Multi-dimensional evaluation method for optimizing content distribution strategy

Abstract

The invention relates to a multi-dimensional evaluation method for optimizing a content distribution strategy, and the method comprises the steps: constructing a comprehensive accelerated distribution effect comprehensive evaluation index system, and analyzing the performance of a whole network within a certain period of time after the distribution strategy takes effect. Quantitative evaluation is carried out from multiple aspects of request hit rate, user time delay, bandwidth consumption, storage occupation, cache updating and the like. The system not only pays attention to the subjective perception of the user in the resource access process, but also quantifies the actual conditions of user experience and system resource utilization through specific performance indexes. Through the multi-dimensional quantitative analysis, the subjective feeling of the user on the service experience can be converted into specific objective data, and a reliable basis is provided for scientific evaluation of the distribution effect.

Description

Technical Field

[0001] This invention belongs to the field of intelligent data chain distribution technology, and in particular relates to a multi-dimensional evaluation method for optimizing content distribution strategies. Background Technology

[0002] In the process of accelerating the distribution of high-traffic services, the diversity of request types, the differences in access frequency, the limitations of network bandwidth, and the uneven performance of nodes all contribute to the instability and complexity of the distribution effect. Faced with massive concurrent requests, the operating status of the distribution system is often affected by many dynamic factors, such as fluctuations in the network environment, limitations of caching mechanisms, and differences in the load of distribution nodes. This makes traditional methods for evaluating the distribution effect difficult to meet actual needs.

[0003] Traditional evaluation methods typically focus on a single metric, such as cache hit rate or bandwidth utilization, neglecting the comprehensive performance of the distribution system across multiple dimensions. This one-dimensional evaluation approach not only fails to accurately capture the adaptability of distribution strategies in different scenarios but may also lead to a one-sided approach to optimization, ultimately impacting user experience and system efficiency.

[0004] Therefore, researching an indicator system that can scientifically and comprehensively evaluate the effect of accelerated distribution has become a key breakthrough in optimizing the distribution strategy for high-traffic services. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-dimensional evaluation method for optimizing content distribution strategies, in order to solve the technical problems of existing evaluation methods that focus on a single indicator and therefore cannot fully reflect the effect of accelerating distribution, and that rely on subjective human evaluation and static data, leading to biased evaluation results.

[0006] To achieve one of the aforementioned objectives, one embodiment of the present invention provides a multi-dimensional evaluation method for optimizing content distribution strategies, including:

[0007] An evaluation index system is constructed based on the observation data of the running status after the distribution strategy takes effect; the evaluation index system includes the request hit rate, average response latency, bandwidth-weighted average peak-to-average ratio, average node storage rate, and update cache access rate.

[0008] The evaluation indicators in the constructed evaluation indicator system are quantified and normalized to eliminate the influence of different dimensions between the evaluation indicators.

[0009] Based on the normalization results of the above evaluation indicators, a comprehensive evaluation function is constructed to form a quantitative assessment of the distribution strategy.

[0010] As a further improvement of one embodiment of the present invention, the expression for the request hit rate is:

[0011] ,

[0012] in, The request hit rate for node i within the scope of the distribution strategy. The number of user requests responded to by node i. The number of user requests generated within the scope of a distribution strategy.

[0013] As a further improvement of one embodiment of the present invention, the expression for the request hit rate is:

[0014] ,

[0015] in, The request hit rate for node i within the scope of the distribution strategy. The number of user requests responded to by node i. The number of user requests generated within the scope of a distribution strategy.

[0016] As a further improvement to one embodiment of the present invention, the expression for the bandwidth-weighted average peak-to-peak ratio is:

[0017] ,

[0018] in, The bandwidth-weighted average peak-to-peak ratio of node i within the scope of the distribution strategy, where K is the number of traffic services from node i to the user during the observation period. Let be the downlink bandwidth of the traffic service k from node i to user k. The bandwidth limit for node i. Let be the peak bandwidth of node i.

[0019] As a further improvement to one embodiment of the present invention, the expression for the average storage rate of the node is:

[0020] ,

[0021] in, The average storage rate of node i within the scope of the distribution strategy. The total storage allocated to node i. Let L be the desired storage utilization ratio, and L be the number of storage samples for node i during the observation period. This is the l-th storage sample for node i.

[0022] As a further improvement to one embodiment of the present invention, the expression for updating the cache access rate is:

[0023] ,

[0024] in, The update cache access rate of node i within the scope of the distribution strategy. This represents the number of times a user accesses the q-th dropdown resource of node i within the observation period. This represents the number of times a user accesses the d-th deleted resource at node i within the observation period. The number of user requests responded to by node i.

[0025] As a further improvement to one embodiment of the present invention, the expression for updating the cache access rate is:

[0026] ,

[0027] in, The update cache access rate of node i within the scope of the distribution strategy. This represents the number of times a user accesses the q-th dropdown resource of node i within the observation period. This represents the number of times a user accesses the d-th deleted resource at node i within the observation period. The number of user requests responded to by node i.

[0028] As a further improvement of one embodiment of the present invention, the comprehensive evaluation function is a weighted sum of the normalized results of the evaluation indicators.

[0029] As a further improvement to one embodiment of the present invention, the method further includes,

[0030] A threshold is set for evaluating the distribution strategy. If the current comprehensive evaluation function value is higher than the threshold, the current distribution strategy is considered reasonable; otherwise, it is considered unreasonable, and the distribution strategy is adjusted based on expert knowledge.

[0031] To achieve one of the above-mentioned objectives, an embodiment of the present invention also provides an electronic device, including a memory and a processor, characterized in that the memory stores a computer program that can run on the processor, and when the program is executed on the processor, it implements the steps in the multi-dimensional evaluation method for optimizing content distribution strategies as described above.

[0032] To achieve one of the above-mentioned objectives, an embodiment of the present invention also provides a storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the steps in the multi-dimensional evaluation method for optimizing content distribution strategies as described above.

[0033] Compared with the prior art, the multi-dimensional evaluation method for optimizing content distribution strategies provided by this invention has the following beneficial effects:

[0034] 1. Multi-dimensional Comprehensive Evaluation: This invention constructs a comprehensive evaluation index system that includes multiple dimensions such as request hit rate, user latency, bandwidth consumption, storage usage, and cache updates. This allows the evaluation to not only focus on single system performance or user experience, but also to comprehensively consider the distribution system's performance under various operating conditions. This method can comprehensively reflect the actual operating status of the system, avoiding the limitations of traditional methods that rely on only a single indicator such as cache hit rate or bandwidth utilization. This makes the evaluation results more objective and comprehensive, thus providing a more precise direction for subsequent optimization.

[0035] 2. Real-time Dynamic Analysis: Traditional distribution effectiveness evaluation often relies on historical data or subjective human judgment, which may lead to evaluation bias due to data lag or human error, and fail to promptly identify system bottlenecks or problems. This invention, based on real-time data analysis, can dynamically and continuously monitor various key performance indicators during the distribution process, ensuring that the evaluation results instantly reflect the current operating status of the distribution system, thereby enabling rapid response and problem repair. This not only enhances the accuracy of the evaluation but also improves the timeliness of system optimization.

[0036] 3. Quantification and Normalization: This invention quantifies and normalizes evaluation data from different dimensions to ensure comparability and consistency among different indicators. Standardizing each evaluation indicator avoids biases caused by differences in data volume or measurement methods, making the evaluation results more scientific and objective. Simultaneously, normalization allows these multi-dimensional indicators to be comprehensively evaluated within the same framework, facilitating unified optimization decisions. This method clearly demonstrates the system's performance in various aspects and provides clear quantitative evidence, offering precise references for subsequent strategy adjustments. Attached Figure Description

[0037] Figure 1 This is a flowchart of the construction process of the evaluation index system of this invention. Detailed Implementation

[0038] The present invention will now be described in detail with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of the present invention.

[0039] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0040] In Embodiment 1 of the present invention, a multi-dimensional evaluation method for optimizing content distribution strategies is provided, such as... Figure 1 As shown, the method includes,

[0041] An evaluation index system is constructed based on the observation data of the running status after the distribution strategy takes effect; the evaluation index system includes the request hit rate, average response latency, bandwidth-weighted average peak-to-average ratio, average node storage rate, and update cache access rate.

[0042] The evaluation indicators in the constructed evaluation indicator system are quantified and normalized to eliminate the influence of different dimensions between the evaluation indicators.

[0043] Based on the normalization results of the above evaluation indicators, a comprehensive evaluation function is constructed to form a quantitative assessment of the distribution strategy.

[0044] This invention aims to construct a comprehensive evaluation index system for accelerated distribution performance. By analyzing the overall network performance within a certain period after the distribution strategy takes effect, it quantitatively evaluates the performance from multiple aspects such as request hit rate, user latency, bandwidth consumption, storage usage, and cache updates.

[0045] This invention, from the two core perspectives of user experience and system performance, clarifies five key evaluation dimensions. At the user experience level, request hit rate and user latency are selected. The former measures whether resources can quickly respond to user needs, while the latter assesses the speed and smoothness of the response process. These two are core indicators that users intuitively perceive regarding service quality. At the system performance level, bandwidth consumption, storage usage, and cache update efficiency are selected, used to measure the efficiency of network resource utilization by the distribution strategy, the rationality of cache resource allocation, and the accuracy during data updates, respectively. These dimensions not only reflect the resource allocation and utilization of the system in different scenarios but also indirectly affect the stability of user experience and service quality. Furthermore, the research further incorporates the typical characteristics of high-traffic services (such as high-concurrency requests and uneven data distribution) and clarifies the definition and calculation methods of each indicator through correlation analysis between distribution strategies and user experience requirements.

[0046] From the perspective of actual business assurance requirements, users' subjective experience with high-traffic services typically focuses on core needs such as fast loading, fast transmission, high stability, and high smoothness. These experience requirements are closely related to the performance of the distribution system. Therefore, when defining the evaluation dimensions for accelerated distribution effectiveness, it is necessary to comprehensively consider two core perspectives: user experience and distribution system performance. From the perspective of user experience, the focus is on the direct perception of users when requesting resources, such as whether resource requests can be responded to quickly (user latency) and whether resources can be obtained from the nearest available location (request hit rate). These indicators directly affect user satisfaction with service quality. From the perspective of distribution system performance, the focus is on the utilization efficiency of network and cache resources, evaluating network bandwidth usage (bandwidth consumption), the rationality of cache resource allocation (storage usage), and the effectiveness and accuracy of the distribution strategy during data update (cache update). By analyzing the correlation between high-traffic service distribution strategies and user business experience requirements, and fully considering the timeliness, accuracy, and convenience of data collection, five key evaluation dimensions—request hit rate, user latency, bandwidth consumption, storage usage, and cache update—were identified, laying a theoretical and practical foundation for subsequent distribution effectiveness evaluation and optimization.

[0047] This invention proposes an operable evaluation index system covering the following core indicators: request hit rate, user latency, bandwidth consumption, storage usage, and cache update efficiency. Each indicator is quantified in conjunction with the actual network environment to ensure its observability and operability. Specific indicator descriptions are as follows:

[0048] 1) Request a hit

[0049] Considering that the response data of each node to user requests is easy to statistically analyze, this invention uses request hit rate to describe the hit rate of user requests under the distribution strategy. Request hit rate is defined as the ratio of user requests responded to by nodes within the observation period. Requests from users can be responded to by edge nodes, regional nodes, and central nodes, and are counted in the hit statistics of their respective nodes. The request hit rate of edge nodes and regional nodes can also be called cache hit rate, while the request hit rate of central nodes can be called origin server hit rate. Generally speaking, hits from edge nodes result in a better user experience.

[0050] 2) User latency

[0051] User latency primarily includes request processing latency, network transmission latency, and distribution processing latency. From a user experience perspective and through statistical analysis of observed data, the average response latency of user requests can be calculated by measuring and analyzing the user request response latency. Average response latency refers to the average latency from initiation to response for all user requests within the effective area of ​​the distribution strategy. Generally, the lower the average latency, the better the user experience. The response latency of a single request can be obtained through proactive measurement or network coordinate prediction.

[0052] 3) Bandwidth consumption

[0053] High-bandwidth traffic consumption refers to the bandwidth usage of resources distributed from upstream nodes to downstream nodes or users. Generally, it's desirable to have controllable peak traffic distribution, to be close to users, and to avoid peak hours. Therefore, a unified bandwidth distribution model is needed to describe system performance. Considering the "top-down" transmission characteristics of high-bandwidth traffic, the downlink bandwidth usage of each node for high-bandwidth traffic can be the primary focus. Given that network bandwidth sampling data is readily available during the observation period, bandwidth consumption can be described by analyzing the relationship between peak bandwidth, average bandwidth, and maximum allocated bandwidth.

[0054] 4) Storage usage

[0055] The storage resources allocated to each node for high-traffic service caching are always limited. The impact of the distribution strategy on storage usage is mainly analyzed by monitoring the storage rate of each node. Node storage rate refers to the ratio of the total allocated storage to the cache resources of each node. An excessively high node storage rate will adversely affect cache updates and resource scheduling; an excessively low node storage rate will result in low storage utilization and a low cache ratio. Therefore, the deviation between the node storage rate and the expected storage rate should be as small as possible.

[0056] 5) Cache update

[0057] The cache updates triggered by the execution of the distribution strategy mainly involve two types of behaviors: first, pulling newly generated data from the origin server to replace existing data or create new cache data; second, deleting caches on nodes that meet certain conditions. By observing the cache updates in the data, the number of pulled and deleted caches can be statistically analyzed. The resource access situation after a cache update is the main basis for judging the effectiveness of the cache update. The ratio of access to updated or deleted cache resources in a node to the total access volume of the node is defined as the update cache access rate. The update cache access rate includes two parts: the pull cache access rate and the delete cache access rate. This metric is used to represent the accuracy of the distribution strategy in updating cache resources. The higher the pull cache access rate and the lower the delete cache access rate, the higher the accuracy of cache release.

[0058] To ensure the feasibility of this indicator system, real-world data from high-traffic business scenarios will be collected. Through data feature extraction and statistical analysis, the calculation methods and weights of each indicator will be clarified. Specifically, actual measurements of network bandwidth, caching strategies, and node performance will be conducted through field tests and simulations to obtain specific indicator values.

[0059] To ensure that the selected evaluation indicators can be effectively integrated into the comprehensive evaluation model, each indicator needs to be quantified and normalized. Since different indicators have different dimensions and value ranges, direct comparison and synthesis are difficult. Therefore, it is necessary to use appropriate methods to make the indicators dimensionless, eliminating their dimensional differences and facilitating subsequent model building and analysis. This study adopts a dimensionless normalization method, mapping the values ​​of each indicator to the [0,1] interval, thereby eliminating the influence of dimensions and facilitating subsequent comprehensive evaluation modeling.

[0060] The specific normalization method depends on the characteristics of the metric: for positive metrics such as request hit rate and cache update efficiency, maximum value normalization is used; for negative metrics such as user latency and bandwidth consumption, minimum value normalization is used; and for storage usage metrics with an optimal value, interval approximation normalization is used. These methods not only achieve unified quantification of each metric but also preserve their differentiated performance in the overall evaluation.

[0061] First, each evaluation indicator is precisely defined mathematically:

[0062] Assuming that after distribution strategy x takes effect, after observation time T, there are a total of [number] people within the scope of the distribution strategy's effect. nodes, generating a total of nodes. A user request, where node i is paired with... Respond to each user request.

[0063] 1) Request hit rate

[0064] According to the definition of request hit rate, the request hit rate of node i can be expressed as:

[0065] .

[0066] 2) Average response time

[0067] After the distribution strategy takes effect, the response latency of all users within the scope of the distribution strategy is statistically analyzed. Since the responses of user requests come from different nodes, the user requests responding to node i are analyzed.

[0068] .

[0069] 3) Peak bandwidth

[0070] In high-bandwidth services, traffic from upstream nodes to downstream nodes is mainly generated by cached push or pull operations based on policies. This type of traffic typically consumes network bandwidth during periods of network idle, and its impact on user experience is negligible. However, peak bandwidth from node to user significantly affects network service quality, thereby impacting user experience. Therefore, peak bandwidth from node to user needs to be given serious consideration.

[0071] Peak bandwidth of node i During the observation period Downlink bandwidth sample for high-traffic services The maximum value.

[0072] .

[0073] Considering only peak bandwidth is insufficient to fully characterize the network load of high-traffic services during the observation period. It is also necessary to consider the proportion of peak bandwidth to the maximum bandwidth allocated to high-traffic services by the node (if no bandwidth is allocated, the node's network bandwidth is used as the maximum bandwidth), and the average bandwidth of high-traffic services during the observation period. Here, we use the bandwidth-weighted average-to-peak ratio for evaluation, assuming that the bandwidth limit for high-traffic services at node i is... The node bandwidth weighted average peak-to-average bandwidth ratio is defined as:

[0074] .

[0075] 4) Node storage rate

[0076] Based on the definition of node storage rate given in the previous section, if the total storage allocated to node i is... The expected storage utilization ratio is During the observation time, L storage samples are taken from node i. Generally, the rate of change in storage caused by adjustments to the distribution strategy is relatively small. Therefore, we use the average node storage rate to evaluate storage performance. The average node storage rate is defined as:

[0077] .

[0078] 5) Change in cache access rate

[0079] During the observation period, the number of resources pulled down by node i was... The number of resources deleted is The number of times the user accessed the dropdown resource was The number of accesses to the deleted resource was The pull-down cache access rate is then... Cache access rate The pull-down cache access rate is a positive indicator, while the delete cache access rate is a negative indicator. Therefore, the update cache access rate can be defined as:

[0080] .

[0081] 6) The impact of node type on evaluation indicators

[0082] From the perspective of the overall system architecture, nodes can be divided into three types: central nodes, regional nodes, and edge nodes. When statistically analyzing the same evaluation indicators for different types of nodes, we use the request hit rate of each type of node as the weight of the same evaluation indicator. That is, the higher the request hit rate of a node, the greater its impact on the system. Therefore, the final evaluation indicator can be expressed as:

[0083] ① Average weighted response time:

[0084] ② Average weighted bandwidth peak-to-sweep ratio:

[0085] ③ Average weighted storage ratio:

[0086] ④ Weighted update cache access rate:

[0087] in, .

[0088] Next, the indicators are normalized to eliminate the influence of dimensions. The range normalization method is used to map the indicator values ​​to the [0,1] interval:

[0089] For positive metrics (the higher the value, the better), such as request hit rate and cache update efficiency, the normalization formula is:

[0090] ;

[0091] For inverse metrics (the smaller the value, the better), such as user latency and bandwidth consumption, the normalization formula is:

[0092] ;

[0093] For metrics with an optimal value (such as storage usage, which is neither too high nor too low is ideal), interval approximation can be used for normalization.

[0094]

[0095] in, This represents the expected storage usage.

[0096] Through the above quantification and normalization processes, all evaluation indicators are converted into dimensionless values ​​in the range [0,1], eliminating the dimensional differences between different indicators and facilitating the construction of subsequent comprehensive evaluation models.

[0097] After quantifying and normalizing each evaluation indicator, the next step is to construct an evaluation function that comprehensively reflects the effect of accelerated distribution. The comprehensive evaluation function aims to integrate the individual indicators according to certain rules to form a quantitative assessment of the overall performance of the distribution system.

[0098] The comprehensive evaluation function E is usually expressed as a weighted sum:

[0099]

[0100] in, For the first A normalized evaluation index For the first The weight of each indicator, Total number of indicators. Weight The determination of the weighted summation model reflects the relative importance of each indicator in the overall evaluation. The reason for adopting the weighted summation model is that it can simply and effectively integrate information from multiple indicators to obtain a comprehensive evaluation value. This model is intuitive and operable, making it easy to understand and apply. To ensure that the comprehensive evaluation function accurately reflects system performance, it is necessary to analyze the interrelationships between the indicators. For example, a high request hit rate and low user latency usually indicate good system performance, while excessive bandwidth consumption may mean low resource utilization efficiency. By reasonably integrating these indicators, the comprehensive evaluation function can provide a balanced assessment. Furthermore, considering that the importance of each indicator may differ under different business scenarios, the comprehensive evaluation function should have a certain degree of flexibility, capable of adjusting the weight allocation according to actual needs. This provides possibilities for subsequent model optimization and strategy formulation. The construction of the comprehensive evaluation function in this study provides a foundation for the quantitative evaluation of accelerated distribution effects. Through this function, multi-dimensional performance indicators can be transformed into a comprehensive evaluation value, providing a scientific basis for identifying system bottlenecks and formulating optimization strategies.

[0101] In a second embodiment of the present invention, the present invention provides an electronic device, including a memory and a processor, characterized in that the memory stores a computer program that can run on the processor, and the program executed on the processor implements the steps in the multi-dimensional evaluation method for optimizing content distribution strategies as described above.

[0102] In a fourth embodiment of the present invention, the present invention provides a storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the steps in the multi-dimensional evaluation method for optimizing content distribution strategies as described above.

[0103] In summary, this invention provides a multi-dimensional evaluation method for optimizing content distribution strategies. From the construction of an evaluation indicator system, quantification and normalization processing, and the development of a comprehensive evaluation function, it systematically addresses the limitations of traditional distribution effectiveness evaluation methods. Through scientifically sound indicator design and real-time data analysis, this invention not only comprehensively reflects the multi-dimensional characteristics of user experience and system performance but also provides clear guidance for the comprehensive optimization of distribution strategies. With the continuous expansion of internet business and the increasing diversification of user needs, the establishment of this indicator system will help drive the distribution system towards greater efficiency and stability.

[0104] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the modules described above can be referred to the corresponding process in the aforementioned method implementation, and will not be repeated here.

[0105] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0106] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or in a combination of hardware and software functional modules.

[0107] The integrated modules implemented as software functional modules described above can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer system (which may be a personal computer, server, or network system, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A multi-dimensional evaluation method for optimizing content distribution strategies, characterized in that: include, An evaluation index system is constructed based on the observation data of the running status after the distribution strategy takes effect; the evaluation index system includes the request hit rate, average response latency, bandwidth-weighted average peak-to-average ratio, average node storage rate, and update cache access rate. The evaluation indicators in the constructed evaluation indicator system are quantified and normalized to eliminate the influence of different dimensions between the evaluation indicators. Based on the normalization results of the above evaluation indicators, a comprehensive evaluation function is constructed to form a quantitative assessment of the distribution strategy.

2. The method according to claim 1, characterized in that: The expression for the request hit rate is: ， in, The request hit rate for node i within the scope of the distribution strategy. The number of user requests responded to by node i. The number of user requests generated within the scope of a distribution strategy.

3. The method according to claim 1, characterized in that: The expression for the average response delay is: ， in, The average response latency of user requests to node i within the scope of the distribution strategy. The number of user requests responded to by node i. The response latency of user request j to node i.

4. The method according to claim 1, characterized in that: The expression for the bandwidth-weighted average peak-to-peak ratio is: ， in, The bandwidth-weighted average peak-to-peak ratio of node i within the scope of the distribution strategy, where K is the number of traffic services from node i to the user during the observation period. Let be the downlink bandwidth of the traffic service k from node i to user k. The bandwidth limit for node i. Let be the peak bandwidth of node i.

5. The method according to claim 1, characterized in that: The expression for the average storage rate of the nodes is: ， in, The average storage rate of node i within the scope of the distribution strategy. The total storage allocated to node i. Let L be the desired storage utilization ratio, and L be the number of storage samples for node i during the observation period. This is the l-th storage sample for node i.

6. The method according to claim 1, characterized in that: The expression for updating the cache access rate is: ， in, The update cache access rate of node i within the scope of the distribution strategy. This represents the number of times a user accesses the q-th dropdown resource of node i within the observation period. This represents the number of times a user accesses the d-th deleted resource at node i within the observation period. The number of user requests responded to by node i.

7. The method according to claim 1, characterized in that: The method also includes, The final evaluation metric is determined by using the request hit rate as the weight for other evaluation metrics: ① Average weighted response time: , ② Average weighted bandwidth peak-to-sweep ratio: , ③ Average weighted storage ratio: , ④ Weighted update cache access rate: , in, The request hit rate for node i within the scope of the distribution strategy. The average response latency of user requests to node i within the scope of the distribution strategy. The bandwidth-weighted average peak-to-peak ratio of node i within the scope of the distribution strategy. The average storage rate of node i within the scope of the distribution strategy. The update cache access rate of node i within the scope of the distribution strategy.

8. The method according to claim 1, characterized in that: The comprehensive evaluation function is a weighted sum of the normalized results of the evaluation indicators.

9. The method according to claim 1, characterized in that: The method also includes, Set a threshold for evaluating the distribution strategy. If the current comprehensive evaluation function value is higher than the threshold, it means that the current distribution strategy is reasonable. Otherwise, it means that the current distribution strategy is unreasonable, and the distribution strategy is adjusted based on expert knowledge.

10. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-9.

Images