Distributed cache data regulation and control safety guarantee management system and method

Abstract

The invention discloses a distributed cache data regulation and control security guarantee management system and method, and relates to the technical field of cache data security management, and the method comprises the steps: obtaining historical node operation records of cache nodes in a network platform, and evaluating node operation states of the cache nodes under different operation indexes; acquiring a historical data access record in the network platform, and analyzing an access association degree between different feature data items in the network platform to obtain feature association data; acquiring historical data access records in the network platform, constructing a data access prediction model, and predicting the data access condition in the network platform by using the constructed data access prediction model in combination with the feature associated data; and loading data in the network platform into the cache node based on the predicted hotspot access data, analyzing the node operation state of the cache node according to the target node operation data, and regulating and controlling the data in the cache node.

Description

Technical Field

[0001] The present invention relates to the technical field of cache data security management, and in particular to a distributed cache data regulation and control security management system and method. Background Art

[0002] Conventional data access requests directly query the back-end database, which not only leads to a significant increase in response time, but also frequent access to the back-end database can easily cause the back-end database to be overloaded, which in turn causes the database to crash. Therefore, distributed caching technology is widely used. The use of distributed storage technology includes but is limited to the following benefits: 1. Improved performance. Distributed caching technology can greatly reduce data access latency by storing data in memory; 2. Improved system scalability. The load can be dispersed by adding cache nodes, thereby improving the system's processing power; 3. More flexible. The application of distributed caching technology enables the system to support different data types, and can also configure corresponding caching strategies according to actual conditions, so as to better adapt to different business scenarios.

[0003] In conventional data storage, data is usually stored in fixed nodes, but in fact different data is accessed and used to different degrees. This method can easily cause load imbalance and may also cause frequently accessed data to be concentrated in a single cache node, affecting system performance. In addition, the current existing technology is unable to analyze the access association between data, and is also unable to predict data access based on historical data. It is impossible to prepare corresponding control strategies in advance and load hot data from the back-end data source to the distributed cache nodes in advance, which is likely to cause data that is not in the cache server to be frequently accessed, thereby causing the system to crash and further trigger a series of security incidents. Summary of the invention

[0004] The object of the present invention is to provide a distributed cache data control security management system and method to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a distributed cache data regulation and security management method, the method comprising: Step S100: Acquire historical node operation records of cache nodes in the network platform, evaluate the node operation status of the cache nodes under different operation indicators, and obtain target node operation data; Step S200: obtaining historical data access records in the network platform, extracting characteristic data items from the historical data access records, analyzing the access correlation degree between different characteristic data items in the network platform, and obtaining characteristic correlation data; Step S300: Acquire historical data access records in the network platform, construct a data access prediction model, use the constructed data access prediction model, and combine with feature association data to predict the data access status in the network platform to obtain predicted hotspot access data; Step S400: Based on the predicted hotspot access data, the data in the network platform is loaded into the cache node, and according to the target node operation data, the node operation status of the cache node is analyzed to regulate the data in the cache node. Furthermore, step S100 includes: Step S101: Acquire historical node operation records of cache nodes in the network platform, and acquire data corresponding to various operation indicators of the cache nodes from the historical node operation records; Step S102: Obtain the thresholds of various performance indicators in the cache node, set a unit time, and collect the performance indicators of the cache nodes in the historical node operation records every unit time to obtain a data set A of the performance indicators in the historical node operation records = {a 1 、a 2 , ..., a n}, where a 1 、a 2 , ..., a n They are respectively represented as the average values ​​of the performance indicators in the 1st, 2nd, ..., nth unit durations in the historical node operation records, where n is the total number of unit durations contained in the historical node operation records; Step S103: Calculate the data mutation value of the performance indicator in each unit time length in the historical node operation record, where the data mutation value B of the performance indicator in the hth unit time length in the historical node operation record is h =a h -a h-1 / a h-1 , a h is the average value of the performance index within the hth unit time in the historical node operation record, a h-1 It is the average value of the performance index within the h-1th unit time length in the historical node operation record; Step S104: When a h ≥a´∧a h <a´∧B h<B', where a' is the threshold of the performance metric in the cache node, B' is the preset data mutation threshold, the (h - 1)-th unit time duration is denoted as the mutation unit time duration of the historical node operation record, obtain each historical node operation record of the cache node, obtain the operation metric of the cache node, and obtain the maximum value of the average value within the mutation unit time duration in each historical node operation record, and denote it as the target threshold of the operation metric. When it is determined that the operation metric is greater than the target threshold, there is an abnormal risk in the operation of the cache node; Step S105: Obtain and aggregate the target thresholds of the operation metrics of the cache node to obtain the target node operation data of the cache node; In the above steps, the data mutation value of the performance metric in the cache node is calculated because it is used to analyze whether the performance metric suddenly exceeds the threshold within this unit time duration. If the data mutation value is less than the preset data mutation threshold, it is determined that the performance metric does not mutate within the unit time duration. For example, if it is normal within the (h - 1)-th unit time duration and abnormal within the h-th unit time duration, it indicates that the operation metric is very likely to cause the performance of the cache node to decline within the (c - 1)-th unit time duration. By obtaining the operation metric within the (c - 1)-th unit time duration, the target threshold of the operation metric in the cache node can be found, providing strong support for ensuring the normal operation and data security of the cache node in the following text.

[0006] Further, step S200 includes: Step S201: Record the access behavior of the backend server in the network platform within the historical period to obtain the historical data access record. From the historical data access record, obtain the data items stored in the backend server of the network platform that are accessed by users, and denote them as characteristic data items; Step S202: Obtain the total number of times the characteristic data items are accessed from the historical data access record, obtain each historical data access record of the network platform, obtain each characteristic data item of the network platform, and analyze the access correlation degree between each characteristic data item in the network platform. Among them, when analyzing the access correlation degree between the d-th characteristic data item and the e-th characteristic data item in the network platform, the specific process is as follows: Calculate the access correlation value F between the d-th characteristic data item and the e-th characteristic data item d,e : , where j is the total number of all historical data access records; G d,i is the total number of times the d-th characteristic data item is accessed in the i-th historical service record of the network platform; G e,i is the total number of times the e-th characteristic data item is accessed in the i-th historical service record; Step S203: When the access correlation value F d,eis greater than a preset access association threshold, and it is determined that the access between the dth feature data item and the eth feature data item is associated; Step S204: Acquire several feature data items with access relevance in the network platform, and aggregate them to obtain feature-related data.

[0007] Furthermore, step S300 includes: Step S301: Acquire historical data access records in each historical period in the network platform, and construct a data access prediction model in the network platform. The model construction process for the characteristic data items in the network platform is as follows: The total number of times the characteristic data items in the historical data access records within the historical period are accessed is used as the output data of the historical period in the data access prediction model, and the total number of times the characteristic data items are accessed in the historical data access records of several historical periods before the historical period are collected to obtain an input data vector, and the input data vector is used as the input data of the historical period in the data access prediction model; Step S302: dividing the historical data access records in each historical period into a training set and a test set based on a preset ratio; Get the optimal hyperplane in the data access model, and get the normal vector w and bias b of the preset hyperplane. The specific formula of the hyperplane is f(z)=w T z+b, where z is the input data vector of the feature data item in the sample in the training set; The data access prediction model is trained using the training set to obtain a preset access threshold. When the total number of accesses to the feature data item of the sample in the training set is greater than the access threshold, the feature data item is determined to be hot data, and the true label y of the sample is set to +s, otherwise it is set to -s, where s is the preset feature value; Step S303: Obtain the objective function of the data access prediction model, and use the Lagrange multiplier method to optimize the objective function to obtain the Lagrange function L(w, b, α): , Among them, α x is the Lagrange multiplier corresponding to the xth sample in the training set; C is the penalty parameter in the data access prediction model; ξ x is the slack variable of the xth sample; z x is the input data vector of the feature data item in the xth sample; y x is the true label of the xth sample; Using the training set, find the optimal solution of the Lagrangian function L(w,b,α) and get the updated w and b; Step S304: Calculate the data access prediction model and the prediction accuracy P of the feature data item: , Among them, TP is the total number of samples predicted as hotspot data with correct feature data items in the test set; TN is the total number of samples predicted as non-hotspot data with correct feature data items in the test set; FP is the total number of samples predicted as hotspot data with incorrect feature data items in the test set; FN is the total number of samples predicted as non-hotspot data with incorrect feature data items in the test set; Step S305: when the prediction accuracy P is greater than a preset prediction accuracy threshold, it is determined that the data access prediction model has completed construction of the feature data item; Step S306: Use the constructed data access prediction model to obtain feature association data, predict the data access status of each feature data item in the network platform, obtain feature data items predicted to be hot data, and record them as hot feature data items, obtain feature data items with access association with hot feature data items, and also record them as hot feature data items, obtain several hot feature data items in the network platform and aggregate them to obtain predicted hotspot access data.

[0008] Furthermore, step S400 includes: Step S401: obtaining predicted hotspot access data in the network platform, obtaining a number of hotspot feature data items from the predicted hotspot access data, and sequentially loading the several hotspot feature data items into a preset cache node; Step S402: Obtain the target node operation data of the cache node, and obtain the target threshold value of each operation indicator from the target node operation data. When the maximum value of an operation indicator in the cache node is greater than the target threshold value of a certain operation indicator, it is determined that the cache node has a node operation risk, and the hot feature data items cached in the cache node are migrated to other cache nodes, and the data in the cache node is regulated.

[0009] In order to better implement the above method, a distributed cache data control security management system is also proposed, the system includes an operation status evaluation module, a feature association data module, a prediction data module, and an intelligent control module; The operation status evaluation module is used to evaluate the node operation status of the cache node under different operation indicators and obtain the target node operation data; The feature association data module is used to analyze the access association degree between different feature data items in the network platform to obtain feature association data; The prediction data module is used to predict the data access status in the network platform and obtain the predicted hotspot access data; The intelligent control module is used to load the data in the network platform into the cache node according to the predicted hot spot access data, and analyze the node operation status of the cache node according to the target node operation data to control the data in the cache node.

[0010] Further, the operation status evaluation module includes a target threshold unit and a target node operation data unit; A target threshold unit is used to obtain the maximum value of the average value of the operation index of the cache node in the operation records of each historical node, and record it as the target threshold of the operation index; The target node operation data unit is used to obtain and aggregate target thresholds of various operation indicators of the cache node to obtain the target node operation data of the cache node.

[0011] Further, the feature association data module includes an access association value unit and a feature association data unit; An access association value unit, used to calculate the access association value between each characteristic data item in the network platform; The feature association data unit is used to analyze the access association degree between each feature data item in the network platform according to the access association value to obtain feature association data.

[0012] Further, the prediction data module includes a model building unit and a prediction data unit; A model building unit, used to obtain historical data access records in each historical period in the network platform and build a data access prediction model in the network platform; The prediction data unit is used to use the constructed data access prediction model and combine it with the feature association data to predict the data access status in the network platform and obtain the predicted hotspot access data.

[0013] Further, the intelligent control module includes an intelligent control unit; The intelligent control unit is used to perform risk assessment on the cache node. When a cache node is judged to have node operation risk, the hot feature data items cached in the cache node are migrated to other cache nodes, and the data in the cache node is regulated.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention realizes the intelligent regulation of cache data by utilizing distributed cache technology. In order to ensure that data is accessed safely and quickly, the node operation status of the cache node and the prediction of hot data are considered. Through historical records, the hot data in the network platform within the current cycle is predicted, and the data predicted as hot spots is loaded into the cache node in advance, the cache is filled in advance, and the number of cache misses during business peak periods is reduced. Moreover, through the analyzed target node operation data, the operation status of the cache node is evaluated in real time, the data in the cache node is migrated, and the intelligent regulation of the data in the cache node is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a method flow chart of a distributed cache data regulation and security management method of the present invention; Figure 2 It is a module schematic diagram of a distributed cache data regulation and security management system of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a distributed cache data regulation and security management method, the method comprising: Step S100: Acquire historical node operation records of cache nodes in the network platform, evaluate the node operation status of the cache nodes under different operation indicators, and obtain target node operation data; Wherein, step S100 includes: Step S101: Acquire historical node operation records of cache nodes in the network platform, and acquire data corresponding to various operation indicators of the cache nodes from the historical node operation records; For example, various operating indicators include memory usage, CPU usage, etc. Step S102: Obtain the thresholds of various performance indicators in the cache node, set a unit time, and collect the performance indicators of the cache nodes in the historical node operation records every unit time to obtain a data set A of the performance indicators in the historical node operation records = {a 1 、a 2 , ..., a n}, where a 1 、a 2 、...、a n respectively represent the average values of the performance metrics in the 1st, 2nd,..., nth unit time periods in the historical node operation records, and n is the total number of unit time periods included in the historical node operation records; For example, the performance metrics include cache hit rate, latency, etc.; Step S103: Calculate the data mutation value of the performance metric in each unit time period in the historical node operation records. Among them, the data mutation value B h =a h -a h-1 / a h-1 , a h is the average value of the performance metric in the hth unit time period in the historical node operation records, and a h-1 is the average value of the performance metric in the (h - 1)th unit time period in the historical node operation records; For example, the average value a 3 of the performance metric in the 3rd unit time period in the historical node operation records is 30, and the average value a 2 of the performance metric in the 2nd unit time period in the historical node operation records is 20. Calculate the data mutation value B 3 =30 - 20 / 20 = 0.5; Step S104: When a h ≥a´ ∧ a h <a´ ∧ B h <B´, where a´ is the threshold of the performance metric in the cache node, and B´ is the preset data mutation threshold. Denote the (h - 1)th unit time period as the mutation unit time period of the historical node operation records. Obtain the various historical node operation records of the cache node, obtain the operation metrics of the cache node, and denote the maximum value of the average values in the mutation unit time periods in the various historical node operation records as the target threshold of the operation metrics. When it is determined that the operation metric is greater than the target threshold, there is an abnormal risk in the operation of the cache node; Step S105: Obtain and collect the target thresholds of the various operation metrics of the cache node to obtain the target node operation data of the cache node; Step S200: Obtain the historical data access records in the network platform, extract the feature data items from the historical data access records, and analyze the access association degree between different feature data items in the network platform to obtain the feature association data; Among them, Step S200 includes: Step S201: Record the access behavior of the backend server in the network platform in the historical period to obtain the historical data access record, and obtain the data items accessed by the user stored in the backend server of the network platform from the historical data access record, and record them as feature data items; Step S202: Obtain the total number of times the characteristic data item is accessed from the historical data access record, obtain each historical data access record of the network platform, obtain each characteristic data item of the network platform, and analyze the access correlation degree between each characteristic data item in the network platform, wherein the access correlation degree between the d-th characteristic data item and the e-th characteristic data item in the network platform is analyzed. The specific process is as follows: Calculate the access association value F between the dth feature data item and the eth feature data item d,e : , Where, j is the total number of historical data access records; G d,i is the total number of times the dth feature data item is accessed in the i-th historical service record of the network platform; G e,i is the total number of times the e-th feature data item is accessed in the i-th historical service record; Step S203: When accessing the associated value F d,e is greater than a preset access association threshold, and it is determined that the access between the dth feature data item and the eth feature data item is associated; Step S204: obtaining a plurality of characteristic data items with access relevance in the network platform, and aggregating them to obtain characteristic relevance data; Step S300: Acquire historical data access records in the network platform, construct a data access prediction model, use the constructed data access prediction model, and combine with feature association data to predict the data access status in the network platform to obtain predicted hotspot access data; Wherein, step S300 includes: Step S301: Acquire historical data access records in each historical period in the network platform, and construct a data access prediction model in the network platform. The model construction process for the characteristic data items in the network platform is as follows: The total number of times the characteristic data items in the historical data access records within the historical period are accessed is used as the output data of the historical period in the data access prediction model, and the total number of times the characteristic data items are accessed in the historical data access records of several historical periods before the historical period are collected to obtain an input data vector, and the input data vector is used as the input data of the historical period in the data access prediction model; Step S302: dividing the historical data access records in each historical period into a training set and a test set based on a preset ratio; Get the optimal hyperplane in the data access model, and get the normal vector w and bias b of the preset hyperplane. The specific formula of the hyperplane is f(z)=w T z+b, where z is the input data vector of the feature data item in the sample in the training set; The data access prediction model is trained using the training set to obtain a preset access threshold. When the total number of accesses to the feature data item of the sample in the training set is greater than the access threshold, the feature data item is determined to be hot data, and the true label y of the sample is set to +s, otherwise it is set to -s, where s is the preset feature value. Step S303: Obtain the objective function of the data access prediction model, and use the Lagrange multiplier method to optimize the objective function to obtain the Lagrangian function L(w, b, α): , Among them, α x is the Lagrange multiplier corresponding to the xth sample in the training set; C is the penalty parameter in the data access prediction model; ξ x is the slack variable of the xth sample; z x is the input data vector of the feature data item in the xth sample; y x is the true label of the xth sample; Using the training set, find the optimal solution of the Lagrangian function L(w,b,α) and get the updated w and b; Step S304: Calculate the data access prediction model and the prediction accuracy P of the feature data item: , Among them, TP is the total number of samples predicted as hotspot data with correct feature data items in the test set; TN is the total number of samples predicted as non-hotspot data with correct feature data items in the test set; FP is the total number of samples predicted as hotspot data with incorrect feature data items in the test set; FN is the total number of samples predicted as non-hotspot data with incorrect feature data items in the test set; For example, TP is 100; TN is 100; FP is 10; FN is 20; calculate the data access prediction model, the prediction accuracy P of the feature data item: , Step S305: when the prediction accuracy P is greater than a preset prediction accuracy threshold, it is determined that the data access prediction model has completed construction of the feature data item; Step S306: using the constructed data access prediction model, obtaining feature association data, predicting the data access status of each feature data item in the network platform, obtaining feature data items predicted to be hot data, and recording them as hot feature data items, obtaining feature data items with access association with the hot feature data items, and also recording them as hot feature data items, obtaining several hot feature data items in the network platform and aggregating them to obtain predicted hot access data; Step S400: Based on the predicted hotspot access data, the data in the network platform is loaded into the cache node, and according to the target node operation data, the node operation status of the cache node is analyzed to regulate the data in the cache node; Wherein, step S400 includes: Step S401: obtaining predicted hotspot access data in the network platform, obtaining a number of hotspot feature data items from the predicted hotspot access data, and sequentially loading the several hotspot feature data items into a preset cache node; Step S402: Obtain target node operation data of the cache node, obtain target thresholds of various operation indicators from the target node operation data, and when the maximum value of a certain operation indicator in the cache node is greater than the target threshold of a certain operation indicator, determine that the cache node has a node operation risk, migrate the hotspot feature data items cached in the cache node to other cache nodes, and regulate the data in the cache node; In order to better implement the above method, a distributed cache data control security management system is also proposed, the system includes an operation status evaluation module, a feature association data module, a prediction data module, and an intelligent control module; The operation status evaluation module is used to evaluate the node operation status of the cache node under different operation indicators and obtain the target node operation data; The feature association data module is used to analyze the access association degree between different feature data items in the network platform to obtain feature association data; The prediction data module is used to predict the data access status in the network platform and obtain the predicted hotspot access data; The intelligent control module is used to load the data in the network platform into the cache node according to the predicted hotspot access data, and analyze the node operation status of the cache node according to the target node operation data, and control the data in the cache node; Among them, the operation status evaluation module includes a target threshold unit and a target node operation data unit; A target threshold unit is used to obtain the maximum value of the average value of the operation index of the cache node in the operation records of each historical node, and record it as the target threshold of the operation index; The target node operation data unit is used to obtain and aggregate the target thresholds of various operation indicators of the cache node to obtain the target node operation data of the cache node; Among them, the feature association data module includes an access association value unit and a feature association data unit; An access association value unit, used to calculate the access association value between each characteristic data item in the network platform; The feature association data unit is used to analyze the access association degree between each feature data item in the network platform according to the access association value to obtain feature association data; Among them, the prediction data module includes a model building unit and a prediction data unit; A model building unit, used to obtain historical data access records in each historical period in the network platform and build a data access prediction model in the network platform; The prediction data unit is used to use the constructed data access prediction model and combine the feature association data to predict the data access status in the network platform and obtain the predicted hotspot access data; Wherein, the intelligent control module includes an intelligent control unit; The intelligent control unit is used to perform risk assessment on the cache node. When a cache node is judged to have node operation risk, the hot feature data items cached in the cache node are migrated to other cache nodes, and the data in the cache node is regulated.

[0018] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A distributed cache data control security management method, characterized in that: The method comprises: Step S100: Acquire historical node operation records of cache nodes in the network platform, evaluate the node operation status of the cache nodes under different operation indicators, and obtain target node operation data; Step S200: obtaining historical data access records in the network platform, extracting characteristic data items from the historical data access records, analyzing the access correlation degree between different characteristic data items in the network platform, and obtaining characteristic correlation data; Step S300: acquiring historical data access records in the network platform, constructing a data access prediction model, using the constructed data access prediction model and combining the feature association data to predict the data access status in the network platform, and obtaining predicted hotspot access data; Step S400: Based on the predicted hotspot access data, the data in the network platform is loaded into the cache node, and according to the target node operation data, the node operation status of the cache node is analyzed to regulate the data in the cache node.

2. A distributed cache data control security management method according to claim 1, characterized in that: The step S100 includes: Step S101: acquiring historical node operation records of cache nodes in the network platform, and acquiring data corresponding to various operation indicators of the cache nodes from the historical node operation records; Step S102: Obtain the thresholds of various performance indicators in the cache node, set a unit time, and collect the performance indicators of the cache node in the historical node operation record every unit time to obtain a data set A of the performance indicators in the historical node operation record = {a1, a2, ..., a n }, where a1, a2, ..., a n They are respectively represented as the average values ​​of the performance indicators in the 1st, 2nd, ..., nth unit durations in the historical node operation records, where n is the total number of unit durations contained in the historical node operation records; Step S103: Calculate the data mutation value of the performance indicator in each unit time length in the historical node operation record, wherein the data mutation value B of the performance indicator in the hth unit time length in the historical node operation record is h =a h -a h-1 / a h-1 , a h is the average value of the performance indicator within the hth unit time length in the historical node operation record, a h-1 is the average value of the performance indicator within the h-1th unit time length in the historical node operation record; Step S104: When a h ≥ a' ∧ a h < a' ∧ B h < B', where a' is the threshold of the performance metric in the cache node, B' is the preset data mutation threshold, the (h - 1)-th unit time duration is recorded as the mutation unit time duration of the historical node operation record, each historical node operation record of the cache node is obtained, the operation metric of the cache node is obtained, the maximum value of the average value within the mutation unit time duration in each historical node operation record is recorded as the target threshold of the operation metric, and when it is determined that the operation metric is greater than the target threshold, there is an abnormal risk in the operation of the cache node; Step S105: acquiring and collecting target thresholds of various operating indicators of the cache node to obtain target node operating data of the cache node.

3. A distributed cache data control security management method according to claim 2, characterized in that: The step S200 includes: Step S201: Recording the access behavior of the backend server in the network platform in a historical period to obtain a historical data access record, and obtaining data items accessed by users stored in the backend server of the network platform from the historical data access record, and recording them as feature data items; Step S202: obtaining the total number of times the characteristic data item is accessed from the historical data access record, obtaining each historical data access record of the network platform, obtaining each characteristic data item of the network platform, and analyzing the access correlation degree between the each characteristic data item in the network platform, wherein the access correlation degree between the d-th characteristic data item and the e-th characteristic data item in the network platform is analyzed, and the specific process is as follows: Calculate the access association value F between the dth feature data item and the eth feature data item d,e : ， Wherein, j is the total number of the historical data access records; G d,i is the total number of times the dth feature data item is accessed in the i-th historical service record of the network platform; G e,i is the total number of times the e-th feature data item is accessed in the i-th historical service record; Step S203: When the access associated value F d,e is greater than a preset access association threshold, determining that the access between the d-th characteristic data item and the e-th characteristic data item is associated; Step S204: Acquire a plurality of feature data items with access relevance in the network platform, and aggregate them to obtain feature-related data.

4. A distributed cache data control security management method according to claim 3, characterized in that: The step S300 includes: Step S301: Acquire historical data access records in each historical period in the network platform, and construct a data access prediction model in the network platform, wherein the model construction process for the characteristic data items in the network platform is: The total number of times the characteristic data items in the historical data access records within the historical period are accessed is used as the output data of the historical period in the data access prediction model, and the total number of times the characteristic data items are accessed in the historical data access records of several historical periods before the historical period are collected to obtain an input data vector, and the input data vector is used as the input data of the historical period in the data access prediction model; Step S302: dividing the historical data access records in each historical period into a training set and a test set based on a preset ratio; Obtain the optimal hyperplane in the data access model, obtain the preset normal vector w and bias b of the hyperplane, where the specific formula of the hyperplane is f(z)=w T z+b, where z is the input data vector of the feature data item in the sample in the training set; The data access prediction model is trained using a training set to obtain a preset access threshold. When the total number of accesses to the feature data item of the sample in the training set is greater than the access threshold, the feature data item is determined to be hot data, and the true label y of the sample is set to +s, otherwise it is set to -s, where s is a preset feature value; Step S303: Obtain the objective function of the data access prediction model, and use the Lagrangian multiplier method to optimize the objective function to obtain the Lagrangian function L(w, b, α): ， Among them, α x is the Lagrange multiplier corresponding to the xth sample in the training set; C is the penalty parameter in the data access prediction model; ξ x is the slack variable of the x-th sample; z x is the input data vector of the feature data item in the x-th sample; x is the true label of the x-th sample; Using the training set, solving the optimal solution of the Lagrangian function L(w, b, α) to obtain updated w and b; Step S304: Calculate the prediction accuracy P of the data access prediction model for the feature data item: ， Among them, TP is the total number of samples predicted as hotspot data when the feature data items in the test set are correct; TN is the total number of samples predicted as non-hotspot data when the feature data items in the test set are correct; FP is the total number of samples predicted as hotspot data when the feature data items in the test set are incorrect; FN is the total number of samples predicted as non-hotspot data when the feature data items in the test set are incorrect. Step S305: when the prediction accuracy rate P is greater than a preset prediction accuracy threshold, it is determined that the data access prediction model has completed construction of the feature data item; Step S306: Use the constructed data access prediction model to obtain the feature association data, predict the data access status of each feature data item in the network platform, obtain feature data items predicted to be hot data, and record them as hot feature data items, obtain feature data items with access association with the hot feature data items, and also record them as hot feature data items, obtain several hot feature data items in the network platform and aggregate them to obtain predicted hotspot access data.

5. A distributed cache data control security management method according to claim 4, characterized in that: The step S400 includes: Step S401: obtaining predicted hotspot access data in the network platform, obtaining a plurality of hotspot feature data items from the predicted hotspot access data, and sequentially loading the plurality of hotspot feature data items into a preset cache node; Step S402: Obtain the target node operation data of the cache node, and obtain the target thresholds of the various operation indicators from the target node operation data. When the maximum value of an operation indicator in the cache node is greater than the target threshold of the operation indicator, it is determined that the cache node has a node operation risk, and the hot feature data items cached in the cache node are migrated to other cache nodes, and the data in the cache node is regulated.

6. A distributed cache data regulation and security management system, used to execute a distributed cache data regulation and security management method according to any one of claims 1 to 5, characterized in that: The system includes an operation status evaluation module, a feature association data module, a prediction data module, and an intelligent control module; The operation status evaluation module is used to evaluate the node operation status of the cache node under different operation indicators to obtain target node operation data; The feature association data module is used to analyze the access association degree between different feature data items in the network platform to obtain feature association data; The prediction data module is used to predict the data access status in the network platform to obtain predicted hotspot access data; The intelligent control module is used to load the data in the network platform into the cache node according to the predicted hotspot access data, and analyze the node operation status of the cache node according to the target node operation data to regulate the data in the cache node.

7. A distributed cache data control and security management system according to claim 6, characterized in that: The operation status evaluation module includes a target threshold unit and a target node operation data unit; The target threshold unit is used to obtain the maximum value of the average values ​​of the operation index of the cache node in the operation records of each historical node, and record it as the target threshold value of the operation index; The target node operation data unit is used to acquire and collect target thresholds of various operation indicators of the cache node to obtain the target node operation data of the cache node.

8. A distributed cache data control and security management system according to claim 6, characterized in that: The feature association data module includes an access association value unit and a feature association data unit; The access association value unit is used to calculate the access association value between each characteristic data item in the network platform; The feature association data unit is used to analyze the access association degree between the various feature data items in the network platform according to the access association value to obtain feature association data.

9. A distributed cache data control and security management system according to claim 6, characterized in that: The prediction data module includes a model building unit and a prediction data unit; The model building unit is used to obtain historical data access records in each historical period in the network platform and build a data access prediction model in the network platform; The prediction data unit is used to use the constructed data access prediction model and combine it with the feature association data to predict the data access status in the network platform and obtain predicted hotspot access data.

10. A distributed cache data control and security management system according to claim 6, characterized in that: The intelligent control module includes an intelligent control unit; The intelligent control unit is used to perform risk assessment on the cache node. When the cache node is determined to have node operation risk, the hot spot feature data items cached in the cache node are migrated to other cache nodes to regulate the data in the cache node.

Images