifying IP nodes that have not been used for a long time or have not been bound to a business, and alerting operation and maintenance personnel; obtaining the bandwidth usage of each IP in each preset time period, and determining the 95 value of each IP by sorting the bandwidth usage of each IP in each preset time period from large to small, and determining the billing amount of each IP according to the ratio of the 95 value of each IP and the billing amount of the shared bandwidth package, and determining the billing amount of each IP by the billing amount of each IP and the IP node and business association map. The present invention does not require human intervention and meets the requirements of automatic and refined sharing of shared bandwidth costs.

Description

Technical Field

[0001] The present invention relates to the field of Internet technology, and in particular to a refined management method, system, device and medium for sharing bandwidth packages. Background Art

[0002] With the rapid development of cloud computing technology, the public network billing methods of domestic public cloud platforms are becoming increasingly diverse, including but not limited to 95% shared bandwidth package billing by region, 95% shared bandwidth package billing by project, and traffic-based billing, etc. These billing methods provide enterprises with a variety of options, but also bring management challenges, especially in terms of cost control, financial account splitting, and daily operation and maintenance.

[0003] The 95 shared bandwidth packages in each region are billed in a billing mode that aggregates all IPs in the same region. When the peak of public network traffic in a business is distributed in different time periods, bandwidth aggregation billing can be achieved through shared bandwidth packages, which greatly reduces public network costs. Businesses are added to shared bandwidth packages in batches for unified management, which supports accurate and timely verification of bandwidth usage, and can quickly locate businesses with abnormal bandwidth usage and risks of exceeding budgets. Therefore, from the perspective of reducing costs, billing by 95 shared bandwidth packages in each region becomes the best choice. Management can meet technical requirements, but solving the financial account splitting needs has become a major problem.

[0004] Although cloud companies have developed some apportionment settlement methods for customers to meet their needs, in a multi-cloud environment, each customer has their own apportionment requirements, and it is difficult for customers to drive cloud companies to provide corresponding settlement methods according to their own business demands. Currently, cloud companies provide ways to solve this apportionment demand: customers divide their business needs according to their own business, establish multiple shared bandwidth packages, and label each shared bandwidth package or create projects to distinguish costs. This method violates the purpose of using shared bandwidth packages to aggregate billing and reduce public network costs; labeling on the cloud relies too much on manual processing, which consumes a lot of manpower and is difficult to ensure accuracy; overly detailed labeling of business purposes on the cloud also brings certain information security issues to enterprises. Summary of the invention

[0005] The purpose of the present invention is to provide a method, system, device and medium for fine-grained management of shared bandwidth packages, which automatically allocates the shared bandwidth cost in each billing period without human intervention, meets the requirements of automatic and fine-grained cost allocation, and solves at least one of the above-mentioned existing technical problems.

[0006] In a first aspect, the present invention provides a method for fine-grained management of a shared bandwidth package, the method specifically comprising:

[0007] Collect IP node usage data, bandwidth flow data, and business information from multiple cloud platforms, align multi-source data based on timestamps, and form a multi-dimensional resource usage data set;

[0008] Constructing an IP node and business association graph based on the multi-dimensional resource usage data set, analyzing IP node usage using a graph traversal algorithm based on the IP node and business association graph, identifying IP nodes that have not been used for a long time or are not bound to a business, and alerting operation and maintenance personnel;

[0009] Obtain the bandwidth usage of each IP in each preset time period according to the bandwidth flow data, determine the 95th value of each IP by sorting the bandwidth usage of each IP in each preset time period from large to small, determine the billing amount of each IP according to the ratio of the 95th value of each IP and the billing amount of the shared bandwidth package, and determine the billing amount of each service according to the billing amount of each IP and the IP node and service association map;

[0010] The bandwidth usage history data is obtained according to the bandwidth flow data, the bandwidth usage history data is clustered and analyzed by a clustering algorithm to determine the bandwidth usage pattern, and an autoregressive moving average model is used to establish a bandwidth usage baseline model, and an abnormal bandwidth fluctuation alarm is implemented according to the bandwidth usage baseline model.

[0011] In a second aspect, the present invention provides a refined management system for shared bandwidth packages, the system specifically comprising:

[0012] The first management module is used to collect IP node usage data, bandwidth flow data and business information from multiple cloud platforms, align multi-source data according to timestamps, and form a multi-dimensional resource usage data set;

[0013] A second management module is used to construct an IP node and business association map based on the multi-dimensional resource usage data set, and analyze the IP node usage using a graph traversal algorithm based on the IP node and business association map, identify IP nodes that have not been used for a long time or are not bound to a business, and issue an alarm to operation and maintenance personnel;

[0014] A third management module is used to obtain the bandwidth usage of each IP in each preset time period according to the bandwidth flow data, determine the 95th value of each IP by sorting the bandwidth usage of each IP in each preset time period from large to small, determine the billing amount of each IP according to the ratio of the 95th value of each IP and the billing amount of the shared bandwidth package, and determine the billing amount of each service according to the billing amount of each IP and the IP node and service association map;

[0015] The fourth management module is used to obtain bandwidth usage history data based on the bandwidth flow data, perform cluster analysis on the bandwidth usage history data through a clustering algorithm, determine the bandwidth usage pattern, and use an autoregressive moving average model to establish a bandwidth usage baseline model, and implement abnormal bandwidth fluctuation alarm based on the bandwidth usage baseline model.

[0016] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor and a computer program stored in the memory, and when the computer program is executed on the processor, a refined management method for shared bandwidth packages as described in any one of the above methods is implemented.

[0017] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for refined management of shared bandwidth packages as described in any one of the above methods is implemented.

[0018] Compared with the prior art, the present invention has at least one of the following technical effects:

[0019] 1. The shared bandwidth cost is automatically allocated in each billing cycle without human intervention, meeting the requirements of automatic and refined cost allocation.

[0020] 2. It can effectively optimize the allocation and use of IP nodes, improve resource utilization efficiency, reduce unnecessary expenses, and enhance network stability and security.

[0021] 3. Through intelligent analysis of bandwidth usage patterns, potential bandwidth problems can be warned in advance, helping the operation and maintenance team to respond faster.

[0022] 4. By constructing a graph of IP nodes and business associations and using a graph traversal algorithm, it is possible to accurately identify IP nodes that have not been used for a long time or are not bound to a business, and promptly issue warnings to operation and maintenance personnel, effectively avoiding waste and idleness of resources, improving resource utilization, and reducing operation and maintenance costs caused by improper resource allocation.

[0023] 5. Through clustering algorithms and autoregressive moving average models, a bandwidth usage baseline model can be established and bandwidth usage can be monitored in real time. Once abnormal bandwidth fluctuations are detected, an alarm will be immediately sent to the operation and maintenance personnel, so that they can respond and solve the problem quickly, ensuring the stable operation of the business. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 It is a flowchart of a refined management method for a shared bandwidth package provided by the first embodiment of the present invention;

[0026] Figure 2 It is a flow chart of a refined management method of a shared bandwidth package provided by the second embodiment of the present invention;

[0027] Figure 3 This is an application example diagram of a refined management method for a shared bandwidth package provided by an embodiment of the present invention;

[0028] Figure 4 It is a structural diagram of a refined management system for shared bandwidth packages provided by an embodiment of the present invention;

[0029] Figure 5 It is a structural schematic diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0031] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0032] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0033] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0034] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0035] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0036] In the embodiment of the present application, the execution subject of the process includes a terminal device, which includes but is not limited to: a server, a computer, a smart phone, a tablet computer, and other devices capable of executing the method disclosed in the present application. Figure 1 A schematic diagram of a process flow of a refined management method for a shared bandwidth package disclosed in the first embodiment of the present invention is shown, and the details are as follows:

[0037] S101, collect IP node usage data, bandwidth flow data and business information from multiple cloud platforms, align multi-source data according to timestamps, and form a multi-dimensional resource usage data set.

[0038] In this embodiment, the solution is regenerated: IP node usage data, bandwidth flow data and business information are collected from multiple cloud platforms, the data format is unified through data normalization, and the multi-source heterogeneous data is aligned according to the timestamp to obtain a standardized multi-dimensional resource usage data set. When collecting data, 5 minutes is set as the basic time granularity to ensure the consistency of subsequent analysis. According to the collected original multi-source heterogeneous data, data cleaning technology is used to remove noise data, and unstructured data is converted into structured data through data conversion. Then, data integration technology is used to integrate the cleaned and converted data in the same format to eliminate semantic ambiguity. The integrated data is subjected to data normalization processing, the measurement unit is unified, the numerical data is normalized to a unified dimension, and the non-numerical data is mapped to numerical data. Normalized data in a unified format is obtained through data normalization. According to the timestamp information in the data, the multi-dimensional data from different sources are aligned on the time axis using data alignment technology, and the data resampling alignment is performed with 5 minutes as the basic unit. If some dimensional data is missing after alignment, the data filling technology is used to complete the data. Through the above data preprocessing operations, a standardized multi-dimensional resource usage data set aligned with the time axis is finally obtained. To ensure data quality, the isolation forest algorithm was used to detect outliers in the dataset and replace them with the median.

[0039] For example, IP node usage data, bandwidth flow data and business information are collected from multiple cloud platforms such as Alibaba Cloud and Tencent Cloud, and the data format is unified through data normalization processing, such as converting the flow data into MB units and unifying the time format into yyyy-MM-dd HH:mm:ss. Multi-source heterogeneous data are aligned according to timestamps, and the data is resampled and aligned with a time granularity of 5 minutes. For example, the data of a certain cloud platform is in a granularity of 1 minute, and it is aligned with the data of 5-minute granularity through resampling. Data cleaning techniques such as outlier detection are used to remove noise data, such as using an isolated forest algorithm to identify and remove abnormal peaks in bandwidth flow data. Unstructured data such as business types are converted into structured data through data conversion such as one-hot encoding. Data integration technology is used to integrate the cleaned and converted data in a unified JSON format to eliminate semantic ambiguity between different cloud platforms. Data normalization is performed on the integrated data, such as normalizing the IP usage rate from a percentage to a numerical range of 0-1, and mapping non-numeric resource types to integer encoding. The normalized data was resampled and aligned according to the 5-minute time granularity. For some dimensional data missing, the data was supplemented by forward filling, and finally a standardized multi-dimensional resource usage dataset with time axis alignment was obtained. To ensure data quality, the isolation forest algorithm was used to detect outliers in the dataset, and the outliers were replaced with the median to remove abnormal noise.

[0040] S102, constructing an IP node and business association graph based on the multi-dimensional resource usage data set, analyzing IP node usage using a graph traversal algorithm based on the IP node and business association graph, identifying IP nodes that have not been used for a long time or have not been bound to a business, and alerting operation and maintenance personnel.

[0041] In this embodiment, a graph database is used to store IP addresses, service identifiers, usage time node information and their relationships. Through the nodes and edges in the graph database, the IP address, service identifier, usage time node and other information are associated to form a complete IP resource and service association graph. Next, the graph traversal algorithm is used to analyze the usage of IP resources. According to the constructed IP resource and service association graph, the graph is traversed using a breadth-first search algorithm or a depth-first search algorithm. During the traversal process, relevant information of each IP address node is obtained, including the associated service identifier, the last usage time, etc. By analyzing this information, it is determined whether the IP address has not been used for a long time or is not bound to a service. If the last usage time of an IP address node exceeds a preset threshold from the current time, it is marked as an IP resource that has not been used for a long time. Similarly, if an IP address node has no associated service identifier information, it is marked as an IP resource that is not bound to a service. Finally, the identified IP resources that have not been used for a long time or are not bound to a service are summarized and counted. A data structure such as a hash table is used to store this IP resource information, and the results are output in a certain format. By analyzing the identification results, we can obtain the current utilization of IP resources and provide data support for subsequent IP resource optimization and recovery work.

[0042] Exemplarily, in the process of constructing the IP resource and business association graph, first, the IP address, business identifier, and usage time node information are input into the graph database. For example, an IP address "19161" is assigned to the financial system on March 1, 2023, and this information will be stored in the graph database as nodes and edges. In this way, a multi-dimensional association network can be formed, in which each node represents an IP address or business identifier, and each edge represents the association relationship between them, such as a usage relationship or an allocation relationship. Next, the graph traversal algorithm is used to analyze the usage of IP resources. Using the breadth-first search (BFS) algorithm, starting from a randomly selected unanalyzed IP address node, all nodes connected to it are traversed layer by layer, and the last usage time and associated business identifier of each node are recorded. For example, starting from the IP address "19161", check all its directly connected nodes, such as the financial system, and then check other IP addresses connected to the financial system, and so on. This method can effectively reveal which IP addresses have not been used for a long time or are not associated with any business. By setting a threshold, such as 180 days, IP addresses that have not been used for more than this period can be identified. If the last use time of an IP address is more than 180 days from the current date, or the IP address is not associated with any business identifier, it will be marked as an IP resource that has not been used for a long time or is not bound to a business. This information is then stored in a hash table, where the key is the IP address and the value is the usage status and the last use time of the IP address. Finally, by analyzing the data in the hash table, a detailed report on the utilization of IP resources can be generated, which includes a list of all IP resources marked as unused for a long time or unbound to a business, as well as their specific information. Such analysis results provide scientific data support for the optimization and recovery of IP resources, helping enterprises to manage and configure their network resources more efficiently.

[0043] In some embodiments, in the above step S102, constructing an IP node and business association graph based on the multi-dimensional resource usage data set specifically includes:

[0044] Map each IP node to a two-dimensional vector space, and calculate the similarity between every two IP nodes using cosine similarity based on the vector of each mapped IP node to obtain a first similarity matrix;

[0045] Obtaining IP node clustering results by performing spectral clustering on the first similarity matrix;

[0046] Based on the multi-dimensional resource usage data set and according to the IP node clustering result, the association rules between IP nodes and services are mined through an association analysis algorithm to form a bipartite graph of IP nodes and services;

[0047] A low-dimensional vector representation of the bipartite graph is obtained by analyzing the heterogeneous graph representation learning algorithm, and the semantic similarity between each IP node and each service is calculated according to the low-dimensional vector representation to form a second similarity matrix;

[0048] The second similarity matrix is ​​used to perform bipartite graph clustering to obtain the association clustering results between IP nodes and services, and an IP node and service association graph is constructed based on the association clustering results.

[0049] In this embodiment, according to the characteristic attribute information of the IP node, a suitable vectorization method is used to map each IP node into a two-dimensional vector space to obtain a two-dimensional vector representation corresponding to each IP node. By calculating the cosine similarity between each two IP node vectors, a first similarity matrix is ​​obtained, which represents the similarity between the IP nodes. According to the obtained first similarity matrix, a spectral clustering algorithm is used to perform cluster analysis on the IP nodes, and a low-dimensional representation of the IP nodes is obtained by performing feature decomposition and dimensionality reduction processing on the similarity matrix. According to the low-dimensional representation of the IP nodes, a clustering algorithm such as K-means is used to cluster the IP nodes, and the final IP node clustering result is obtained by iteratively optimizing the cluster center and adjusting the cluster label.

[0050] According to the multi-dimensional resource usage data set, a clustering algorithm is used to perform cluster analysis on IP nodes, obtain the clustering results of IP nodes, and obtain IP node groups with similar resource usage patterns. Through the association analysis algorithm, the IP node clustering results are used to mine the association rules between IP nodes and services, obtain the association strength between IP nodes and services, and form a bipartite graph of IP nodes and services to represent the association relationship between IP nodes and services. A heterogeneous graph representation learning algorithm is used to analyze the bipartite graph of IP nodes and services, obtain the low-dimensional vector representation of the bipartite graph, map IP nodes and services to the same semantic space, and capture the potential semantic relationship between them. According to the low-dimensional vector representation, the semantic similarity between each IP node and each service is calculated. Through measurement methods such as cosine similarity, the similarity between IP nodes and services in the semantic space is quantified to form a second similarity matrix to represent the semantic correlation between IP nodes and services.

[0051] According to the obtained second similarity matrix, the spectral clustering algorithm is used to jointly cluster IP nodes and business nodes. By performing spectral decomposition on the similarity matrix, the associated clustering results of IP nodes and businesses are obtained, revealing the inherent association between IP and business. According to the associated clustering results of IP nodes and businesses, the graph construction algorithm is used to generate the association map of IP nodes and businesses, and the association relationship between IP and business is intuitively displayed in a visual way, so as to obtain a business association map that is easy to analyze and make decisions.

[0052] Exemplarily, each IP node is first mapped to a two-dimensional vector space, for example, the first two bytes of the IP address can be used as the horizontal coordinate and the last two bytes as the vertical coordinate. Then the cosine similarity is used to calculate the similarity between each two IP nodes to form a first similarity matrix. For example, the cosine similarity between IP1 (19161) and IP2 (19161) is 95, and the cosine similarity between IP1 and IP3 (11) is 1. Then, the first similarity matrix is ​​subjected to spectral clustering, and the K-means algorithm can be used to set the number of clusters K=3 to obtain the clustering results of the IP nodes, such as IP1 and IP2 belong to the same class, and IP3 belongs to a separate class. Then, based on multi-dimensional resource usage data sets, such as CPU usage, memory usage, etc., the association rules between IP nodes and services are mined through the Apriori association analysis algorithm, such as the association between IP1 and service A is 80%, and the association between IP2 and service B is 90%, forming a bipartite graph of IP nodes and services. A heterogeneous graph representation learning algorithm such as Metapath2vec is used to analyze the bipartite graph to obtain a 128-dimensional low-dimensional vector representation. The semantic similarity between IP nodes and services is calculated based on the vector representation to form a second similarity matrix. The second similarity matrix is ​​used to cluster the bipartite graph. The spectral clustering algorithm can be used to set the number of clusters to 4 to obtain the clustering results of the association between IP nodes and services, such as IP1 and services A and B are clustered into one category, IP2 and service C are clustered into one category, etc., and the IP node and service association map is constructed accordingly.

[0053] In some embodiments, in the above step S102, based on the IP node and service association graph, the use of the graph traversal algorithm is used to analyze the IP node usage, identify IP nodes that have not been used for a long time or are not bound to a service, and alert the operation and maintenance personnel, specifically including:

[0054] Based on the IP node and business association graph, a graph traversal algorithm is used to traverse each IP node in the graph to obtain the last use timestamp of each IP node and a list of business nodes associated with it;

[0055] According to the last use timestamp of the current IP node, determine whether the current IP node is a long-term unused resource. If the last use timestamp is more than the preset time threshold from the current time, mark the current IP node as a long-term unused resource.

[0056] According to the service node list associated with the current IP node, determine whether the current IP node is an unbound service resource. If the service node list associated with the current IP node is empty, mark the current IP node as an unbound service resource.

[0057] Acquire the attribute information of the IP nodes marked as long-term unused resources or unbound service resources, generate alarm information according to the attribute information, and send the alarm information to the operation and maintenance personnel.

[0058] In this embodiment, according to the IP node and business association graph, a graph traversal algorithm is used to traverse each IP node in the graph, obtain the last use timestamp of each IP node and the business node list associated therewith, and obtain the usage status and associated business information of the IP node. By traversing the list, it is determined whether the list is empty. If the business node list associated with the current IP node is empty, it is determined that the IP node is not bound to any business resources. According to the judgment result of the previous step, if the IP node is not bound to a business resource, the IP node is marked as an "unbound business resource" state. The "unbound business resource" state obtained in the previous step is synchronized to the relevant attributes of the IP node by updating the IP node attributes, so as to be identified and processed in the subsequent business logic. The IP nodes that have been judged to be unused for a long time and unbound to the business are marked. The alarm generation algorithm is used to generate corresponding alarm information according to the attribute information of the marked IP node. The alarm information includes the IP address, the long-term unused time or the unbound business, etc. The generated alarm information is sent to the relevant operation and maintenance personnel through the alarm sending module, prompting them to deal with these IP nodes that have not been used for a long time or have not been bound to services in a timely manner to avoid wasting resources. Based on the received IP node alarm information, the operation and maintenance personnel will check and handle the IP nodes that have not been used for a long time or have not been bound to services, and determine the usage status and existing problems of these IP nodes. For IP nodes that have been confirmed to be useless and have not been used for a long time, the operation and maintenance personnel will perform recycling operations to release them for use by new services and improve the utilization rate of IP resources. For IP nodes that are not bound to services, the operation and maintenance personnel will analyze the reasons, recycle the useless IP nodes, and supplement the binding of the IP nodes that have been missed to ensure the accurate association between the IP nodes and the services. After the recycling and binding processing is completed, the operation and maintenance personnel will synchronously update the latest status of the IP node to the IP attribute information library to ensure the accuracy and timeliness of the IP node attribute information, and facilitate the subsequent management and use of IP nodes.

[0059] S103, obtaining the bandwidth usage of each IP in each preset time period according to the bandwidth flow data, determining the 95 value of each IP by sorting the bandwidth usage of each IP in each preset time period from large to small, determining the billing amount of each IP according to the ratio of the 95 value of each IP and the billing amount of the shared bandwidth package, and determining the billing amount of each service according to the billing amount of each IP and the IP node and service association map.

[0060] In this embodiment, the bandwidth usage data of each IP in each preset time period is obtained, and the bandwidth usage data of each IP in each preset time period is sorted from large to small using a quick sort algorithm. According to the sorted bandwidth usage data, the 95th percentile value of each IP is determined by the quantile algorithm to obtain the 95th value bandwidth usage of each IP. According to the billing amount of the shared bandwidth package and the ratio of the 95th value bandwidth usage of each IP to the total 95th value bandwidth usage of all IPs, the weighted average algorithm is used to calculate the billing amount of each IP. Finally, a statistical chart of the billing amount of each IP is generated by data visualization technology to intuitively display the billing situation of each IP, which is convenient for user analysis and management. At the same time, the billing details data of each IP are stored in the database for subsequent query and analysis.

[0061] For example, suppose A shares a bandwidth package, and the fee after the monthly 95% fee is 100 yuan; there are 5 IPs in the bandwidth package, the 95 value of IP1 is equal to x1, and the service path to find the host is "Game 1", and the fee y1 = x1 / (x1+x2+x3+x4+x5)*100; the 95 value of IP2 is equal to x2, and the service path to find the host is "Game 2", and the fee y2 = x2 / (x1+x2+x3+x4+x5)*100; the 95 value of IP3 is equal to x3, and the service path to find the host is "Game 1", and the fee y3 = x3 / (x1+x2+x3+x4+x5)*100; the 95 value of IP4 is equal to x4, the service path to find the host is "Game 1", and the cost y4=x4 / (x1+x2+x3+x4+x5)*100; the 95 value of IP5 is equal to x5, the service path to find the host is "Game 2", and the cost y5=x5 / (x1+x2+x3+x4+x5)*100; then the bandwidth cost borne by "Game 1" is (y1+y3+y4), and the cost borne by "Game 2" is (y2+y5).

[0062] In some embodiments, in the above step S103, the 95 value of each IP is determined by sorting the bandwidth usage of each IP in each preset time period from large to small, and the billing amount of each IP is determined according to the ratio of the 95 value of each IP and the billing amount of the shared bandwidth package, specifically including:

[0063] A quick sorting algorithm is used to sort the bandwidth usage data of each IP in each preset time period from large to small to obtain sorted bandwidth usage data;

[0064] According to the sorted bandwidth usage data, the 95th percentile value of each IP is determined by a quantile algorithm to obtain the 95th value bandwidth usage of each IP;

[0065] The billing amount for each IP is calculated using a weighted average algorithm based on the billing amount of the shared bandwidth package and the ratio of each IP's 95% bandwidth usage to the total 95% bandwidth usage of all IPs.

[0066] If the billing amount of any IP is less than 0, its billing amount will be set to 0, and the difference in its billing amount will be proportionally allocated to the billing amounts of other IPs, and the billing amounts of other IPs will be recalculated until the billing amounts of all IPs are greater than or equal to 0.

[0067] In this embodiment, according to the bandwidth usage data of each IP in each preset time period, the quick sorting algorithm is used to sort these bandwidth usage data from large to small to obtain the sorted bandwidth usage data, which reflects the bandwidth usage of each IP in different time periods. Through the quantile algorithm, the 95th percentile value of each IP is determined according to the sorted bandwidth usage data, and the 95th value bandwidth usage of each IP is obtained. This value represents the upper limit of the bandwidth usage of the IP in 95% of the time. According to the billing amount of the shared bandwidth package and the ratio of the 95th value bandwidth usage of each IP to the total 95th value bandwidth usage of all IPs, the weighted average algorithm is used to calculate the billing amount of each IP, and the fee that each IP should bear in the shared bandwidth package is obtained. This fee is related to the ratio of the bandwidth usage of the IP to the bandwidth usage of all IPs. Get the calculated fees for each IP, and add up these fees to determine whether the total accumulated fees are equal to the billing amount of the shared bandwidth package. If they are equal, the billing amount for each IP is accurate. If they are not equal, adjust the weights in the weighted average algorithm and recalculate the billing amount for each IP until the total fee is equal to the billing amount of the shared bandwidth package. Generate billing data based on the billing amount of each IP. This billing data includes information such as IP address, billing amount, and billing cycle. Through this billing data, customers can be shown the cost of each IP in the shared bandwidth package, making it easier for customers to understand and manage their bandwidth usage and expense expenditures.

[0068] Get the billing amount data of all IPs and store it in a list, recorded as ip_amounts. By traversing the ip_amounts list, determine whether the billing amount of each IP is less than 0. If there is an IP with a billing amount less than 0, set its billing amount to 0 and calculate its difference, recorded as diff. According to the number of IPs with billing amounts greater than or equal to 0 in the ip_amounts list, calculate the difference amount that each IP needs to share, that is, divide diff by the remaining number of IPs to obtain the amount share that each IP needs to share. By traversing the ip_amounts list again, add share to the billing amount of each IP with a billing amount greater than or equal to 0, and update its billing amount. Repeat steps 2 to 4 until the billing amount of all IPs is greater than or equal to 0. Obtain the final ip_amounts list, which contains the recalculated billing amount of each IP, ensuring that the billing amount of all IPs is non-negative.

[0069] S104, obtaining bandwidth usage history data according to the bandwidth flow data, performing cluster analysis on the bandwidth usage history data through a clustering algorithm to determine the bandwidth usage pattern, and establishing a bandwidth usage baseline model using an autoregressive moving average model, and implementing an abnormal bandwidth fluctuation alarm according to the bandwidth usage baseline model.

[0070] In this embodiment, bandwidth usage history data is obtained according to the bandwidth flow data, and the bandwidth usage history data is clustered and analyzed by the K-means clustering algorithm, and the bandwidth usage pattern is determined according to the results of the cluster analysis. The bandwidth usage baseline model is established by using the autoregressive moving average model. The autoregressive moving average model takes the bandwidth usage history data as input, and obtains the bandwidth usage baseline by the autoregressive and moving average method. According to the bandwidth usage baseline model, it is judged whether the current bandwidth usage data exceeds the range predicted by the baseline model. If the current bandwidth usage data exceeds the range predicted by the bandwidth usage baseline model, it is judged as abnormal bandwidth fluctuation and an alarm is triggered. When judging the abnormal bandwidth fluctuation, the current bandwidth usage data is compared with the threshold by setting a threshold. If the current bandwidth usage data exceeds the threshold, it is judged as an abnormal fluctuation. The setting of the threshold needs to comprehensively consider the prediction results of the bandwidth usage baseline model and business needs to avoid too many false positives and false negatives. After the alarm is triggered, the system automatically generates an abnormal report, which contains information such as the time when the abnormality occurred, the duration of the abnormality, and the amplitude of the abnormal fluctuation, providing data support for the operation and maintenance personnel to analyze and handle the abnormality. At the same time, the alarm information is promptly notified to relevant personnel via SMS, email, etc., so that they can intervene and handle it at the first time and minimize the impact of the abnormality.

[0071] For example, first obtain the bandwidth usage data of the last month, collect it every 5 minutes, and a total of 8640 data points. Use the K-means clustering algorithm, set the number of clusters k=3, calculate the Euclidean distance from each data point to the cluster center, assign the data point to the cluster with the closest distance, and update the cluster center. Continue to iterate until the cluster center no longer changes, and finally get 3 clusters, representing low, medium, and high bandwidth usage modes. According to the clustering results, select the class with the highest cluster center value as the high bandwidth usage mode, and extract the data points under this mode as training data, and use the autoregressive moving average model ARMA (p, q) for modeling, where p=2, q=1. After the model training is completed, use the 120 data points before the current moment as input to predict the bandwidth usage baseline value at the current moment. If the current actual bandwidth usage value exceeds 5 times the baseline value, it is determined to be an abnormal fluctuation and trigger an alarm. The alarm information includes the time of abnormal occurrence, the duration of the abnormality, the amplitude of abnormal fluctuation, etc., and the operation and maintenance personnel are notified by email and SMS. For example, the ARMA model predicts that the current bandwidth usage baseline value is 80Mbps, while the actual monitoring value is 150Mbps, which is 5 times higher than the baseline value. It is considered an abnormal fluctuation and triggers an alarm. The alarm information shows that the abnormality occurred at 14:30 on April 20, 2023, lasted for 10 minutes, and the abnormal fluctuation range was 85%. After receiving the alarm, the operation and maintenance personnel will check the abnormal report in time, analyze the cause of the abnormality, and take corresponding measures, such as limiting the bandwidth usage of non-critical services to ensure the normal operation of critical services until the bandwidth usage returns to normal.

[0072] In some embodiments, in the above step S104, clustering analysis is performed on the bandwidth usage history data by using a clustering algorithm to determine the bandwidth usage pattern, and an autoregressive moving average model is used to establish a bandwidth usage baseline model, and an abnormal bandwidth fluctuation alarm is implemented according to the bandwidth usage baseline model, specifically including:

[0073] Performing cluster analysis on the bandwidth usage history data according to a K-means clustering algorithm, calculating the Euclidean distance from each data point to the cluster center, assigning the data point to the cluster with the closest distance, and updating the cluster center, iterating continuously until the cluster center no longer changes, and obtaining a bandwidth usage pattern clustering result, wherein the bandwidth usage pattern clustering result includes high bandwidth usage pattern data, medium bandwidth usage pattern data, and low bandwidth usage pattern data;

[0074] Inputting the high bandwidth usage pattern data, the medium bandwidth usage pattern data and the low bandwidth usage pattern data into an autoregressive moving average model for modeling training respectively, to obtain a first bandwidth usage baseline model, a second bandwidth usage baseline model and a third bandwidth usage baseline model;

[0075] According to the bandwidth demand values ​​corresponding to different levels of services, the first bandwidth usage baseline model, the second bandwidth usage baseline model or the third bandwidth usage baseline model is used to determine whether the current bandwidth usage data exceeds the normal bandwidth fluctuation range. If it exceeds, it is determined to be a bandwidth abnormal state and an alarm is sent to the operation and maintenance personnel.

[0076] In this embodiment, based on the historical bandwidth usage data, the K-means clustering algorithm is used to perform cluster analysis on the data, and the data is divided into three categories: high-bandwidth usage pattern data, medium-bandwidth usage pattern data, and low-bandwidth usage pattern data. By calculating the Euclidean distance from each bandwidth usage data point to the cluster center, the data point is assigned to the cluster where the nearest cluster center is located, and the initial clustering result is obtained. According to the current clustering result, the center point of each cluster is recalculated to obtain the updated cluster center. Using the updated cluster center, the above steps are repeated, and the clustering results are continuously iterated and optimized until the cluster center no longer changes, that is, the convergence condition is reached. The final bandwidth usage pattern clustering result is obtained, including three categories of high-bandwidth usage pattern data, medium-bandwidth usage pattern data, and low-bandwidth usage pattern data, and each category of data represents a typical bandwidth usage pattern.

[0077] According to the high bandwidth usage pattern data, an autoregressive moving average model is used to train the model. Through the model training process, a first bandwidth usage baseline model is obtained, which can reflect the bandwidth usage baseline under high bandwidth usage. According to the medium bandwidth usage pattern data, an autoregressive moving average model is used to train the model. Through the model training process, a second bandwidth usage baseline model is obtained, which can reflect the bandwidth usage baseline under medium bandwidth usage. According to the low bandwidth usage pattern data, an autoregressive moving average model is used to train the model. Through the model training process, a third bandwidth usage baseline model is obtained, which can reflect the bandwidth usage baseline under low bandwidth usage. According to the bandwidth demand values ​​corresponding to services of different levels, the level of the current service and the corresponding bandwidth demand threshold are obtained. The first bandwidth usage baseline model is used to obtain the first bandwidth usage baseline value through the actual bandwidth usage data of the current service. The second bandwidth usage baseline model is used to obtain the second bandwidth usage baseline value through the actual bandwidth usage data of the current service. The third bandwidth usage baseline model is used to obtain the third bandwidth usage baseline value through the actual bandwidth usage data of the current service. According to the bandwidth demand threshold corresponding to the service level, an appropriate bandwidth usage baseline value is determined as a judgment basis. Obtain the real-time bandwidth usage data of the current service, and compare it with the selected bandwidth usage baseline value to determine whether the current bandwidth usage exceeds the normal fluctuation range. If the current bandwidth usage data exceeds the selected bandwidth usage baseline value, it is determined to be a bandwidth abnormal state, and the priority of the alarm is determined according to the importance of the service level. Obtain the contact information of the operation and maintenance personnel, and send bandwidth abnormality alarm information to the operation and maintenance personnel through SMS, email or other means, including the service level, current bandwidth usage data and abnormality judgment results. Continuously monitor the real-time bandwidth usage of the service, and dynamically adjust the selected bandwidth usage baseline model according to the bandwidth demand thresholds of different service levels to ensure the accuracy and timeliness of abnormality judgment. Record detailed information of each bandwidth abnormality event, including service level, abnormal occurrence time, abnormal duration, abnormal bandwidth usage data, etc., for subsequent data analysis and optimization.

[0078] Reference Figure 2 , Figure 2 The following is a flow chart of a refined management method for a shared bandwidth package disclosed in the second embodiment of the present invention, which is described in detail as follows:

[0079] S201, obtaining historical bandwidth usage data, and performing statistical analysis on the bandwidth usage of each IP node using a weighted moving average method based on the historical bandwidth usage data to obtain an average bandwidth usage rate of each IP node.

[0080] In this embodiment, the historical bandwidth usage data of all IP nodes in the network, including the bandwidth usage rate at each time point, is obtained to form time series data. Based on the historical bandwidth usage data obtained, the weighted moving average method is used to perform statistical analysis on the bandwidth usage of each IP node. In the weighted moving average method, a time window is set to perform weighted average on the bandwidth usage rate within the window, where the closer the data is to the current moment, the greater the weight, and the farther the data is, the smaller the weight. Through the weighted moving average calculation, the average bandwidth usage rate of each IP node at the current moment is obtained as the statistical result of the bandwidth usage of the node.

[0081] S202: According to the historical bandwidth usage data, a least square method is used to fit the bandwidth usage trend of each IP node to obtain a bandwidth usage trend coefficient of each IP node.

[0082] In this embodiment, the bandwidth usage of each IP node is counted and analyzed based on the historical bandwidth usage data to obtain time series data reflecting the bandwidth usage trend of each node. The bandwidth usage trend time series data of each IP node is fitted using the least squares method, and the bandwidth usage trend characteristics of each node are determined by the function coefficient obtained by fitting. The bandwidth usage trend coefficients of all IP nodes are obtained, and the bandwidth usage trend of each node is determined to be rising, falling or stable based on the size and sign of the coefficients. By comparing the bandwidth usage trend coefficients of different IP nodes, the nodes with the fastest and slowest bandwidth usage growth in the entire network, as well as the nodes with stable bandwidth usage, are determined.

[0083] S203: Determine a resource utilization score of each IP node according to the average bandwidth utilization and the bandwidth utilization trend coefficient.

[0084] In this embodiment, according to the average bandwidth utilization rate of each IP node, a normalization processing method is used to map the average bandwidth utilization rate to between 0 and 1, and the average bandwidth utilization rate score of each IP node is obtained. According to the bandwidth utilization trend coefficient of each IP node, a normalization processing method is used to map the bandwidth utilization trend coefficient to between 0 and 1, and the bandwidth utilization trend score of each IP node is obtained. By weighted average, the average bandwidth utilization rate score and the bandwidth utilization trend score of each IP node are weighted and summed to obtain the resource utilization score of each IP node. The weight of the average bandwidth utilization rate score is 7, and the weight of the bandwidth utilization trend score is 3. According to the resource utilization score of each IP node, a sorting algorithm is used to arrange all IP nodes in descending order to determine the IP nodes with higher resource utilization scores. By setting a resource utilization score threshold, it is judged that the IP node with a resource utilization score higher than the threshold is a high resource utilization node, and it is added to the high resource utilization node set.

[0085] S204, performing cluster analysis on the resource utilization scores of all IP nodes using a Gaussian mixture model according to the resource utilization scores of all IP nodes, and obtaining the distribution of resource utilization scores of IP nodes of different categories.

[0086] In this embodiment, according to the resource utilization data of all IP nodes, a data preprocessing method is used to clean and standardize the data, and a standardized IP node resource utilization score data set is obtained. By extracting features from the standardized IP node resource utilization score data set, a feature vector that can characterize the resource utilization characteristics of the IP node is obtained. According to the extracted feature vector, a Gaussian mixture model is used to perform cluster analysis on the IP node resource utilization score, and the optimal clustering result is obtained by iteratively optimizing the parameters of the Gaussian mixture model, and the IP nodes are divided into different categories. According to the clustering results, the resource utilization score distribution of the IP nodes in each category is obtained, and the distribution characteristics of the resource utilization scores of IP nodes in different categories are obtained by calculating the mean, variance and other statistical indicators of each category. Visualization technology is used to display the resource utilization score distribution of IP nodes in different categories in the form of histograms, box plots, etc., intuitively presenting the differences and characteristics of resource utilization of IP nodes in each category, and providing decision support for subsequent IP node resource optimization and management.

[0087] S205: Determine the score value corresponding to the first trough of the resource utilization score distribution as a potential idle resource threshold, and mark the IP nodes whose resource utilization scores are lower than the potential idle resource threshold as potential idle resources.

[0088] In this embodiment, based on the resource utilization data of the IP nodes, a statistical analysis method is used to obtain the resource utilization score distribution of all IP nodes; by performing a trough analysis on the resource utilization score distribution, a score value corresponding to the first trough of the resource utilization score distribution is obtained; the score value corresponding to the first trough obtained is determined as a potential idle resource threshold as a basis for judging whether the IP node is a potential idle resource; the resource utilization score is compared with the potential idle resource threshold, and for IP nodes whose resource utilization scores are lower than the potential idle resource threshold, they are judged to be potential idle resources; based on the judgment result, the IP nodes whose resource utilization scores are lower than the potential idle resource threshold are marked as potential idle resources, and a list of IP nodes of potential idle resources is obtained.

[0089] Furthermore, the second embodiment further includes:

[0090] According to the business income and number of users of each IP node and whether each IP node is a potential idle resource, a business importance evaluation matrix is ​​established by using the hierarchical analysis method, and the business importance score of each IP node is determined by the business importance evaluation matrix;

[0091] The cost sharing weight is determined according to the ratio of the 95 value of each IP and the business importance score, and the billing amount of each IP is determined by the cost sharing weight and the billing amount of the shared bandwidth package.

[0092] In this embodiment, according to the business income data of each IP node, a data collection and analysis method is used to obtain the business income amount of each IP node, and the business income score of each IP node is obtained by normalizing the business income amount. According to the user number data of each IP node, a data collection and analysis method is used to obtain the number of users of each IP node, and the user number score of each IP node is obtained by normalizing the number of users. According to the data of whether each IP node is a potential idle resource, a data collection and analysis method is used to obtain the idle resource status of each IP node, and the idle resource score of each IP node is obtained by quantifying the idle resource status. According to the business income score, the user number score and the idle resource score, the hierarchical analysis method is used to determine the weight of each evaluation index by constructing a hierarchical analysis model with a three-layer structure. According to the weight of each evaluation index, a weighted summation method is used to multiply the business income score, the user number score and the idle resource score of each IP node by the corresponding weight and sum them to obtain the business importance score of each IP node. According to the business importance score of each IP node, a data sorting method is used to sort the business importance scores in descending order to obtain the IP node sorting results from high to low business importance. According to the business importance scores and IP node sorting results, a classification method is used to set the threshold of the business importance score to determine the business importance level of each IP node, and obtain the classification results of key IP nodes, important IP nodes and general IP nodes.

[0093] Get the 95th value of each IP in the shared bandwidth package as the basic data for calculating the cost sharing weight. According to the business importance score of each IP provided by the business system, combined with the 95th value data, use the weighted average algorithm to calculate the cost sharing weight of each IP. Multiply the billing amount of the shared bandwidth package by the cost sharing weight of each IP to obtain the cost amount to be shared by each IP. Record the cost amount shared by each IP in the billing system, bind it to the IP, and determine the bandwidth usage fee that each IP needs to pay in the end.

[0094] As an application example of the first embodiment, Figure 3As shown in the figure, multiple services share a shared bandwidth package. To achieve refined management and accounting, it is mainly divided into two aspects: daily IP library management and maintenance and monthly refined cost accounting. In terms of daily IP library management and maintenance, IP resources in a multi-cloud environment can be integrated into CMDB for unified management by establishing an IP library, which can facilitate operation and maintenance and ensure information security. Daily operation and maintenance personnel need to manage servers, load balancing, NAT and other resources due to operation and maintenance needs. Therefore, in order to avoid duplication of work by operation and maintenance personnel, IP automatically binds the corresponding "business path" by obtaining resource binding relationships from the cloud. For IPs that have not found the "business path", daily alarms are used to remind operation and maintenance personnel to judge the operation status of the IP and handle it in a timely manner, such as: timely deletion of missed resources to reduce cost waste; timely supplementary binding of resources of the missed "business path" to ensure that all IPs have their own "business path". In terms of monthly refined cost sharing, the bandwidth value of each IP every 5 minutes is obtained through the cloud interface and saved in the table; the bandwidth value of each IP every 5 minutes is sorted from large to small, and the number ranked 5% is taken to calculate the 95th value of all IPs; the monthly billing 95th value and amount of the shared bandwidth package are obtained through the cloud interface, and the amount is allocated according to the proportion of the 95th value of each IP to obtain the billing amount for each IP; after each IP obtains the share amount, the "business path" can be obtained from the IP library, and finally the refined sharing of multi-business shared bandwidth packages is completed.

[0095] Reference Figure 4 An embodiment of the present invention provides a refined management system 4 for shared bandwidth packages, and the system 4 specifically includes:

[0096] The first management module 401 is used to collect IP node usage data, bandwidth flow data and business information from multiple cloud platforms, align multi-source data according to timestamps, and form a multi-dimensional resource usage data set;

[0097] The second management module 402 is used to construct an IP node and service association map based on the multi-dimensional resource usage data set, analyze the IP node usage based on the IP node and service association map using a graph traversal algorithm, identify IP nodes that have not been used for a long time or are not bound to a service, and issue an alarm to the operation and maintenance personnel;

[0098] The third management module 403 is used to obtain the bandwidth usage of each IP in each preset time period according to the bandwidth flow data, determine the 95th value of each IP by sorting the bandwidth usage of each IP in each preset time period from large to small, determine the billing amount of each IP according to the ratio of the 95th value of each IP and the billing amount of the shared bandwidth package, and determine the billing amount of each service according to the billing amount of each IP and the IP node and service association map;

[0099] The fourth management module 404 is used to obtain bandwidth usage history data based on the bandwidth flow data, perform cluster analysis on the bandwidth usage history data through a clustering algorithm, determine the bandwidth usage pattern, and use an autoregressive moving average model to establish a bandwidth usage baseline model, and implement abnormal bandwidth fluctuation alarm based on the bandwidth usage baseline model.

[0100] It is understandable that if Figure 1 The contents of the embodiment of the refined management method of the shared bandwidth package shown in the figure are applicable to the embodiment of the refined management system of the shared bandwidth package. The functions specifically implemented by the embodiment of the refined management system of the shared bandwidth package are similar to those of the embodiment of the shared bandwidth package management system shown in the figure. Figure 1 The embodiment of the method for fine-grained management of shared bandwidth packages shown in FIG. 1 is the same as that of FIG. 1 , and the beneficial effects achieved are the same as those of FIG. Figure 1 The beneficial effects achieved by the illustrated embodiment of the method for refined management of shared bandwidth packages are also the same.

[0101] It should be noted that the information interaction, execution process and other contents between the above-mentioned systems are based on the same concept as the embodiment of the method of the present invention. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0102] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0103] Reference Figure 5 The embodiment of the present invention further provides a computer device 5, comprising: a memory 502 and a processor 501 and a computer program 503 stored in the memory 502. When the computer program 503 is executed on the processor 501, a refined management method for a shared bandwidth package as described in any one of the above methods is implemented.

[0104] The computer device 5 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The computer device 5 may include, but is not limited to, a processor 501 and a memory 502. Those skilled in the art will appreciate that Figure 5 It is only an example of the computer device 5 and does not constitute a limitation on the computer device 5. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include input and output devices, network access devices, etc.

[0105] The processor 501 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0106] In some embodiments, the memory 502 may be an internal storage unit of the computer device 5, such as a hard disk or memory of the computer device 5. In other embodiments, the memory 502 may also be an external storage device of the computer device 5, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device 5. Further, the memory 502 may also include both an internal storage unit of the computer device 5 and an external storage device. The memory 502 is used to store an operating system, an application program, a boot loader (Boot Loader), data, and other programs, such as the program code of the computer program, etc. The memory 502 may also be used to temporarily store data that has been output or is to be output.

[0107] 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, the method for fine-grained management of shared bandwidth packages as described in any one of the above methods is implemented.

[0108] In this embodiment, if the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the camera / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0109] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0110] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0111] In the embodiments disclosed in the present application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0112] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

Claims

1. A refined management method for a shared bandwidth package, characterized in that: The method specifically comprises: Collect IP node usage data, bandwidth flow data, and business information from multiple cloud platforms, align multi-source data based on timestamps, and form a multi-dimensional resource usage data set; Construct an IP node and service association graph based on the multi-dimensional resource usage data set, analyze IP node usage using a graph traversal algorithm based on the IP node and service association graph, identify IP nodes that have not been used for a long time or are not bound to a service, and issue an alarm to operation and maintenance personnel; Obtain the bandwidth usage of each IP in each preset time period according to the bandwidth flow data, determine the 95th value of each IP by sorting the bandwidth usage of each IP in each preset time period from large to small, determine the billing amount of each IP according to the ratio of the 95th value of each IP and the billing amount of the shared bandwidth package, and determine the billing amount of each service according to the billing amount of each IP and the IP node and service association map; Acquire bandwidth usage history data according to the bandwidth flow data, perform cluster analysis on the bandwidth usage history data by using a clustering algorithm to determine the bandwidth usage pattern, and establish a bandwidth usage baseline model by using an autoregressive moving average model, and implement abnormal bandwidth fluctuation alarm according to the bandwidth usage baseline model; The method of clustering the bandwidth usage history data using a clustering algorithm to determine the bandwidth usage pattern, and using an autoregressive moving average model to establish a bandwidth usage baseline model, and implementing an abnormal bandwidth fluctuation alarm based on the bandwidth usage baseline model, specifically includes: Performing cluster analysis on the bandwidth usage history data according to a K-means clustering algorithm, calculating the Euclidean distance from each data point to the cluster center, assigning the data point to the cluster with the closest distance, and updating the cluster center, iterating continuously until the cluster center no longer changes, and obtaining a bandwidth usage pattern clustering result, wherein the bandwidth usage pattern clustering result includes high bandwidth usage pattern data, medium bandwidth usage pattern data, and low bandwidth usage pattern data; Inputting the high bandwidth usage pattern data, the medium bandwidth usage pattern data and the low bandwidth usage pattern data into an autoregressive moving average model for modeling training respectively, to obtain a first bandwidth usage baseline model, a second bandwidth usage baseline model and a third bandwidth usage baseline model; According to the bandwidth demand values ​​corresponding to different levels of services, the first bandwidth usage baseline model, the second bandwidth usage baseline model or the third bandwidth usage baseline model is used to determine whether the current bandwidth usage data exceeds the normal bandwidth fluctuation range. If it exceeds, it is determined to be a bandwidth abnormal state and an alarm is sent to the operation and maintenance personnel.

2. The method according to claim 1, characterized in that The step of constructing an IP node and service association graph based on the multi-dimensional resource usage data set specifically includes: Map each IP node to a two-dimensional vector space, and calculate the similarity between every two IP nodes using cosine similarity based on the vector of each mapped IP node to obtain a first similarity matrix; Obtaining IP node clustering results by performing spectral clustering on the first similarity matrix; Based on the multi-dimensional resource usage data set and according to the IP node clustering result, the association rules between IP nodes and services are mined through an association analysis algorithm to form a bipartite graph of IP nodes and services; A low-dimensional vector representation of the bipartite graph is obtained by analyzing the heterogeneous graph representation learning algorithm, and the semantic similarity between each IP node and each service is calculated according to the low-dimensional vector representation to form a second similarity matrix; The second similarity matrix is ​​used to perform bipartite graph clustering to obtain the association clustering results between IP nodes and services, and an IP node and service association graph is constructed based on the association clustering results.

3. The method according to claim 1, characterized in that Based on the IP node and service association graph, the graph traversal algorithm is used to analyze the usage of the IP nodes, identify IP nodes that have not been used for a long time or are not bound to services, and issue an alarm to the operation and maintenance personnel, specifically including: Based on the IP node and business association graph, a graph traversal algorithm is used to traverse each IP node in the graph to obtain the last use timestamp of each IP node and a list of business nodes associated with it; According to the last use timestamp of the current IP node, determine whether the current IP node is a long-term unused resource. If the last use timestamp is more than the preset time threshold from the current time, mark the current IP node as a long-term unused resource. According to the service node list associated with the current IP node, determine whether the current IP node is an unbound service resource. If the service node list associated with the current IP node is empty, mark the current IP node as an unbound service resource. Acquire the attribute information of the IP nodes marked as long-term unused resources or unbound service resources, generate alarm information according to the attribute information, and send the alarm information to the operation and maintenance personnel.

4. The method according to claim 1, characterized in that: The method determines the 95th value of each IP by sorting the bandwidth usage of each IP in each preset time period from large to small, and determines the billing amount of each IP according to the ratio of the 95th value of each IP and the billing amount of the shared bandwidth package, specifically including: A quick sorting algorithm is used to sort the bandwidth usage data of each IP in each preset time period from large to small to obtain sorted bandwidth usage data; According to the sorted bandwidth usage data, the 95th percentile value of each IP is determined by a quantile algorithm to obtain the 95th value bandwidth usage of each IP; The billing amount for each IP is calculated using a weighted average algorithm based on the billing amount of the shared bandwidth package and the ratio of each IP's 95% bandwidth usage to the total 95% bandwidth usage of all IPs. If the billing amount of any IP is less than 0, its billing amount will be set to 0, and the difference in its billing amount will be proportionally allocated to the billing amounts of other IPs, and the billing amounts of other IPs will be recalculated until the billing amounts of all IPs are greater than or equal to 0.

5. The method according to claim 1, characterized in that The method further comprises: Obtain historical bandwidth usage data, and based on the historical bandwidth usage data, use a weighted moving average method to perform statistical analysis on the bandwidth usage of each IP node to obtain an average bandwidth usage rate of each IP node; According to the historical bandwidth usage data, the bandwidth usage trend of each IP node is fitted using the least square method to obtain a bandwidth usage trend coefficient of each IP node; Determining a resource utilization score of each IP node according to the average bandwidth utilization and the bandwidth utilization trend coefficient; According to the resource utilization scores of all IP nodes, a Gaussian mixture model is used to perform cluster analysis on the resource utilization scores of all IP nodes to obtain the distribution of resource utilization scores of different categories of IP nodes; The score value corresponding to the first trough of the resource utilization score distribution is determined as a potential idle resource threshold, and the IP nodes whose resource utilization scores are lower than the potential idle resource threshold are marked as potential idle resources.

6. The method according to claim 5, characterized in that The method further comprises: According to the business income and number of users of each IP node and whether each IP node is a potential idle resource, a business importance evaluation matrix is ​​established by using the hierarchical analysis method, and the business importance score of each IP node is determined by the business importance evaluation matrix; The cost sharing weight is determined according to the ratio of the 95 value of each IP and the business importance score, and the billing amount of each IP is determined by the cost sharing weight and the billing amount of the shared bandwidth package.

7. A refined management system for shared bandwidth packages, characterized in that: The system specifically comprises: The first management module is used to collect IP node usage data, bandwidth flow data and business information from multiple cloud platforms, align multi-source data according to timestamps, and form a multi-dimensional resource usage data set; A second management module is used to construct an IP node and business association map based on the multi-dimensional resource usage data set, and based on the IP node and business association map, use a graph traversal algorithm to analyze the IP node usage, identify IP nodes that have not been used for a long time or are not bound to a business, and issue an alarm to the operation and maintenance personnel; A third management module is used to obtain the bandwidth usage of each IP in each preset time period according to the bandwidth flow data, determine the 95th value of each IP by sorting the bandwidth usage of each IP in each preset time period from large to small, determine the billing amount of each IP according to the ratio of the 95th value of each IP and the billing amount of the shared bandwidth package, and determine the billing amount of each service according to the billing amount of each IP and the IP node and service association map; a fourth management module, configured to obtain bandwidth usage history data according to the bandwidth flow data, perform cluster analysis on the bandwidth usage history data by using a clustering algorithm, determine a bandwidth usage pattern, establish a bandwidth usage baseline model by using an autoregressive moving average model, and implement an abnormal bandwidth fluctuation alarm according to the bandwidth usage baseline model; The method of clustering the bandwidth usage history data using a clustering algorithm to determine the bandwidth usage pattern, and using an autoregressive moving average model to establish a bandwidth usage baseline model, and implementing an abnormal bandwidth fluctuation alarm based on the bandwidth usage baseline model, specifically includes: Performing cluster analysis on the bandwidth usage history data according to a K-means clustering algorithm, calculating the Euclidean distance from each data point to the cluster center, assigning the data point to the cluster with the closest distance, and updating the cluster center, iterating continuously until the cluster center no longer changes, and obtaining a bandwidth usage pattern clustering result, wherein the bandwidth usage pattern clustering result includes high bandwidth usage pattern data, medium bandwidth usage pattern data, and low bandwidth usage pattern data; Inputting the high bandwidth usage pattern data, the medium bandwidth usage pattern data and the low bandwidth usage pattern data into an autoregressive moving average model for modeling training respectively, to obtain a first bandwidth usage baseline model, a second bandwidth usage baseline model and a third bandwidth usage baseline model; According to the bandwidth demand values ​​corresponding to different levels of services, the first bandwidth usage baseline model, the second bandwidth usage baseline model or the third bandwidth usage baseline model is used to determine whether the current bandwidth usage data exceeds the normal bandwidth fluctuation range. If it exceeds, it is determined to be a bandwidth abnormal state and an alarm is sent to the operation and maintenance personnel.

8. A computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory, when the computer program is executed on the processor, implements the refined management method of the shared bandwidth package according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the refined management method of the shared bandwidth package according to any one of claims 1 to 6 is implemented.

Images