Distributed cache optimization method and device, storage medium and product

Abstract

The invention discloses a distributed cache optimization method and device, a storage medium and a product, and relates to the technical field of distributed caches, and the method comprises the steps: determining the overall distribution type of a data access mode through a low-precision sorter; the overall distribution types are classified through a high-precision sorter; and determining an allocation strategy of the cache resources through the cache allocator according to the distribution parameter combination under each overall distribution type, and performing optimal allocation of the cache resources according to the allocation strategy. According to the method, the overall distribution type of data access is identified through the low-precision sorter, so that the consumption of computing resources is reduced; the distributed parameters are refined through the high-precision sorter, the accuracy of data feature capture is improved, the hierarchical processing mode not only improves the accuracy and efficiency of decision making, but also can adapt to different access modes and maintain efficient cache management, and the problem that various existing algorithms are insufficient in universality in specific scenes is solved.

Description

Technical Field

[0001] The present application relates to the field of distributed cache technology, and in particular to a distributed cache optimization method, device, storage medium and product. Background Art

[0002] In a distributed cache system, effective use of cache resources and improvement of cache hit rate are key factors in improving system performance. With the rapid growth of data size and increasingly complex access patterns, how to achieve efficient data access with limited cache resources has become an important issue that needs to be solved. Efficient cache management can not only significantly reduce data access latency, but also greatly improve the overall throughput of the system, which is of vital importance for the performance optimization of large-scale distributed systems.

[0003] Existing cache optimization technologies are mainly divided into two categories: cache strategies based on classic algorithms and intelligent cache methods based on machine learning. Existing cache strategies such as Least Recently Used (LRU) and Least Frequently Used (LFU), although simple to implement and with low computational overhead, often fail to fully tap the potential of the cache when faced with complex and changing access patterns. Although machine learning-based methods have improved cache efficiency to a certain extent, they still face problems such as high model complexity, high training overhead, and lack of real-time performance when dealing with large-scale data and dynamic environments. How to solve the low cache hit rate and poor technical versatility in existing cache optimization methods has become a problem that needs to be solved urgently. Summary of the invention

[0004] The main purpose of this application is to provide a distributed cache optimization method, device, storage medium and product, aiming to solve the technical problems of low cache hit rate and poor technical versatility in existing cache optimization methods.

[0005] To achieve the above objectives, the present application proposes a distributed cache optimization method, the method comprising:

[0006] Determine the overall distribution type of data access patterns through low-precision sorters;

[0007] Classifying the overall distribution type by a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of the data access pattern;

[0008] The cache allocator determines the allocation strategy of cache resources according to the distribution parameter combination under each of the overall distribution types, and optimizes the allocation of cache resources according to the allocation strategy.

[0009] In one embodiment, the step of determining the overall distribution type of the data access pattern by a low-precision sorter includes:

[0010] Counting access characteristic parameters of the data access pattern by means of a low-precision sorter, wherein the access characteristic parameters include data access frequency, access interval time, number of accesses of a data item, and time series of a data item;

[0011] Performing parameter estimation on a plurality of preset distribution types according to the access characteristic parameters according to a maximum likelihood estimation method, and obtaining a likelihood value corresponding to each of the preset distribution types;

[0012] Calculating the Bayesian information value corresponding to each of the preset distribution types according to the likelihood value by using the Bayesian information criterion;

[0013] The preset distribution type with the smallest Bayesian information value is used as the overall distribution type of the data access pattern.

[0014] In one embodiment, the step of classifying the overall distribution type by a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of the data access pattern includes:

[0015] Extracting high-dimensional features corresponding to the overall distribution type and the access characteristic parameters according to the autoencoder;

[0016] Clustering the extracted high-dimensional features in a low-dimensional feature space to obtain a clustering result;

[0017] The preset distribution key parameters under each of the overall distribution types are updated according to the clustering results to obtain a distribution parameter combination under each of the overall distribution types of the data access pattern.

[0018] In one embodiment, the step of determining, by a cache allocator, a cache resource allocation strategy according to the distribution parameter combination under each of the overall distribution types, and performing optimal allocation of cache resources according to the allocation strategy includes:

[0019] Determining the priority weight of each of the overall distribution types according to the distribution parameter combination and the access characteristic parameter by a cache allocator;

[0020] Determining a cache space for each of the overall distribution types according to the priority weight;

[0021] Determine the data unit value of each data item according to the data access frequency and data size of each data item in each of the overall distribution types;

[0022] The target data item in each of the overall distribution types is determined according to the data unit value, and the target data item is loaded into the cache space corresponding to each of the overall distribution types.

[0023] In one embodiment, after the step of determining the allocation strategy of cache resources according to the distribution parameter combination under each of the overall distribution types by the cache allocator and optimizing the allocation of cache resources according to the allocation strategy, the method further includes:

[0024] Monitor the data access pattern in real time through a dynamic enhanced cache optimizer to obtain real-time data and data access characteristics corresponding to the real-time data, wherein the data access characteristics include data access frequency, access interval time, number of accesses of a data item, size of a data item, and time series of a data item;

[0025] Perform prediction based on a preset multivariate time series analysis model, the real-time data, and the data access characteristics to obtain a predicted data access pattern;

[0026] Modeling the cache resource allocation strategy as a convex optimization problem based on the predicted data access pattern;

[0027] Solving the convex optimization problem to obtain an optimized cache allocation strategy;

[0028] The current cache space is adjusted according to the cache resource allocation strategy and the optimized cache allocation strategy.

[0029] In one embodiment, the step of solving the convex optimization problem to obtain an optimized cache allocation strategy includes:

[0030] Determining an initial point within the feasible domain of the convex optimization problem;

[0031] Determine the gradient and Hessian matrix of the current solution by an interior point method based on the initial point;

[0032] The direction and step size of the current solution are iteratively updated according to the gradient and the Hessian matrix until an optimal solution is found, and the optimal solution is used as the optimized cache allocation strategy.

[0033] In one embodiment, the step of adjusting the current cache space according to the cache resource allocation strategy and the optimized cache allocation strategy includes:

[0034] Calculating objective function values ​​corresponding to various objective functions for each cache allocation scheme in the cache resource allocation strategy and the optimized cache allocation strategy, wherein the various objective functions include objective functions corresponding to cache hit rate, resource utilization rate, and system response time, respectively;

[0035] Sort each cache allocation scheme according to the objective function value to obtain a non-dominated solution;

[0036] The non-dominated solution is added to the Pareto front, and the current cache space is adjusted according to the Pareto front.

[0037] In addition, to achieve the above purpose, the present application also proposes a distributed cache optimization device, which includes:

[0038] An overall distribution type determination module, used for determining the overall distribution type of the data access pattern by a low-precision sorter;

[0039] A distribution parameter combination determination module, used for classifying the overall distribution type by a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of a data access mode;

[0040] The cache resource optimization module is used to determine the cache resource allocation strategy according to the distribution parameter combination under each overall distribution type through a cache allocator, and optimize the cache resource allocation according to the allocation strategy.

[0041] In addition, to achieve the above-mentioned purpose, the present application also proposes a distributed cache optimization device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the distributed cache optimization method described above.

[0042] In addition, to achieve the above objectives, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the distributed cache optimization method described above are implemented.

[0043] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the distributed cache optimization method described above are implemented.

[0044] The present application provides a distributed cache optimization method, which determines the overall distribution type of data access patterns through a low-precision sorter; classifies the overall distribution type through a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of the data access pattern; determines the allocation strategy of cache resources according to the distribution parameter combination under each overall distribution type through a cache allocator, and optimizes the allocation of cache resources according to the allocation strategy. The present application identifies the overall distribution type of data access through a low-precision sorter to reduce computing resource consumption; refines the distribution parameters through a high-precision sorter to improve the accuracy of capturing data features. This hierarchical processing method not only improves the accuracy and efficiency of decision-making, but also can adapt to different access patterns, maintain efficient cache management, and solve the problem of insufficient versatility of various existing algorithms in specific scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0046] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0047] Figure 1 A flow chart of the first embodiment of the distributed cache optimization method of the present application;

[0048] Figure 2 An example diagram of each preset distribution type in the distributed cache optimization method of this application;

[0049] Figure 3 A flow chart of the second embodiment of the distributed cache optimization method of the present application;

[0050] Figure 4 A schematic diagram of the process flow provided for the third embodiment of the distributed cache optimization method of the present application;

[0051] Figure 5 A flowchart of the fourth embodiment of the distributed cache optimization method of the present application is provided;

[0052] Figure 6 This is the overall architecture diagram of the distributed cache optimization method of this application;

[0053] Figure 7 This is a schematic diagram of the module structure of the distributed cache optimization device according to an embodiment of the present application;

[0054] Figure 8This is a schematic diagram of the device structure of the hardware operating environment involved in the distributed cache optimization method in the embodiment of the present application.

[0055] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0056] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0057] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0058] The main solution of the embodiment of the present application is: determine the overall distribution type of the data access pattern through a low-precision sorter; classify the overall distribution type through a high-precision sorter to obtain the distribution parameter combination under each of the overall distribution types of the data access pattern; determine the allocation strategy of cache resources according to the distribution parameter combination under each of the overall distribution types through a cache allocator, and optimize the allocation of cache resources according to the allocation strategy.

[0059] Existing cache optimization technologies are mainly divided into two categories: cache strategies based on classic algorithms and intelligent cache methods based on machine learning. Existing cache strategies such as LRU (Least Recently Used) and LFU (Least Frequently Used), although simple to implement and with low computational overhead, often fail to fully tap the potential of the cache when faced with complex and changing access patterns. Although machine learning-based methods have improved cache efficiency to a certain extent, they still face problems such as high model complexity, high training overhead, and lack of real-time performance when dealing with large-scale data and dynamic environments. How to solve the low cache hit rate and poor technical versatility in existing cache optimization methods has become an urgent problem to be solved.

[0060] The present application provides a solution, which determines the overall distribution type of data access patterns through a low-precision sorter; classifies the overall distribution type through a high-precision sorter to obtain the distribution parameter combination under each overall distribution type of the data access pattern; determines the allocation strategy of cache resources according to the distribution parameter combination under each overall distribution type through a cache allocator, and optimizes the allocation of cache resources according to the allocation strategy. The present application identifies the overall distribution type of data access through a low-precision sorter to reduce computing resource consumption; refines the distribution parameters through a high-precision sorter to improve the accuracy of data feature capture. This hierarchical processing method not only improves the accuracy and efficiency of decision-making, but also can adapt to different access patterns, maintain efficient cache management, and solve the problem of insufficient versatility of various existing algorithms in specific scenarios.

[0061] It should be noted that the execution subject of the method of this embodiment can be a computing service device with distributed cache optimization, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc.; it can also be a distributed cache optimization device with the same or similar functions. This embodiment and the following embodiments will be described by taking the distributed cache optimization device as an example.

[0062] Based on this, the embodiment of the present application provides a distributed cache optimization method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the distributed cache optimization method of the present application.

[0063] In this embodiment, the distributed cache optimization method includes steps S10 to S30:

[0064] Step S10, determining the overall distribution type of the data access pattern by a low-precision sorter.

[0065] It can be understood that the distributed cache optimization method proposed in this embodiment realizes the accurate capture and management of data access patterns through the design of two-layer sorters, thereby optimizing the allocation of cache resources. The main task of the low-precision sorter is to quickly identify the overall distribution type of data access. By utilizing combined methods such as maximum likelihood estimation, chi-square test, Bayesian information criterion, etc., the low-precision sorter can quickly locate the most suitable distribution category in the data access pattern, such as uniform distribution, normal distribution, exponential distribution, etc. This process not only reduces the consumption of computing resources, but also provides accurate input for subsequent high-precision sorters.

[0066] It should be understood that the low-precision sorter output is the only overall distribution type after the sorter recognizes it. The overall distribution type can be one of multiple preset distribution types, which include but are not limited to the eight basic distribution types in the following table. Users can also further enrich the distribution types according to their scene characteristics:

[0067] Table 1: Preset distribution types

[0068]

[0069]

[0070] In order to more intuitively demonstrate the characteristics of each preset distribution type of the device, the example diagram of each distribution can be referred to Figure 2 .

[0071] Step S20, classifying the overall distribution type by a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of the data access pattern.

[0072] It should be understood that after the low-precision sorter determines the overall distribution type, the high-precision sorter further refines the distribution parameters, such as mean, standard deviation, etc., to obtain the distribution parameter combination under each overall distribution type of the data access pattern, thereby constructing an accurate access pattern model. This layer of refinement provides the cache allocator with detailed distribution information and optimizes the allocation strategy of cache resources.

[0073] Step S30: determining, by a cache allocator, a cache resource allocation strategy according to the distribution parameter combination under each of the overall distribution types, and optimizing the allocation of cache resources according to the allocation strategy.

[0074] It can be understood that the cache allocator formulates a cache resource allocation strategy based on the access mode output by the high-precision sorter, that is, information related to the overall distribution type and the distribution parameter combination under each overall distribution type. According to the characteristics of different distribution types, the cache allocator allocates appropriate cache space based on the frequency of data access and the inherent properties of the distribution, formulates a corresponding cache resource allocation strategy, and optimizes the allocation of cache resources according to the allocation strategy.

[0075] This embodiment provides a distributed cache optimization method, which determines the overall distribution type of data access patterns through a low-precision sorter; classifies the overall distribution type through a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of the data access pattern; determines the allocation strategy of cache resources according to the distribution parameter combination under each overall distribution type through a cache allocator, and optimizes the allocation of cache resources according to the allocation strategy. This embodiment identifies the overall distribution type of data access through a low-precision sorter to reduce computing resource consumption; refines the distribution parameters through a high-precision sorter to improve the accuracy of capturing data features. This hierarchical processing method not only improves the accuracy and efficiency of decision-making, but also can adapt to different access patterns, maintain efficient cache management, and solve the problem of insufficient versatility of various existing algorithms in specific scenarios.

[0076] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can refer to the above introduction, and will not be repeated later. Figure 2 , step S10, the distributed cache optimization method further includes steps S101 to S104:

[0077] Step S101, collecting access characteristic parameters of the data access pattern through a low-precision sorter, wherein the access characteristic parameters include data access frequency, access interval time, number of accesses to a data item, and time series of a data item.

[0078] It should be noted that the core task of the low-precision sorter is to quickly identify the overall distribution type of data access and provide accurate input for the subsequent high-precision sorter. The sorter uses mathematical methods to quickly determine the distribution type of data by statistically analyzing the data access frequency, interval time and other characteristics. Below, the core principles of the low-precision sorter are explained from three levels: input, distribution fitting method and output.

[0079] It can be understood that the input of the low-precision sorter includes the following access characteristic parameters: data access frequency (λ), access interval time (Δt), number of accesses of the data item (n), and time series of the data item (T = {t1, t2, ..., t n}). Among them, data access frequency (λ) refers to the number of accesses to a data item per unit time. It is defined as: λ i =n i / T i , where n i represents the total number of accesses to data item i, T i represents the observation time window. The access interval (Δt), that is, the time interval between two consecutive accesses, is defined as:

[0080] Δtij =t ij -t i,(j-1) ,

[0081] Among them, t ij represents the jth access time of the ith data item. The number of accesses of a data item (n) is the total number of accesses to a data item during the observation period. The time series of a data item (T) is the sequence of all access time points of a data item.

[0082] Step S102: performing parameter estimation on a plurality of preset distribution types according to the access characteristic parameters using a maximum likelihood estimation method to obtain a likelihood value corresponding to each of the preset distribution types.

[0083] It is worth noting that in this embodiment, a low-precision sorter uses a combination of maximum likelihood estimation (MLE) and Bayesian Information Criterion (BIC) to quickly identify the distribution type of data. Maximum likelihood estimation fits the data distribution by finding the parameter value that is most likely to produce the observed data, while the Bayesian Information Criterion selects the optimal distribution type by penalizing the complexity of the model. In actual operation, the low-precision sorter first estimates the parameters of each preset distribution type (such as uniform distribution, normal distribution, exponential distribution, etc.) through maximum likelihood estimation to obtain their respective likelihood values.

[0084] Specifically, maximum likelihood estimation fits the data distribution by finding the parameter values ​​that are most likely to produce the observed data. For example, for a normal distribution, MLE is used to estimate the mean μ and standard deviation σ, then

[0085] as well as

[0086] where x i Represents the i-th observation value, which is obtained by the input information of the low-precision sorter. For each preset distribution type, the MLE method is used to estimate the parameters and obtain the respective likelihood values ​​(L).

[0087] Step S103, calculating the Bayesian information value corresponding to each of the preset distribution types according to the likelihood value using the Bayesian Information Criterion.

[0088] It should be understood that the Bayesian Information Criterion (BIC) selects the optimal distribution type by penalizing the model complexity. The Bayesian Information Criterion is used to calculate the Bayesian Information Quantity (BIC value) corresponding to each preset distribution type according to the likelihood value. The calculation formula of BIC is:

[0089] BIC = kln(n)-2ln(L),

[0090] Where k is the number of model parameters, n is the number of samples, and L is the maximum likelihood value.

[0091] Step S104: taking the preset distribution type with the smallest Bayesian information value as the overall distribution type of the data access pattern.

[0092] It can be understood that the BIC values ​​of all distribution types can be compared, and the distribution type with the smallest BIC value can be selected as the identification result. This process can be expressed as:

[0093] F(λ,Δt,n,T)=argmin d∈D BIC(d),

[0094] Where D represents the set of all candidate distribution types, and the F function returns the distribution type with the smallest BIC value. The final output is the only distribution type identified by the low-precision sorter, which is one of the eight basic distribution types shown in Table 1 above.

[0095] In a feasible implementation, step S20 may include steps S201 to S203:

[0096] Step S201: extracting high-dimensional features corresponding to the overall distribution type and the access characteristic parameters according to the autoencoder.

[0097] It should be noted that the main task of the high-precision sorter is to further refine the classification of various distributions identified by the low-precision sorter and determine the specific parameter combination under each distribution in order to accurately capture the data access pattern. On the one hand, this process needs to pursue the accuracy of classification, and on the other hand, it is necessary to avoid excessive computational consumption. To this end, this embodiment proposes an adaptive distribution classification model DCM (Deep Clustering Model) based on deep clustering.

[0098] It can be understood that the DCM model extracts high-dimensional features of data (including the overall distribution type identified by the low-precision sorter, the data access frequency λ, the access interval time Δt, the access frequency n of the data item, and the time series T of the data item), specifically through the autoencoder. The autoencoder consists of an encoder and a decoder. The input data is mapped to a low-dimensional feature space through the encoder and then reconstructed back to the original space through the decoder. The goal of the autoencoder is to minimize the difference between the input data and the reconstructed data, so as to effectively extract the feature representation of the data. Assume that the input data is, and the autoencoder extracts features through the following process:

[0099] Z=fencoder (X),

[0100]

[0101] Among them, Z is the low-latitude feature representation, To reconstruct the data. The autoencoder minimizes the reconstruction error Optimize the feature extraction process.

[0102] Step S202: clustering the extracted high-dimensional features in a low-dimensional feature space to obtain a clustering result.

[0103] It is understandable that after feature extraction, the DCM model clusters the extracted high-dimensional features in the low-dimensional feature space and divides the data into the most appropriate categories, that is, the distribution under different parameter combinations (such as the binomial distribution under a certain (a, b) parameter combination or the normal distribution under a certain (μ, σ) combination, etc.). The advantage of clustering in the feature space is that the high-dimensional features extracted by the above process can be used to improve the accuracy and robustness of clustering. In the actual operation process, the choice of clustering method is very flexible, such as using clustering methods such as K-Means or Gaussian mixture model (GMM).

[0104] Step S203: updating preset distribution key parameters under each of the overall distribution types according to the clustering result, and obtaining distribution parameter combinations under each of the overall distribution types of the data access pattern.

[0105] It is understandable that in order to determine the specific distribution parameters of each category, the DCM algorithm updates the parameters of the clustering results. For example, for uniform distribution, the lower and upper limits of each sub-distribution are updated. For normal distribution, the mean and standard deviation of each sub-distribution are updated. The following Table 2 summarizes the key parameter values ​​of the eight basic distributions that the high-precision sorter needs to learn and estimate. These parameters affect the characteristics of the distribution and will in turn affect the subsequent allocation of cache space. The table also shows the parameter update formula for each distribution.

[0106] Table 2: Key parameters of the estimated distribution of high-precision sorters

[0107]

[0108]

[0109] In this embodiment, the high-precision sorter further refines the classification of various distributions identified by the low-precision sorter and determines the specific parameter combination under each distribution in order to accurately capture the data access pattern, thereby improving the classification accuracy and avoiding excessive computing consumption.

[0110] In this embodiment, a low-precision sorter is used to quickly identify the overall distribution type of data access, providing accurate input for the subsequent high-precision sorter. The sorter uses mathematical methods to quickly determine the distribution type of data by statistically analyzing the data access frequency, interval time and other characteristics. Then, the high-precision sorter further refines and classifies the various distributions identified by the low-precision sorter, and determines the specific parameter combination under each distribution in order to accurately capture the data access pattern. This layer of refined processing provides detailed distribution information for the cache allocator and optimizes the allocation strategy of cache resources. The multi-precision sorter is set to improve the accuracy of fitting on the basis of ensuring the high efficiency of data distribution fitting.

[0111] Based on the first embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the first embodiment can be referred to the above description, and will not be described in detail later. Figure 4 , step S30, the distributed cache optimization method further includes steps S301 to S304:

[0112] Step S301: determining the priority weight of each of the overall distribution types according to the distribution parameter combination and access characteristic parameters by a cache allocator.

[0113] It is worth noting that the main task of the cache allocator in this embodiment is to reasonably allocate the limited cache space according to the distribution information provided by the first two layers of sorters to maximize the cache hit rate. The new cache allocation strategy proposed in this embodiment is called the Priority Distribution Allocation (PDA) method. The PDA method combines the distribution characteristics of data access and specific parameter combinations, and accurately allocates cache resources and improves the cache hit rate through a priority weight mechanism. The core idea of ​​the PDA method is to determine the priority weight of each distribution based on different distribution types and their specific parameter combinations, and allocate cache space accordingly.

[0114] It can be understood that, first, for each overall distribution type, its priority weight is calculated according to its access characteristic parameters (including access frequency, access interval, access times, data item size and time series characteristics) and specific parameter combinations. Assuming there are N types of distributions in total, the sum of all sub-categories in the eight basic distributions is N. i represents the number of sub-distributions in the i-th basic distribution, then

[0115]

[0116] The number of data in each distribution is n i, the total cache space is C. The calculation of priority weights needs to be combined with the characteristics of the distribution, such as mean, variance, etc. Here, we take uniform distribution as an example. If the mean is high, it means that data arrives frequently and regularly, so a larger cache space should be allocated. Specifically, for the i-th uniform distribution, the priority weight can be expressed as:

[0117] w i =αμ i +βσ i +γn i ,

[0118] Among them, μ i represents the mean of the distribution, σ i represents the standard deviation of the distribution, n i Indicates the number of data, α, β, γ are weight coefficients, which are used to balance the influence of different parameters. The other basic distributions adopt similar ideas, and their priority weights can be obtained by combining their parameter characteristics.

[0119] Step S302: determining a cache space for each of the overall distribution types according to the priority weight.

[0120] It is understandable that the priority weight w i , allocate cache space C. The cache space c allocated to the i-th distribution i for:

[0121]

[0122] Step S303: determining the data unit value of each data item according to the data access frequency and data size of each data item in each of the overall distribution types.

[0123] It is worth noting that when allocating cache, in addition to the frequency of data access, the size of the data itself is also an important consideration. Frequently accessed big data items are not necessarily suitable for caching, because they may take up a lot of cache space, thus affecting the overall cache efficiency. This embodiment introduces an indicator that comprehensively measures the frequency of data access and the size of data, namely, the value per unit size (VPUS), and defines its calculation method:

[0124] VPUS i =F i / S i ,

[0125] Among them, F i represents the access frequency of the i-th data item, S iIndicates the size of the i-th data item. The higher the unit value, the higher the cost performance of the data item in the cache. The cache allocation method in this embodiment is performed based on the unit value of the data.

[0126] Step S304: determining a target data item in each of the overall distribution types according to the data unit value, and loading the target data item into a cache space corresponding to each of the overall distribution types.

[0127] It should be understood that after determining the cache space for each distribution, it is necessary to further select specific data items to load into the cache. For each distribution, the target data items in each overall distribution type are determined based on the data unit value, and the target data items are loaded into the cache space corresponding to each overall distribution type.

[0128] In this embodiment, the cache allocator determines the priority weight of each overall distribution type according to the distribution parameter combination and the access characteristic parameter; determines the cache space of each overall distribution type according to the priority weight; determines the data unit value of each data item according to the data access frequency and data size of each data item in each overall distribution type; determines the target data item in each overall distribution type according to the data unit value, and loads the target data item into the cache space corresponding to each overall distribution type. According to the distribution information provided by the first two layers of sorters, the limited cache space is reasonably allocated to maximize the cache hit rate. Combined with the distribution characteristics of data access and the specific parameter combination, the cache resources are accurately allocated through the priority weight mechanism, thereby improving the cache hit rate.

[0129] Based on the first embodiment of the present application, in the fourth embodiment of the present application, the same or similar contents as those in the first embodiment can refer to the above description, and will not be described in detail later. Figure 5 After step S30, the distributed cache optimization method further includes steps S40 to S80:

[0130] Step S40, monitor the data access pattern in real time through the dynamic enhanced cache optimizer to obtain real-time data and data access characteristics corresponding to the real-time data, wherein the data access characteristics include data access frequency, access interval time, number of accesses of data items, size of data items and time series of data items.

[0131] It is worth noting that in this embodiment, a dynamic enhanced cache optimizer is also provided for efficient management and dynamic adjustment of cache resources. The main function of the dynamic enhanced cache optimizer (DECO) is to monitor the changes in data access patterns in real time and dynamically adjust the cache strategy according to these changes, thereby improving the cache hit rate and system performance. DECO achieves efficient management and dynamic adjustment of cache resources by combining real-time data flow analysis technology and mathematical optimization methods.

[0132] First, DECO uses an efficient data stream monitoring system to capture data access characteristics in real time (including access frequency, access interval, access times, size of data items, and time series characteristics of real-time data and data stored in the cache). In order to process these massive amounts of data, DECO adopts a real-time data stream processing method based on sliding windows. The sliding window technology allows the system to process and analyze data access characteristics within a fixed time interval, thereby capturing the latest changes in access patterns.

[0133] Step S50, performing prediction based on a preset multivariate time series analysis model, the real-time data and the data access characteristics to obtain a predicted data access pattern.

[0134] It can be understood that after obtaining the real-time data and the data access features corresponding to the real-time data, DECO applies a prediction model based on multivariate time series analysis to predict future data access patterns. Future data access patterns refer to the predicted access frequency and regularity of data items in the future. The specific implementation steps include: (1) collecting historical access logs (i.e., the real-time data and the data access features corresponding to the real-time data obtained in the above steps) to form time series data; (2) preprocessing the time series data to remove noise and outliers; (3) applying multivariate time series analysis models (such as ARIMA, LSTM, etc.) for training to predict future access frequencies and patterns.

[0135] Step S60: modeling the cache resource allocation strategy as a convex optimization problem according to the predicted data access pattern.

[0136] It is understandable that DECO adopts a cache allocation adjustment and optimization strategy based on convex optimization based on predicting future access patterns. The goal of this strategy is to maximize the cache hit rate while minimizing the waste of cache resources. Specifically, DECO models the cache allocation problem as a convex optimization problem and determines the optimal cache allocation solution by solving the optimization problem, that is, adjusting the existing cache space.

[0137] Step S70, solving the convex optimization problem to obtain an optimized cache allocation strategy.

[0138] It should be understood that the optimal cache allocation solution is determined by solving the optimization problem, that is, adjusting the existing cache space. The objective function of the optimization problem is defined as:

[0139]

[0140] Where c={c1,c2,…,c n} represents the cache allocation for each data item (including the data items currently in the cache and the data items that may need to be cached according to prediction), α i ,β i ,γ i They respectively represent the weight coefficient, cache hit rate gain coefficient and cache resource consumption coefficient of the i-th data item. Indicates the current cache allocation. C indicates the total cache space size. δ i Indicates the maximum adjustment allowed for each data item (if there is no special restriction, it can be set to a sufficiently large number M). By solving the above optimization problem, DECO can dynamically adjust the cache allocation amount under different data access modes to achieve optimal cache management. DECO obtains a new cache allocation scheme c by solving this optimization problem.

[0141] Step S80: adjusting the current cache space according to the cache resource allocation strategy and the optimized cache allocation strategy.

[0142] It should be understood that after solving this optimization problem and obtaining a new cache allocation scheme c, the difference between the new and old allocation schemes (c i and ) to adjust the existing cache space.

[0143] In a feasible implementation manner, step S70 may further include steps S701 to S703:

[0144] Step S701, determining an initial point within the feasible domain of the convex optimization problem.

[0145] It should be noted that this embodiment describes the specific process of solving convex optimization problems. DE CO adopts an efficient solution algorithm based on the interior point method. The interior point method is an iterative algorithm for solving convex optimization problems. Its main idea is to find an initial point in the feasible domain and then gradually approach the optimal solution.

[0146] Step S702, determining the gradient and Hessian matrix of the current solution by an interior point method based on the initial point.

[0147] Step S703, iteratively updating the direction and step size of the current solution according to the gradient and the Hessian matrix until an optimal solution is found, and using the optimal solution as the optimized cache allocation strategy.

[0148] It can be understood that in each iteration, the interior point method calculates the gradient and Hessian matrix of the current point, and uses this information to update the direction and step size of the solution, thereby gradually approaching the optimal solution and using the optimal solution as the optimized cache allocation strategy.

[0149]

[0150] Among them, c k represents the solution of the kth iteration, η k is the step length, and are the gradient and Hessian matrix of the objective function respectively.

[0151] In a feasible implementation manner, step S80 may further include steps S801 to S803:

[0152] Step S801, calculating objective function values ​​corresponding to various objective functions for each cache allocation scheme in the cache resource allocation strategy and the optimized cache allocation strategy, wherein the various objective functions include objective functions corresponding to cache hit rate, resource utilization rate and system response time respectively.

[0153] It should be noted that in order to further improve the dynamic adaptability of the system, DECO combines an adaptive adjustment mechanism based on multi-objective optimization. DECO adopts a multi-objective optimization algorithm based on Pareto optimization. By constructing the Pareto frontier, it searches for a set of non-inferior solutions that perform well on all objective functions. The algorithm can make trade-offs between different optimization objectives to achieve a globally optimal cache allocation strategy. The mechanism makes trade-offs between different optimization objectives by introducing multiple objective functions (such as cache hit rate, resource utilization, and system response time). The objective function values ​​corresponding to each objective function are calculated for each cache allocation scheme in the cache resource allocation strategy and the optimized cache allocation strategy.

[0154] Step S802: sort each cache allocation scheme according to the objective function value to obtain a non-dominated solution.

[0155] Specifically, each cache allocation scheme is sorted according to the objective function value to find a non-dominated solution.

[0156] Step S803: adding the non-dominated solution to the Pareto front, and adjusting the current cache space according to the Pareto front.

[0157] It is understandable that the non-dominated solutions are added to the Pareto frontier and the current cache space is adjusted according to the Pareto frontier. The Pareto frontier consists of the points in the point set that show the best trade-off solution, that is, no other solution is better than the points in the point set at any goal. The boundary formed by these points is the Pareto frontier.

[0158] In this embodiment, the data access pattern is monitored in real time by a dynamic enhanced cache optimizer to obtain real-time data and data access features corresponding to the real-time data; predictions are made based on a preset multivariate time series analysis model, the real-time data and the data access features to obtain a predicted data access pattern; the cache resource allocation strategy is modeled as a convex optimization problem based on the predicted data access pattern; the convex optimization problem is solved to obtain an optimized cache allocation strategy; the current cache space is adjusted based on the cache resource allocation strategy and the optimized cache allocation strategy. The dynamic enhanced cache optimization method based on multivariate time series analysis and convex optimization in this embodiment adopts a strategy that combines real-time monitoring, predictive analysis and optimization adjustment to address the problem of dynamic changes in data access patterns, thereby achieving real-time adjustment of cache strategies in a dynamically changing environment, significantly improving cache hit rates and system performance.

[0159] For example, to help understand the overall implementation process of the distributed cache optimization method obtained by combining this embodiment with the above-mentioned embodiment 1, please refer to Figure 6 , Figure 6 This is the overall architecture diagram of the distributed cache optimization method of this application, specifically:

[0160] This application proposal proposes a distributed cache optimization method based on a multi-precision hierarchical sorter. Through the collaborative work of two-layer sorters and the dynamic enhanced cache optimizer (DECO), efficient cache management in a complex dynamic environment is achieved. Specifically, it includes the following contents:

[0161] (1) End-to-end cache optimization device: This application constructs an end-to-end cache optimization system framework, including a low-precision sorter, a high-precision sorter, a cache allocator, and a dynamically enhanced cache optimizer (DECO). Through the collaborative work of these modules, efficient cache management under complex data access patterns is achieved, ensuring the efficient operation of the system and the real-time update of the cache strategy. The distributed cache setting scheme has significant advantages in theory, and can accurately capture the characteristics of data access and perform scientific and reasonable cache allocation. Compared with existing cache strategies, this application can better adapt to different access patterns through the application of distributed models, ensure the optimal utilization of cache resources, and improve cache hit rate and system performance.

[0162] (2) High- and low-precision sorters: The multi-precision sorter in this proposal includes a two-layer structure of a low-precision sorter and a high-precision sorter. The main task of the low-precision sorter is to quickly identify the overall distribution type of data access and reduce the consumption of computing resources. After the low-precision sorter determines the distribution type, the high-precision sorter further refines the distribution parameters, such as the mean, standard deviation, etc., to construct an accurate access pattern model. This layer of refinement provides detailed distribution information for the cache allocator and optimizes the allocation strategy of cache resources. The multi-precision sorter is set to improve the accuracy of fitting on the basis of ensuring the high efficiency of data distribution fitting.

[0163] (3) DECO dynamic optimization method: This application designs a dynamic enhanced cache optimization method based on multivariate time series analysis and convex optimization. To address the problem of dynamic changes in data access patterns, a strategy combining real-time monitoring, predictive analysis and optimization adjustment is adopted to achieve real-time adjustment of cache strategies in a dynamically changing environment, significantly improving cache hit rates and system performance.

[0164] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the distributed cache optimization method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0165] This application also provides a distributed cache optimization device, please refer to Figure 7 , the distributed cache optimization device comprises:

[0166] An overall distribution type determination module 10, for determining the overall distribution type of the data access pattern by a low-precision sorter;

[0167] A distribution parameter combination determination module 20, configured to classify the overall distribution type by a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of a data access pattern;

[0168] The cache resource optimization module 30 is used to determine the cache resource allocation strategy according to the distribution parameter combination under each of the overall distribution types through a cache allocator, and optimize the cache resource allocation according to the allocation strategy.

[0169] The distributed cache optimization device provided by the present application adopts the distributed cache optimization method in the above embodiment to solve the technical problem. Compared with the prior art, the beneficial effects of the distributed cache optimization device provided by the present application are the same as the beneficial effects of the distributed cache optimization method provided by the above embodiment, and the other technical features in the distributed cache optimization device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0170] The present application provides a distributed cache optimization device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the distributed cache optimization method in the above-mentioned embodiment one.

[0171] Reference below Figure 8 , which shows a schematic diagram of the structure of a distributed cache optimization device suitable for implementing the embodiment of the present application. The distributed cache optimization device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 8 The distributed cache optimization device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0172] like Figure 8As shown, the distributed cache optimization device may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. Various programs and data required for the operation of the distributed cache optimization device are also stored in the RAM 1004. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the distributed cache optimization device to communicate with other devices wirelessly or by wire to exchange data. Although the distributed cache optimization device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have instead.

[0173] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0174] The distributed cache optimization device provided by the present application adopts the distributed cache optimization method in the above embodiment to solve the technical problems of distributed cache optimization. Compared with the prior art, the beneficial effects of the distributed cache optimization device provided by the present application are the same as the beneficial effects of the distributed cache optimization method provided by the above embodiment, and the other technical features in the distributed cache optimization device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0175] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0176] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0177] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the distributed cache optimization method in the above-mentioned embodiment.

[0178] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0179] The computer-readable storage medium may be included in the distributed cache optimization device; or may exist independently without being assembled into the distributed cache optimization device.

[0180] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the distributed cache optimization device, the distributed cache optimization device: determines the overall distribution type of the data access pattern through a low-precision sorter; classifies the overall distribution type through a high-precision sorter to obtain the distribution parameter combination under each of the overall distribution types of the data access pattern; determines the allocation strategy of cache resources according to the distribution parameter combination under each of the overall distribution types through a cache allocator, and optimizes the allocation of cache resources according to the allocation strategy.

[0181] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0182] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0183] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0184] The readable storage medium provided in this application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned distributed cache optimization method, and can solve technical problems. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the distributed cache optimization method provided in the above-mentioned embodiment, and will not be repeated here.

[0185] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned distributed cache optimization method when executed by a processor.

[0186] The computer program product provided by this application can solve technical problems. Compared with the prior art, the beneficial effects of the computer program product provided by this application are the same as the beneficial effects of the distributed cache optimization method provided by the above embodiment, which will not be repeated here.

[0187] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A distributed cache optimization method, characterized in that: The method includes: Determine the overall distribution type of data access patterns through low-precision sorters; Classifying the overall distribution type by a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of the data access pattern; The cache allocator determines the allocation strategy of cache resources according to the distribution parameter combination under each of the overall distribution types, and optimizes the allocation of cache resources according to the allocation strategy.

2. The method according to claim 1, characterized in that The step of determining the overall distribution type of the data access pattern by a low-precision sorter comprises: Counting access characteristic parameters of the data access pattern by means of a low-precision sorter, wherein the access characteristic parameters include data access frequency, access interval time, number of accesses of a data item, and time series of a data item; Performing parameter estimation on a plurality of preset distribution types according to the access characteristic parameters according to a maximum likelihood estimation method, and obtaining a likelihood value corresponding to each of the preset distribution types; Calculating the Bayesian information value corresponding to each of the preset distribution types according to the likelihood value by using the Bayesian information criterion; The preset distribution type with the smallest Bayesian information value is used as the overall distribution type of the data access pattern.

3. The method according to claim 2, characterized in that The step of classifying the overall distribution type by a high-precision sorter to obtain a distribution parameter combination under each overall distribution type of the data access mode includes: Extracting high-dimensional features corresponding to the overall distribution type and the access characteristic parameters according to the autoencoder; Clustering the extracted high-dimensional features in a low-dimensional feature space to obtain a clustering result; The preset distribution key parameters under each of the overall distribution types are updated according to the clustering results to obtain a distribution parameter combination under each of the overall distribution types of the data access pattern.

4. The method according to claim 1, characterized in that The step of determining the allocation strategy of cache resources according to the distribution parameter combination under each of the overall distribution types by the cache allocator, and optimizing the allocation of cache resources according to the allocation strategy includes: Determining the priority weight of each of the overall distribution types according to the distribution parameter combination and the access characteristic parameter by a cache allocator; Determining a cache space for each of the overall distribution types according to the priority weight; Determine the data unit value of each data item according to the data access frequency and data size of each data item in each of the overall distribution types; The target data item in each of the overall distribution types is determined according to the data unit value, and the target data item is loaded into the cache space corresponding to each of the overall distribution types.

5. The method according to claim 1, characterized in that After the step of determining the allocation strategy of cache resources according to the distribution parameter combination under each of the overall distribution types by the cache allocator and optimizing the allocation of cache resources according to the allocation strategy, the method further includes: Monitor the data access pattern in real time through a dynamic enhanced cache optimizer to obtain real-time data and data access characteristics corresponding to the real-time data, wherein the data access characteristics include data access frequency, access interval time, number of accesses of a data item, size of a data item, and time series of a data item; Perform prediction based on a preset multivariate time series analysis model, the real-time data, and the data access characteristics to obtain a predicted data access pattern; Modeling the cache resource allocation strategy as a convex optimization problem based on the predicted data access pattern; Solving the convex optimization problem to obtain an optimized cache allocation strategy; The current cache space is adjusted according to the cache resource allocation strategy and the optimized cache allocation strategy.

6. The method according to claim 5, characterized in that The step of solving the convex optimization problem to obtain an optimized cache allocation strategy includes: Determining an initial point within the feasible domain of the convex optimization problem; Determine the gradient and Hessian matrix of the current solution by an interior point method based on the initial point; The direction and step size of the current solution are iteratively updated according to the gradient and the Hessian matrix until an optimal solution is found, and the optimal solution is used as the optimized cache allocation strategy.

7. The method according to claim 5, characterized in that The step of adjusting the current cache space according to the cache resource allocation strategy and the optimized cache allocation strategy includes: Calculating objective function values ​​corresponding to various objective functions for each cache allocation scheme in the cache resource allocation strategy and the optimized cache allocation strategy, wherein the various objective functions include objective functions corresponding to cache hit rate, resource utilization rate, and system response time, respectively; Sort each cache allocation scheme according to the objective function value to obtain a non-dominated solution; The non-dominated solution is added to the Pareto front, and the current cache space is adjusted according to the Pareto front.

8. A distributed cache optimization device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the distributed cache optimization method according to any one of claims 1 to 7.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the distributed cache optimization method according to any one of claims 1 to 7 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the distributed cache optimization method according to any one of claims 1 to 7 are implemented.

Images