CDN resource intelligent scheduling cost optimization method and device based on AI

Abstract

The invention relates to an AI-based CDN resource intelligent scheduling cost optimization method and device. Obtaining the current resource use condition of each resource node in the distribution network and the cost strategy of each resource affiliation party to which each resource node belongs; adopting an AI cost strategy model to analyze an estimated resource demand required by a demand side in a prediction time period, a current resource use condition of each resource node and a cost strategy of each resource affiliation side to which each resource node belongs, and obtaining a target node selected by the AI cost model from each resource node; and providing corresponding resources for the demand side in the prediction time period by adopting the target node.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to an AI-based CDN resource intelligent scheduling cost optimization method and device. Background Art

[0002] With the rapid growth of Internet content, content distribution network (CDN) has become a key technology to improve content transmission efficiency and improve user experience. CDN shortens the distance of content transmission by copying content to multiple geographically distributed servers, thereby reducing transmission delays and improving access speeds. However, different CDN service providers have different pricing models. For example, some charge by bandwidth peak, some charge by traffic usage, and some charge by service duration. This makes CDN fusion vendors face complex cost management issues when selecting and managing CDN resources. CDN resource scheduling in existing technologies mostly focuses on performance and service quality, but often ignores cost optimization. Therefore, CDN fusion vendors are in urgent need of a method that can optimize the cost of CDN resource usage while ensuring service quality. Summary of the invention

[0003] Based on this, it is necessary to provide an AI-based CDN resource intelligent scheduling cost optimization method and device that can combine resource scheduling costs for scheduling optimization in response to the above technical problems.

[0004] In a first aspect, the present application provides an AI-based CDN resource intelligent scheduling cost optimization method, the method comprising:

[0005] Obtain the current resource usage of each resource node in the distribution network, as well as the cost strategy of each resource owner to which each resource node belongs;

[0006] The AI ​​cost strategy model is used to analyze the estimated resource demand required by the demand side during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtain the target node selected from each resource node by the AI ​​cost model;

[0007] Use target nodes to provide corresponding resources to demanders within the forecast period.

[0008] In one embodiment, obtaining the cost strategy of any resource owner includes:

[0009] Obtain the resource types and resource purchasing modes that the resource owner can provide;

[0010] Determine the cost strategy of the resource owner based on the cost rules corresponding to each resource type, the charging model corresponding to each resource purchase model, and the regional cost rules of the resource owner;

[0011] Among them, resource types include traffic resources and bandwidth resources; resource purchase modes include real-time purchase and prepaid purchase.

[0012] In one embodiment, obtaining the current resource usage of any resource node includes:

[0013] Obtain the traffic consumption records, request records, and bandwidth usage records of the resource nodes to determine the current resource usage of the resource nodes.

[0014] In one embodiment, the estimated resource requirements include various service types and service resource requirements corresponding to the various service types.

[0015] In one embodiment, an AI cost strategy model is used to analyze the estimated resource demand required by the demander during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtain the target node selected from each resource node by the AI ​​cost model, including:

[0016] For each service type, determine the candidate nodes corresponding to the service type, and select the service node corresponding to the service type from the candidate nodes according to the service resource requirements of the service type, the current resource usage of the candidate nodes corresponding to the service type, and the cost strategy of each resource owner to which each candidate node belongs;

[0017] According to the service priority of each service type, a target node is selected from each service node.

[0018] In one embodiment, selecting a service node corresponding to the service type from each candidate node according to the service resource demand of the service type, the current resource usage of the candidate node corresponding to the service type, and the cost strategy of each resource owner to which each candidate node belongs includes:

[0019] Compare the current resource usage of each candidate node with the resource requirements of the business type to obtain the candidate node that meets the resource requirements;

[0020] For each candidate node that meets the resource requirements, the total cost required to deploy the service on the candidate node that meets the resource requirements is calculated according to the cost strategy; the total cost includes at least the resource usage cost and the node startup cost;

[0021] The candidate node that has the smallest total cost and meets the resource requirements is selected as the business node corresponding to the business type.

[0022] In one embodiment, determining a candidate node corresponding to a service type includes:

[0023] According to the association between the service type and each resource node, a candidate node corresponding to the service type is determined from each resource node.

[0024] In one embodiment, according to the association relationship between the service type and each resource node, determining the candidate node corresponding to the service type from each resource node includes:

[0025] According to the relationship between the business type and each resource node, the ability score of each resource node in processing the business type is determined;

[0026] The resource nodes whose capability scores meet the score threshold are selected as candidate nodes corresponding to the business type;

[0027] In one embodiment, obtaining the association relationship between the service type and each resource node includes:

[0028] Obtain the core processing function corresponding to the business type and verify whether each resource node has the core processing function;

[0029] If so, it is determined that the association relationship between the service type and each resource node is an association;

[0030] If not, it is determined that the association relationship between the service type and each resource node is non-existent.

[0031] In one embodiment, the current resource usage includes remaining resource amount, resource usage rate, performance status and maintenance record;

[0032] According to the business resource requirements of the business type, the current resource usage of the candidate nodes corresponding to the business type, and the cost strategy of each resource owner to which each candidate node belongs, a business node corresponding to the business type is selected from each candidate node, including:

[0033] Determine the revenue value of each candidate node based on the business resource requirements of the business type, the current resource usage of the candidate node corresponding to the business type, and the cost strategy of each resource owner to which each candidate node belongs;

[0034] According to the revenue value of each candidate node, a business node corresponding to the business type is selected from each candidate node.

[0035] In a second aspect, the present application also provides a resource scheduling device, including:

[0036] An acquisition module is used to obtain the current resource usage of each resource node in the distribution network and the cost strategy of each resource owner to which each resource node belongs;

[0037] A node selection module is used to use an AI cost strategy model to analyze the estimated resource demand required by the demand side during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtain a target node selected from each resource node by the AI ​​cost model;

[0038] The resource provision module is used to adopt the target node and provide corresponding resources to the demand side within the forecast period.

[0039] The above-mentioned AI-based CDN resource intelligent scheduling cost optimization method and device, by obtaining the current resource usage of each resource node in the distribution network, the present application can understand the resource status of each node in real time, including the remaining resource amount, resource utilization rate, etc. This helps to avoid idle or overused resources, thereby achieving optimal allocation of resources. Combined with the cost strategy of the resource owner to which each resource node belongs, cost factors can be comprehensively considered when selecting the target node. By selecting nodes with a high cost-effectiveness ratio, the total cost of resource use can be reduced and the overall economic benefit can be improved. Based on the estimated resource requirements required by the demand side during the forecast period, it can be ensured that the selected target node can meet the actual needs of the demand side. This helps to avoid service interruptions or performance degradation caused by insufficient resources, thereby improving user satisfaction. By dynamically selecting the target node, resource allocation can be flexibly adjusted according to changes in the resource requirements of the demand side. This helps to improve resource utilization and flexibility, enabling the distribution network to better adapt to different business scenarios and demand changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions 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 related drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 A flowchart of an AI-based CDN resource intelligent scheduling cost optimization method in one embodiment;

[0042] Figure 2 A flowchart of steps for obtaining a cost strategy of any resource owner in an embodiment;

[0043] Figure 3 A flowchart of the step of obtaining a target node selected from various resource nodes by the AI ​​cost model in an embodiment;

[0044] Figure 4A schematic diagram of a flow chart of the steps of obtaining the association relationship between the service type and each resource node in an embodiment;

[0045] Figure 5 A schematic diagram of a flow chart of a step of selecting a service node corresponding to a service type from among the nodes to be selected in one embodiment;

[0046] Figure 6 FIG. 4 is a structural block diagram of a resource scheduling device in an embodiment. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0048] In an exemplary embodiment, Figure 1 As shown, a method for intelligent scheduling cost optimization of CDN resources based on AI is provided, and the method includes:

[0049] S101, obtaining the current resource usage of each resource node in the distribution network and the cost strategy of each resource owner to which each resource node belongs.

[0050] The process of obtaining the cost strategy of each resource owner is as follows: Collect the pricing strategies of each CDN service provider, including factors such as traffic-based billing, bandwidth-based billing, prepaid model, and regional differences. It is also necessary to collect discount information from each service provider to evaluate the actual cost of each service provider under specific conditions. Based on the collected pricing data, the customer SLA is taken into consideration to build a comprehensive AI cost model. This model takes into account different billing methods, regional differences, and discount strategies, and can dynamically calculate the costs of different CDN service providers under various conditions.

[0051] Furthermore, when obtaining the current resource usage of each resource node in the distribution network, it includes: real-time monitoring of the current resource usage of each resource node, including traffic consumption, number of requests, bandwidth occupancy, etc. This information will be transmitted to the cost monitoring system in real time to calculate the resource usage fee.

[0052] S102, using an AI cost strategy model, analyzes the estimated resource demand required by the demander during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtains the target node selected by the AI ​​cost model from each resource node.

[0053] First, obtain the estimated resource requirements required by the demand side during the forecast period.

[0054] Then, communicate with the demand side to clarify the resource type, quantity, quality requirements, etc. required during the forecast period. Based on historical data and trend analysis, predict the changes in future resource demand. Organize the estimated resource demand into a document, including key information such as the demand list, demand quantity, and demand time.

[0055] Resource usage assessment: Check the current resource usage data of each resource node collected in the S101 stage. Analyze the remaining resources, utilization rate, performance status, etc. of each node to determine whether it can meet the estimated resource demand. Based on resource usage and business needs, determine which resource nodes are available during the forecast period and which may be busy or under maintenance.

[0056] The cost strategy acquisition process of each resource owner includes:

[0057] Collect pricing strategies: Billing by traffic: Understand the charging standards of CDN service providers for different traffic levels, including tiered pricing or monthly pricing. Bandwidth billing: Collect the billing rules for bandwidth usage, such as peak bandwidth, average bandwidth, or guaranteed bandwidth. Prepaid model: Understand the content of the prepaid package, including the prepaid amount, the corresponding service period, and the resource quota. Regional differences: Consider the pricing differences in different geographic regions (such as domestic and international), as well as possible special regional discounts.

[0058] Collect discount information: Long-term cooperation discount: Ask the CDN service provider about its discount policy for long-term cooperation customers. Bulk purchase discount: Find out whether you can get discounts when purchasing resources in large quantities. Promotional activities: Pay attention to the CDN service provider's promotional activities, such as limited-time discounts, new customer discounts, etc.

[0059] Build an AI cost model: Incorporate customer SLA: Consider the customer's service level agreement (SLA) in the AI ​​cost model, such as response time, availability, and fault recovery time, which may affect the service cost. Dynamically calculate costs: Use the collected pricing data and discount information to build an AI cost model that can dynamically calculate the costs of different CDN service providers under various conditions (such as different traffic, different regions, and different service levels).

[0060] Specifically, the current resource usage of each resource node in the distribution network is obtained, including:

[0061] Traffic consumption: Real-time monitoring of the traffic consumption of each resource node, including uplink and downlink traffic.

[0062] Number of requests: Record and monitor the number of requests for each node, including HTTP requests, HTTPS requests, etc.

[0063] Bandwidth usage: Real-time monitoring of node bandwidth usage, including instantaneous bandwidth and average bandwidth.

[0064] Data transmission to the cost monitoring system: Real-time transmission: Ensure that the monitored resource usage data can be transmitted to the cost monitoring system in real time.

[0065] Data connection: The cost monitoring system needs to connect with the resource monitoring system to accurately calculate resource usage costs.

[0066] Using the previously constructed AI cost model, combined with the real-time monitored resource usage data, dynamically calculate the usage fee of each resource node. Generate a cost report to show the cost situation of different time periods and different nodes for cost analysis and optimization. Consult the cost strategy information of each resource owner collected in the S101 stage. Analyze the cost structure of different resource nodes, including fixed costs, variable costs, preferential policies, etc. Combine the estimated resource demand and the cost strategy of the resource node to evaluate the cost-effectiveness ratio of different nodes. Consider possible cost-saving measures such as long-term cooperation and bulk procurement.

[0067] Establish a comprehensive evaluation system based on estimated resource demand, resource usage, and cost strategy. The evaluation system should include multiple dimensions, such as resource matching, cost-effectiveness, service quality, risk level, etc. Score each resource node and sort them according to the results of the comprehensive evaluation system. Select the node with the highest score or the top ranking as the target node. Make a decision on the target node based on the evaluation results.

[0068] S103, using the target node to provide corresponding resources to the demand side within the forecast period.

[0069] The above-mentioned AI-based CDN resource intelligent scheduling cost optimization method, this application can understand the resource status of each node in real time, including the remaining resource amount, resource utilization rate, etc., by obtaining the current resource usage of each resource node in the distribution network. This helps to avoid idle or overused resources, thereby achieving optimal allocation of resources. Combined with the cost strategy of the resource owner to which each resource node belongs, cost factors can be comprehensively considered when selecting the target node. By selecting nodes with a high cost-effectiveness ratio, the total cost of resource use can be reduced and the overall economic benefit can be improved. Based on the estimated resource requirements required by the demand side during the forecast period, it can be ensured that the selected target node can meet the actual needs of the demand side. This helps to avoid service interruptions or performance degradation caused by insufficient resources, thereby improving user satisfaction. By dynamically selecting the target node, resource allocation can be flexibly adjusted according to changes in the resource requirements of the demand side. This helps to improve resource utilization and flexibility, enabling the distribution network to better adapt to different business scenarios and demand changes.

[0070] In an exemplary embodiment, Figure 2 As shown, obtain the cost strategy of any resource owner, including:

[0071] S201, obtaining various resource types and resource purchasing modes that the resource owner can provide.

[0072] Among them, resource types include traffic resources and bandwidth resources. Traffic resources usually refer to the amount of data transmission, which can be measured in units such as bytes, megabytes or gigabytes. Bandwidth resources refer to the capacity or speed of network connections, usually expressed in bit rates (such as Mbps, Gbps).

[0073] Among them, resource purchase modes include real-time purchase and prepaid purchase. Real-time purchase means that resources can be purchased immediately as needed, usually charged according to actual usage. Prepaid purchase requires a certain fee to be paid in advance to obtain the right to use resources within a specific period of time.

[0074] S202, determining the cost strategy of the resource owner according to the cost rules corresponding to each resource type, the charging mode corresponding to each resource purchase mode, and the regional cost rules of the resource owner.

[0075] Understandably, for each resource type, you need to understand the corresponding cost rules. For example, traffic resources may be charged per GB, while bandwidth resources may be charged per guaranteed bandwidth per month; you also need to understand the charging model corresponding to each resource purchase model. Real-time purchases may involve dynamic pricing based on usage, while prepaid purchases may provide fixed prices and service terms. Considering the cost differences in different regions, you also need to understand the cost rules of the region where the resource owner is located. This includes regional pricing differences, taxes, policy subsidies and other factors.

[0076] Furthermore, the current resource usage of any resource node is obtained, including: obtaining the flow consumption record, request record and bandwidth occupancy record of the resource node, and determining the current resource usage of the resource node.

[0077] For example, through monitoring tools or API interfaces, obtain the traffic consumption records of resource nodes in a specific time period. This helps to understand the data transmission volume and usage of the node. Record and analyze the number of requests processed by the resource node, including HTTP requests, HTTPS requests, etc. This helps to evaluate the load and performance of the node. Monitor the bandwidth usage of the resource node, including instantaneous bandwidth and average bandwidth. This helps to understand the network capacity and utilization of the node.

[0078] Furthermore, the estimated resource requirements include various business types and business resource requirements corresponding to each business type.

[0079] Among them, service types, such as video streaming, file downloading, web browsing, etc. Different service types may have different requirements for resources. For each service type, estimate the amount of resources it will need during the forecast period. This includes key resources such as traffic, bandwidth, and storage. Combine the resource requirements of each service type to derive the total estimated resource requirements. This helps make reasonable arrangements in resource procurement and scheduling.

[0080] In an exemplary embodiment, Figure 3 As shown, the AI ​​cost strategy model is used to analyze the estimated resource demand required by the demand side during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtain the target node selected from each resource node by the AI ​​cost model, including:

[0081] S301, for each business type, determine the candidate nodes corresponding to the business type, and select the business node corresponding to the business type from the candidate nodes based on the business resource requirements of the business type, the current resource usage of the candidate nodes corresponding to the business type, and the cost strategy of each resource owner to which each candidate node belongs.

[0082] First, for each service type, it is necessary to determine the candidate nodes associated with the service type, which is usually based on a predefined association relationship between the service type and the resource node. For example, some service types may be more suitable for using a specific type of resource node.

[0083] For each business type, the business node that best suits the business type is selected from the candidate nodes based on its business resource requirements (such as traffic, bandwidth, etc.), the current resource usage of the candidate nodes (such as remaining traffic, bandwidth occupancy, etc.), and the cost strategy of the resource owner of each candidate node (such as pricing rules, discount policy, etc.).

[0084] Specifically, the method selects a business node corresponding to the business type from each candidate node based on the business resource demand of the business type, the current resource usage of the candidate nodes corresponding to the business type, and the cost strategy of each resource owner to which each candidate node belongs, including: comparing the current resource usage of each candidate node with the resource demand of the business type to obtain the candidate nodes that meet the resource demand; for each candidate node that meets the resource demand, calculating the total cost required to deploy the business on the candidate node that meets the resource demand according to the cost strategy; the total cost includes at least the resource usage cost and the node startup cost; and selecting the candidate node that meets the resource demand with the smallest total cost as the business node corresponding to the business type.

[0085] Exemplarily, selecting the service node that best suits the service type from the candidate nodes based on their service resource requirements (such as traffic, bandwidth, etc.), the current resource usage of the candidate nodes (such as remaining traffic, bandwidth occupancy, etc.), and the cost strategy (such as pricing rules, discount policies, etc.) of the resource owner of each candidate node may include:

[0086] 1. Define input parameters

[0087] Business type set businessTypes: contains all business types that need to be assigned.

[0088] Node collection nodes: contains all nodes to be selected.

[0089] Business resource requirements resourceRequirements: The amount of resources required for each business type, which can be a vector or dictionary, indicating the requirements for different resources.

[0090] Current resource usage currentResourceUsage: The current resource usage of each node, which can also be a vector or dictionary.

[0091] CostStrategies: The cost strategy of each resource owner, which can be a function or a lookup table, used to calculate the cost under different resource usage.

[0092] 2. Initialize data structure

[0093] Candidate node set candidateNodes: For each business type, initialize an empty candidate node set.

[0094] Optimal node selection optimalNodes: used to store the nodes finally selected for each business type.

[0095] 3. Algorithm steps

[0096] Determine the nodes to be selected:

[0097] For each business type bt in businessTypes:

[0098] According to the characteristics and strategies of the business type, select nodes that meet the preliminary requirements from the nodes and add them to candidateNodes[bt].

[0099] Evaluate node fitness:

[0100] Define a fitness function fitness(node,bt,resourceRequirements,currentResourceUsage,costStrategies) to evaluate the fitness of a node for a certain business type. The function can consider the following factors:

[0101] Are resources sufficient to meet demand?

[0102] The amount of remaining resources (the more the better).

[0103] Cost (lower is better), calculated according to costStrategies.

[0104] Select the best node:

[0105] For each business type bt:

[0106] Initialize the best node bestNode to None and the best fitness bestFitness to negative infinity (or a maximum value if it is a minimization problem).

[0107] Traverse each node node in candidateNodes[bt]:

[0108] Calculate fitnessValue = fitness(node,bt,resourceRequirements[bt],currentResourceUsage[node],costStrategies).

[0109] If fitnessValue is better than bestFitness, update bestNode to the current node and update bestFitness to fitnessValue.

[0110] Add bestNode to optimalNodes[bt].

[0111] Output:

[0112] Returns optimalNodes, which is the optimal node corresponding to each business type.

[0113] S302: Select a target node from each service node according to the service priority of each service type.

[0114] Specifically, determining the candidate node corresponding to the service type includes: determining the candidate node corresponding to the service type from each resource node according to the association relationship between the service type and each resource node.

[0115] Secondly, after selecting the business nodes corresponding to each business type, it is necessary to further screen the target nodes according to the business priority of each business type. Business priority may be based on a variety of factors, such as the importance, urgency, expected benefits, etc. Considering the resource satisfaction, cost-effectiveness and business priority of each business node, the final target node is selected from each business node. These target nodes will provide corresponding resources to the demand side within the forecast period.

[0116] According to the association between the business type and each resource node, determining the candidate node corresponding to the business type from each resource node includes: determining the capability score of each resource node when processing the business type according to the association between the business type and each resource node. The resource node whose capability score meets the scoring threshold is used as the candidate node corresponding to the business type.

[0117] Further, such as Figure 4 As shown, the association relationship between the business type and each resource node is obtained, including:

[0118] S401, obtaining the core processing function corresponding to the service type, and verifying whether each resource node has the core processing function.

[0119] For each business type, it is necessary to clarify its core processing functions. These functions may be specific computing, storage, network transmission and other requirements, or specific processing capabilities closely related to business logic. Describe these core processing functions in a clear and accurate manner to form a functional definition or specification. This will help to verify whether the resource nodes have these functions in the future.

[0120] S402: If yes, determine that the association relationship between the service type and each resource node is an association.

[0121] For each resource node, it is necessary to verify whether it has the core processing function required by the business type. This can be achieved by querying the function list of the resource node, performing functional testing, or viewing relevant documents. Record the verification results for subsequent use. If the resource node has the required core processing function, proceed to the next step; if not, jump to S403.

[0122] S403: If not, determine that the association relationship between the service type and each resource node is non-existent.

[0123] If the resource node has the core processing function required by the service type, it is determined that there is an association relationship between the service type and the resource node. This means that the resource node can be used as a candidate node for processing the service type. If the resource node does not have the core processing function required by the service type, it is determined that there is no association relationship between the service type and the resource node. This means that the resource node cannot be used as a candidate node for processing the service type and needs to be searched in other resource nodes.

[0124] In an exemplary embodiment, the current resource usage includes remaining resource amount, resource usage rate, performance status and maintenance record;

[0125] Correspondingly, such as Figure 5 As shown, according to the business resource requirements of the business type, the current resource usage of the candidate nodes corresponding to the business type, and the cost strategy of each resource owner to which each candidate node belongs, a business node corresponding to the business type is selected from each candidate node, including:

[0126] S501, determining the revenue value of each candidate node according to the service resource demand of the service type, the current resource usage of the candidate node corresponding to the service type, and the cost strategy of each resource owner to which each candidate node belongs.

[0127] First, evaluate whether the current resource usage of each candidate node (including remaining resources, resource utilization, performance status, and maintenance records) can meet the business resource requirements of the business type. Next, consider the cost strategy of the resource owner of each candidate node, including pricing, discounts, promotions and other factors, and calculate the cost-effectiveness of using each candidate node. Combined with the satisfaction of business resource requirements and cost-effectiveness, a benefit value is calculated for each candidate node. This benefit value can be a comprehensive score that reflects the comprehensive performance of the node in meeting business needs and cost control.

[0128] S502: Select a service node corresponding to the service type from among the nodes to be selected according to the revenue value of each node to be selected.

[0129] Specifically, the revenue values ​​of each candidate node are sorted from high to low. According to the sorting result, the candidate node with the highest revenue value is selected as the business node corresponding to the business type. If there are multiple nodes with the same or similar revenue values, other factors (such as the geographical location of the node, network latency, etc.) can be further considered to make a decision.

[0130] Through the above steps, you can reasonably select the business nodes corresponding to the business type based on business resource requirements, current resource usage and cost strategy, thereby optimizing resource allocation and cost control.

[0131] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0132] Based on the same inventive concept, the embodiment of the present application also provides a resource scheduling device for implementing the above-mentioned AI-based CDN resource intelligent scheduling cost optimization method. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above-mentioned method, so the specific limitations in one or more resource scheduling device embodiments provided below can refer to the limitations of the AI-based CDN resource intelligent scheduling cost optimization method above, and will not be repeated here.

[0133] In an exemplary embodiment, Figure 6 As shown, a method and device for optimizing the cost of intelligent scheduling of CDN resources based on AI is provided, comprising: an acquisition module 11, a node selection module 12 and a resource provision module 13, wherein:

[0134] The acquisition module 11 is used to acquire the current resource usage of each resource node in the distribution network and the cost strategy of each resource owner to which each resource node belongs;

[0135] The node selection module 12 is used to use the AI ​​cost strategy model to analyze the estimated resource demand required by the demand side during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtain the target node selected by the AI ​​cost model from each resource node;

[0136] The resource providing module 13 is used to adopt the target node and provide corresponding resources to the demand side within the forecast period.

[0137] Each module in the resource scheduling device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0138] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0139] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0140] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. An AI-based CDN resource intelligent scheduling cost optimization method, characterized in that: The method comprises: Obtain the current resource usage of each resource node in the distribution network, as well as the cost strategy of each resource owner to which each resource node belongs; The AI ​​cost strategy model is used to analyze the estimated resource demand required by the demand side during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtain the target node selected from each resource node by the AI ​​cost model; The target node is used to provide corresponding resources for the demander within the forecast period.

2. The method according to claim 1, characterized in that Strategies for obtaining the cost of any resource ownership include: Obtaining various resource types and resource purchasing modes that the resource owner can provide; Determine the cost strategy of the resource owner according to the cost rules corresponding to each resource type, the charging mode corresponding to each resource purchase mode, and the regional cost rules of the resource owner; Among them, the resource types include traffic resources and bandwidth resources; the resource purchase modes include real-time purchase and prepaid purchase.

3. The method according to claim 2, characterized in that Get the current resource usage of any resource node, including: The flow consumption record, request record and bandwidth occupancy record of the resource node are obtained to determine the current resource usage of the resource node.

4. The method according to claim 1, characterized in that: The AI ​​cost strategy model is used to analyze the estimated resource demand required by the demand side during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtain the target node selected from each resource node by the AI ​​cost model, including: Adopting the AI ​​cost strategy model, for each service type, determine the candidate nodes corresponding to the service type, and select the service node corresponding to the service type from the candidate nodes according to the service resource requirements of the service type, the current resource usage of the candidate nodes corresponding to the service type, and the cost strategy of each resource owner to which each candidate node belongs; According to the service priority of each service type, a target node is selected from each service node.

5. The method according to claim 4, characterized in that The selecting a service node corresponding to the service type from each candidate node according to the service resource demand of the service type, the current resource usage of the candidate node corresponding to the service type, and the cost strategy of each resource owner to which each candidate node belongs includes: Compare the current resource usage of each candidate node with the resource requirements of the business type to obtain the candidate node that meets the resource requirements; For each candidate node that meets the resource requirements, the total cost required to deploy the service on the candidate node that meets the resource requirements is calculated according to the cost strategy; the total cost includes at least the resource usage cost and the node startup cost; The candidate node that meets the resource requirements and has the smallest total cost is used as the service node corresponding to the service type.

6. The method according to claim 4, characterized in that The determining of the candidate node corresponding to the service type includes: According to the association relationship between the service type and each resource node, a candidate node corresponding to the service type is determined from each resource node.

7. The method according to claim 4, characterized in that The determining, from the resource nodes, a candidate node corresponding to the service type according to the association relationship between the service type and the resource nodes, comprises: Determining the capability score of each resource node when processing the service type according to the association relationship between the service type and each resource node; The resource nodes whose capability scores meet the score threshold are used as candidate nodes corresponding to the business type.

8. The method according to claim 6, characterized in that The obtaining of the association relationship between the service type and each resource node includes: Obtain the core processing function corresponding to the service type, and verify whether each resource node has the core processing function; If so, determining that the association relationship between the service type and each resource node is an association; If not, it is determined that the association relationship between the service type and each resource node is non-existent.

9. The method according to claim 5, characterized in that The current resource usage includes remaining resource amount, resource utilization rate, performance status and maintenance record; The selecting a service node corresponding to the service type from each candidate node according to the service resource demand of the service type, the current resource usage of the candidate node corresponding to the service type, and the cost strategy of each resource owner to which each candidate node belongs includes: Determining the revenue value of each candidate node according to the business resource demand of the business type, the current resource usage of the candidate node corresponding to the business type, and the cost strategy of each resource owner to which each candidate node belongs; According to the revenue value of each candidate node, a service node corresponding to the service type is selected from each candidate node.

10. A resource scheduling device, characterized in that: The device comprises: An acquisition module is used to obtain the current resource usage of each resource node in the distribution network and the cost strategy of each resource owner to which each resource node belongs; A node selection module is used to use an AI cost strategy model to analyze the estimated resource demand required by the demand side during the forecast period, the current resource usage of each resource node, and the cost strategy of each resource owner to which each resource node belongs, and obtain a target node selected from each resource node by the AI ​​cost model; A resource providing module is used to adopt the target node to provide corresponding resources to the demander within the forecast period.

Images