A method and system for intelligent edge dynamic security optimization

Abstract

The present application discloses a method and system for dynamic security optimization of intelligent edge, which relates to the technical field of data security. The method includes: obtaining the use environment of edge computing equipment, evaluating and obtaining the probability of data leakage; obtaining the data accessed by the edge computing equipment, analyzing and obtaining the data necessity; extracting the data features of the data, analyzing and obtaining the data sensitivity; combining the probability of data leakage, the data necessity and the data sensitivity to comprehensively analyze and obtain the data security; obtaining all security policies, evaluating and obtaining the policy security and policy convenience; according to the policy security and policy convenience, combined with the data security, analyzing and obtaining the policy matching, selecting the security policy with the highest policy matching to run; using Docker container technology to deploy lightweight AI models, train AI models, and realize dynamic security optimization of intelligent edge through AI models. The present application improves the security of dynamic security optimization of intelligent edge.

Description

Technical Field

[0001] The present application relates to the technical field of data security, and in particular to a method and system for dynamic security optimization of an intelligent edge. Background Art

[0002] Intelligent edge computing is a distributed computing architecture that migrates computing and data storage to the edge of the network, that is, computing and storing near the device or terminal, rather than transmitting data to the cloud or central server for processing. This computing model can greatly improve the speed and efficiency of data processing, reduce transmission costs, and reduce network congestion, and achieve real-time feedback and decision-making. In the prior art, dynamic security is usually achieved through login verification, face recognition, etc., but these security policy settings are very simple and are activated under specific operation actions. A single fixed security policy setting is difficult to cover multiple scenarios, resulting in low security of edge computing. Summary of the invention

[0003] The purpose of the present invention is to provide an intelligent edge dynamic security optimization method and system to solve the problems raised in the above background technology.

[0004] In the first aspect, the present application provides an intelligent edge dynamic security optimization method, which adopts the following technical solutions:

[0005] Obtain the usage environment of the edge computing device, evaluate the probability of data leakage based on the usage environment and record it as the data leakage probability;

[0006] Obtain the data accessed by the edge computing device, and analyze the necessity of accessing the data based on the usage environment to obtain the data necessity;

[0007] Extract data features of the data, and analyze the sensitivity of the data according to the data features to obtain data sensitivity;

[0008] Combine the probability of data leakage, data necessity and data sensitivity to comprehensively analyze the security of the data and obtain the data security level;

[0009] Obtain all security policies, evaluate the security and convenience of the security policies, and obtain the policy security and policy convenience;

[0010] According to the policy security and convenience, combined with data security, the policy matching is analyzed and the security policy with the highest policy matching is selected for operation;

[0011] Use Docker container technology to deploy lightweight AI models, train AI models according to the selection method of security policy operation, and realize dynamic security optimization of intelligent edge through AI models.

[0012] Preferably, the step of obtaining the use environment of the edge computing device, evaluating the probability of data leakage according to the use environment and recording it as the data leakage probability is specifically:

[0013] Obtaining a usage environment of the edge computing device, wherein the usage environment includes a usage location and device usage information;

[0014] Obtain historical location data of edge computing devices, and extract common locations based on the historical location data;

[0015] Determine whether the usage location is a common location. If the usage location is not a common location, obtain the usage habits of the user holding the edge computing device, analyze the usage habits, and obtain the data leakage probability based on the usage habits and record it as the first leakage probability;

[0016] If the usage location is a common location, obtaining a preset data leakage probability as a first leakage probability;

[0017] A second leakage probability is obtained according to the device usage information analysis, and the average of the first leakage probability and the second leakage probability is calculated to obtain the data leakage probability.

[0018] Preferably, the step of determining whether the usage location is a common location, and if the usage location is an uncommon location, obtaining the usage habits of the user holding the edge computing device, and obtaining the data leakage probability according to the usage habits and recording it as the first leakage probability, is specifically as follows:

[0019] Get the average crowd density SM of the usage location;

[0020] The usage habits include the usage height SG, the font size SZ, and the usage duration SS of the edge computing device;

[0021] According to the first leakage correlation function The first leakage probability SY is calculated, where , is the scaling factor and is greater than 0.

[0022] Preferably, the step of obtaining the second leakage probability according to the device usage information analysis is specifically:

[0023] The device usage information includes physical usage information and network usage information;

[0024] The physical usage information includes real-time temperature AW, real-time humidity AS and electromagnetic intensity AD;

[0025] Obtain the standard temperature BW and standard humidity BS used by edge computing devices;

[0026] According to the damage correlation function The equipment damage probability AL is calculated, where , , is the scale factor and is greater than 0;

[0027] The network security degree AA of the network used by the edge computing device is obtained based on the network usage information;

[0028] According to the second leakage correlation function The second leakage probability BE is calculated, where , is the scaling factor and is greater than 0.

[0029] Preferably, the step of obtaining the data accessed by the edge computing device and analyzing the necessity of accessing the data according to the usage environment to obtain the data necessity degree is specifically as follows:

[0030] Obtain the data accessed by the edge computing device and record it as access data, and extract the user privacy data in the access data;

[0031] Find the usage business associated with the user's privacy data and form a standard business library;

[0032] Find all the businesses that have been handled using location history and form a real-time business database;

[0033] The intersection of the business libraries is formed based on the standard business library and the real-time business library. If the intersection of the business libraries is 0, it is determined that the data is not needed, and the data necessity DB is 0;

[0034] If the intersection of the business libraries is not 0, the number of businesses DY in the intersection of the business libraries is counted, and the data necessity DB is obtained according to the number of businesses.

[0035] Preferably, if the intersection of the business libraries is not 0, the step of counting the number of businesses DY in the intersection of the business libraries and obtaining the data necessity DB according to the number of businesses is specifically as follows:

[0036] Record the businesses in the intersection of the business database as pending businesses, and count the ratio of the number of pending businesses among all businesses handled using the location history. , where i is the number of different pending businesses;

[0037] Obtain the historical processing data of pending business and extract the frequency ratio of access data used in the historical processing data , where i is the number of different pending businesses;

[0038] According to the necessary degree of data correlation function The data necessity DB is calculated, where i is the number of different pending businesses, and n is the maximum number of different pending businesses. , is the scaling factor and is greater than 0.

[0039] Preferably, the step of extracting data features and analyzing the sensitivity of the data according to the data features to obtain the data sensitivity is specifically:

[0040] Extract the data features of the access data and determine whether the data features are unique to the holding user. If the data features are unique to the holding user, determine that they are privacy features.

[0041] If the data feature is not unique to the user holding it, the data feature is judged to be a non-privacy feature;

[0042] If the data feature is a privacy feature, the risk level of privacy feature leakage is obtained and recorded as data sensitivity;

[0043] If the data feature is a non-privacy feature, the probability of obtaining the privacy feature through the non-privacy feature is determined and recorded as the privacy probability;

[0044] The risk level of privacy feature leakage corresponding to the privacy probability is obtained, and the data sensitivity is obtained by multiplying it by the privacy probability.

[0045] Preferably, the steps of obtaining all security policies, evaluating the security and convenience of the security policies, and obtaining the policy security and policy convenience are specifically as follows:

[0046] Obtain the data encryption complexity FZ, response speed FX and attack success rate FL of the security policy;

[0047] According to the policy safety correlation function The strategy security degree FA is calculated, where , , is the scale factor and is greater than 0;

[0048] Determine whether the security policy is an automatic operation. If the security policy is an automatic operation, obtain the occupied resources and data processing efficiency of the security policy, set the automatic weight ratio corresponding to the occupied resources and data processing efficiency, and calculate the policy convenience based on the automatic weight ratio;

[0049] If the security policy is a manual operation, the manual operation time and the number of manual operation steps are obtained, a manual weight ratio corresponding to the manual operation time and the number of manual operation steps is set, and the policy convenience is calculated according to the manual weight ratio.

[0050] Preferably, the step of analyzing the policy matching degree according to the policy security and policy convenience degree in combination with the data security degree, and selecting the security policy with the highest policy matching degree to run is specifically as follows:

[0051] Match the corresponding policy security level according to the data security level and record it as the standard security level;

[0052] Screen the security policies whose policy safety reaches the standard safety and record them as candidate security policies;

[0053] Set the policy weight ratio of the policy security and policy convenience of the security policy to be selected, and calculate the policy matching degree of the security policy to be selected according to the policy security and policy convenience and the corresponding policy weight ratio;

[0054] The candidate security policy with the highest policy matching degree is screened and run.

[0055] In the second aspect, the present application provides an intelligent edge dynamic security optimization system, which adopts the following technical solutions:

[0056] An intelligent edge dynamic security optimization system, comprising:

[0057] The leakage probability module obtains the usage environment of the edge computing device, evaluates the probability of data leakage based on the usage environment and records it as the data leakage probability;

[0058] The data necessity module obtains the data accessed by the edge computing device and analyzes the necessity of accessing the data according to the usage environment to obtain the data necessity degree;

[0059] The data sensitivity module extracts the data features of the data and analyzes the sensitivity of the data according to the data features to obtain the data sensitivity;

[0060] The data security module comprehensively analyzes the security level of data and obtains the data security level by combining the probability of data leakage, data necessity and data sensitivity;

[0061] The security policy module obtains all security policies, evaluates the security and convenience of the security policies, and obtains the policy security and policy convenience;

[0062] The policy matching module analyzes the policy matching degree based on the policy security and policy convenience, combined with the data security, and selects the security policy with the highest policy matching degree to run;

[0063] The security optimization module uses Docker container technology to deploy lightweight AI models, trains AI models according to the selection method of security policy operation, and realizes dynamic security optimization of intelligent edge through AI models.

[0064] In summary, the present application includes at least one of the following beneficial technical effects:

[0065] 1. Evaluate the probability of data leakage based on the use environment of the edge computing device, confirm the necessity of data access, and combine the data sensitivity of the accessed data to obtain data security. According to the data security, select a security policy that meets the requirements and is safe and convenient to run. At the same time, through AI training, automatic and high-speed intelligent edge dynamic security optimization is achieved, so that edge computing devices can better cope with more scenarios, and give the best security strategy in line with the use scenario, which improves the security of intelligent edge dynamic security optimization.

[0066] 2. Use location to distinguish between common locations and uncommon locations. For uncommon locations, obtain the first leakage probability based on the average crowd density and user usage habits. Obtain the second leakage probability of data based on device usage information including physical usage information and network usage information. Combine the first leakage probability and the second leakage probability to obtain the data leakage probability. The data leakage probability is conducive to selecting more appropriate and accurate security strategies and improves the accuracy of dynamic security optimization of intelligent edge.

[0067] 3. Evaluate the security level of the security policy based on its data encryption complexity, response speed, and attack success rate, and obtain the convenience of the policy based on whether it is manual or automatic. Combined with data security, select a convenient security policy that meets security requirements for operation. Without affecting data security, select a more convenient security policy, which reduces the inconvenience caused to users by the operation of the security policy and improves the convenience of dynamic security optimization of the intelligent edge. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 It is a schematic diagram of the specific steps of an embodiment of an intelligent edge dynamic security optimization method of the present invention.

[0069] Figure 2 It is a module connection diagram of an embodiment of an intelligent edge dynamic security optimization system of the present invention. DETAILED DESCRIPTION

[0070] Below is a combination of the embodiments and Figure 1-Figure 2 The present invention is further described in detail, but the embodiments of the present invention are not limited thereto.

[0071] The present invention discloses a smart edge dynamic security optimization method, which specifically includes the following steps:

[0072] Step S1, obtain the use environment of the edge computing device, evaluate the probability of data leakage according to the use environment and record it as the data leakage probability.

[0073] Step S2, obtaining the data accessed by the edge computing device, and analyzing the necessity of accessing the data according to the usage environment to obtain the data necessity.

[0074] Step S3, extracting data features of the data, and analyzing the sensitivity of the data according to the data features to obtain data sensitivity.

[0075] Step S4, comprehensively analyzing the security level of the data by combining the data leakage probability, data necessity and data sensitivity to obtain the data security level.

[0076] Set the weight ratio corresponding to the data leakage probability, data necessity and data sensitivity for comprehensive analysis of the data, and calculate the data security based on the weight ratio. The greater the probability of data leakage, the less secure the data is, because the data is in a dangerous environment. The lower the data necessity, the less necessary it is to access. At this time, the device requires access, and there is a possibility of data leakage. The higher the data sensitivity, the more serious the consequences of data leakage.

[0077] Step S5, obtaining all security policies, evaluating the security degree and convenience degree of the security policies, and obtaining the policy security degree and policy convenience degree.

[0078] Step S6, according to the policy security and policy convenience, combined with the data security, the policy matching degree is analyzed and the security policy with the highest policy matching degree is selected for execution.

[0079] Step S7: Use Docker container technology to deploy a lightweight AI model, train the AI ​​model according to the selection method of security policy operation, and realize dynamic security optimization of the intelligent edge through the AI ​​model.

[0080] Using Docker container technology to deploy lightweight AI models means packaging the lightweight AI model and its required operating environment, dependencies, etc. into an independent Docker image, and deploying and running it through the Docker container. This approach can ensure that the AI ​​model has consistent operating performance in different environments, improve the portability and deployment efficiency of the model. The establishment of the AI ​​model can obtain results more quickly after training, and realize the dynamic security of edge computing more quickly.

[0081] In actual application, intelligent edge refers to the technology that moves tasks such as computing and data storage from central servers to edge devices. These devices usually include IoT devices such as smartphones, tablets, routers, cameras, and smart TVs. Edge computing devices may also pose threats to data during use due to external malicious attacks, improper user operations, and other reasons. By selecting appropriate security policies based on the security level of the data, the defense capabilities of edge computing devices can be enhanced and data security can be improved. According to different usage environments, the necessity and sensitivity of data, the corresponding security policies are matched, and a more convenient security policy is selected without affecting the protection of edge computing security. Selecting the corresponding security policy according to the actual usage scenario is more flexible, adapts to the user usage scenario, and improves data security.

[0082] The steps of obtaining the use environment of the edge computing device, evaluating the probability of data leakage based on the use environment and recording it as the data leakage probability are as follows:

[0083] Step S11, obtaining the usage environment of the edge computing device, where the usage environment includes the usage location and device usage information.

[0084] Step S12, obtaining historical location data of the edge computing device, and extracting commonly used locations based on the historical location data.

[0085] Step S13, determine whether the usage location is a common location. If the usage location is not a common location, obtain the usage habits of the user holding the edge computing device, analyze the usage habits, and obtain the data leakage probability based on the usage habits and record it as the first leakage probability.

[0086] Step S14: if the usage location is a common location, obtaining a preset data leakage probability as a first leakage probability.

[0087] Step S15, obtaining a second leakage probability based on the device usage information analysis, calculating the average of the first leakage probability and the second leakage probability, and obtaining a data leakage probability.

[0088] In actual use, many security policies are currently triggered in a single fixed scenario and cannot adapt well to multiple scenarios. For example, all logins require passwords or verification codes for verification, but after logging into the page, all information in the software platform can be obtained. In this way, when the edge computing device is stolen by others, it is easy for others to steal information. Different security policies are adopted for different usage locations of edge computing devices to improve data security. For example, when users use edge computing devices at home or in the company, which are common locations, the probability of edge computing devices being stolen by others is low. Therefore, the probability of users personally holding devices is higher, and the probability of data leakage is relatively low. The average data leakage probability of this location is used as the preset data to obtain the first leakage probability. The average data leakage probability of the company will be greater than the average data leakage probability at home. When the usage location is not a common location, there are many hidden dangers, and the probability of data leakage needs to be further analyzed.

[0089] Determine whether the usage location is a common location. If the usage location is not a common location, obtain the usage habits of the user holding the edge computing device, analyze the usage habits to obtain the data leakage probability and record it as the first leakage probability. Specifically, the steps are:

[0090] Step S131, obtaining the average crowd density SM of the usage location.

[0091] Step S132, usage habits include the usage height SG, the font size SZ, and the usage time SS of the edge computing device.

[0092] Step S133: according to the first leakage correlation function The first leakage probability SY is calculated, where , is the scaling factor and is greater than 0.

[0093] In actual use, in uncommon locations, data leakage is more likely to occur due to the distribution of strangers around. For example, the height of the phone can be obtained using the built-in sensor of the phone. The lower the height of the phone, the easier it is for the people around to see the content in the phone. The larger the font used and the longer the usage time, the higher the probability of data being stolen by the people around. At the same time, the density of the surrounding crowd will also increase the risk of data being stolen, thereby increasing the probability of data leakage. At the same time, the high density of people in the use location will also increase the risk of the phone being stolen by others and increase the probability of data leakage.

[0094] The steps of obtaining the second leakage probability according to the device usage information analysis are specifically:

[0095] Step S151, the device usage information includes physical usage information and network usage information.

[0096] Step S152: Physical usage information includes real-time temperature AW, real-time humidity AS and electromagnetic intensity AD.

[0097] Step S153, obtaining the standard temperature BW and standard humidity BS used by the edge computing device.

[0098] Step S154: according to the damage correlation function The equipment damage probability AL is calculated, where , , is the scaling factor and is greater than 0.

[0099] Step S155: extract the network security level AA of the network used by the edge computing device based on the network usage information.

[0100] Step S156: according to the second leakage correlation function The second leakage probability BE is calculated, where , is the scaling factor and is greater than 0.

[0101] In actual use, edge computing devices are often affected by the environment, which can cause changes in their performance and thus affect the probability of data leakage. For example, when the temperature is too high or too low, the mobile phone will freeze or freeze. At this time, the security policy cannot be activated and the mobile phone protection is weak. If it is attacked at this time, it is easy to cause data leakage. Similarly, humidity and electromagnetic interference will also have an impact on edge computing devices. Edge computing devices usually need to interact with the cloud or other devices through the network for data. The location of the device will affect the stability and security of its network connection. In an unstable network environment or vulnerable network area, the data transmission of edge computing devices may be more easily intercepted, tampered with, or stolen. When users use public networks, the network is insecure and easily attacked due to reasons such as too many users, resulting in data leakage.

[0102] The steps to obtain the data accessed by the edge computing device and analyze the necessity of accessing the data according to the usage environment to obtain the data necessity are as follows:

[0103] Step S21, obtain the data accessed by the edge computing device and record it as access data, and extract the user privacy data in the access data.

[0104] Step S22, searching for usage services associated with the user's privacy data and forming a standard service library.

[0105] Step S23, searching for all services handled using the location history and forming a real-time service database.

[0106] Step S24: forming a business library intersection according to the standard business library and the real-time business library. If the business library intersection is 0, it is determined that the data is not needed and the data necessity DB is 0.

[0107] Step S25: If the intersection of the business libraries is not 0, the number of businesses DY in the intersection of the business libraries is counted, and the data necessity DB is obtained according to the number of businesses.

[0108] In actual use, the necessity of edge computing devices to access data is used to determine whether the platform or others will maliciously obtain data. For example, a user's medical record information can be used for services such as seeing a doctor and buying medicine, but the current user is in a clothing store, and the clothing store's services include buying clothes and returning clothes. However, the services that can be used for medical records are not related to the services handled by the clothing store, and the intersection of the business database is 0. The more business intersections there are, the more necessary it is to retrieve data. The necessity of data can be used to determine whether data needs to be retrieved. The lower the necessity of data, the higher the probability of being used by others. For example, when a user accesses bank card information in a milk tea shop, it is likely that someone else is accessing it maliciously, and a higher level of verification is required at this time.

[0109] Step S251, if the intersection of the business libraries is not 0, then the number of businesses DY in the intersection of the business libraries is counted, and the step of obtaining the data necessity DB according to the number of businesses is specifically as follows:

[0110] Step S252: record the businesses in the intersection of the business database as pending businesses, and count the ratio of the number of pending businesses to all businesses handled using the location history. , where i is the number of different pending businesses.

[0111] Step S253: Obtain historical processing data of pending business, and extract the frequency ratio of using access data in the historical processing data. , where i is the number of different pending businesses.

[0112] Step S254: Correlation function based on data necessity The data necessity DB is calculated, where i is the number of different pending businesses, and n is the maximum number of different pending businesses. , is the scaling factor and is greater than 0.

[0113] In actual application, the larger the ratio of the number of pending businesses, the more reasonable it is to retrieve the data at this location. At the same time, the higher the frequency of accessing data in pending businesses, the more reasonable it is to access the data. For example, when a user is in a bank, there are many intersections between the businesses that can be handled by the ID card number and the businesses handled by the bank, such as loans, card processing, financial management, etc., among which pending businesses account for 70% of the total bank business. However, the number of intersections between the businesses that can be handled by medical record information and the businesses handled by the bank is small, and the ratio of the number of each business is small, so the ratio after summing is also small, so it is more reasonable to call the ID card number. Among the pending businesses, such as card processing business, the ID card number is 100% required, so the rationality of accessing the data is increased. The frequency of using the ID card number in the withdrawal business is not 100%, so the rationality will be lower. Judging the rationality of accessing data based on the necessity of data, the risk of data leakage can be obtained. The more unreasonable the use of data, the greater the possibility of others retrieving it.

[0114] The steps of extracting data features and analyzing the sensitivity of the data according to the data features to obtain the data sensitivity are as follows:

[0115] Step S31, extracting data features of access data, and determining whether the data features are unique to the holding user. If the data features are unique to the holding user, it is determined to be a privacy feature.

[0116] Step S32: If the data feature is not unique to the user, the data feature is determined to be a non-privacy feature.

[0117] Step S33: If the data feature is a privacy feature, the risk level of privacy feature leakage is obtained and recorded as data sensitivity.

[0118] The risk level is determined based on the severity of the consequences brought about by the information, and thus the data sensitivity is obtained.

[0119] Step S34: If the data feature is a non-privacy feature, the probability of obtaining the privacy feature through the non-privacy feature is determined and recorded as the privacy probability.

[0120] Step S35, obtaining the risk level of privacy feature leakage corresponding to the privacy probability, and multiplying it by the privacy probability to obtain the data sensitivity.

[0121] In actual application, the consequences of different data leaks are also different. The leakage of private data can endanger the personal safety and property safety of users, while the consequences of non-private data leakage are smaller. According to the privacy feature, the data is judged whether it is private data, so as to obtain the data sensitivity, and the data sensitivity reflects the privacy importance of the data. For example, information such as ID card number and phone number are used exclusively by the user and are not shared with others, so they belong to privacy features. However, the weather conditions, news information and other messages in the mobile phone are information shared by the public, which belong to non-private features. The privacy feature obtains the data sensitivity according to the severity of the consequences. For example, the consequences of ID card number leakage are more serious than those of phone number leakage, so the data sensitivity of ID card number is greater. Non-private features need to multiply the probability of privacy features. For example, the traffic conditions in the mobile phone are non-private features, but the probability of obtaining the user's current location based on the traffic conditions is 50%, and the data sensitivity of the location is 80, so the data sensitivity of the traffic conditions is 40.

[0122] The steps of obtaining all security policies, evaluating the security and convenience of the security policies, and obtaining the policy security and policy convenience are as follows:

[0123] Step S51, obtaining the data encryption complexity FZ, response speed FX and attack success rate FL of the security policy.

[0124] Step S52: according to the policy security correlation function The strategy security degree FA is calculated, where , , is the scaling factor and is greater than 0.

[0125] Step S53, determine whether the security policy is an automatic operation. If the security policy is an automatic operation, obtain the occupied resources and data processing efficiency of the security policy, set the automatic weight ratio corresponding to the occupied resources and data processing efficiency, and calculate the policy convenience based on the automatic weight ratio.

[0126] For example, the automatic weight ratios of resource usage and data processing efficiency are 80% and 20% respectively, the resource usage is 500MB, and the data processing efficiency is 10M / S, then the policy convenience is 500×80%+10×20%=402.

[0127] Step S54: If the security policy is a manual operation, the manual operation time and the number of manual operation steps are obtained, a manual weight ratio corresponding to the manual operation time and the number of manual operation steps is set, and the policy convenience is calculated according to the manual weight ratio.

[0128] In actual application, the use of security policies needs to consider security performance on the one hand, and convenience performance on the other hand. When the encryption of the security policy is more complex, the response speed is faster, and the attack success rate is lower, it means that the security performance of the security policy is better. Security policies are divided into manual and automatic. Security policies such as face verification and fingerprint verification belong to manual operations, while security policies such as access restrictions belong to automatic operations. Automatic operations need to consider the amount of resources occupied by security policies and data processing efficiency, because they will affect the use of edge computing devices and increase the burden on the operation of edge computing devices. Manual operations need to consider the manual operation time and the number of steps, and consider the trouble and inconvenience that manual operations bring to users.

[0129] According to the policy security and convenience, combined with the data security, the policy matching is analyzed and the steps of running the security policy with the highest policy matching are selected. Specifically,

[0130] Step S61, matching the corresponding policy security level according to the data security level and recording it as the standard security level.

[0131] Step S62, screening security policies whose policy safety reaches the standard safety and recording them as candidate security policies.

[0132] Step S63, setting a policy weight ratio of the policy safety and policy convenience of the security policy to be selected, and calculating the policy matching degree of the security policy to be selected according to the policy safety and policy convenience and the corresponding policy weight ratio.

[0133] Step S64, screening and running the candidate security policy with the highest policy matching degree.

[0134] In actual use, first of all, it is necessary to ensure that the data operation is safe. For example, when a user needs to access a bank card number in a milk tea shop, a higher level of face verification is required, while when accessing a bank card number at home, verification can be performed through a verification code. In addition, when ensuring the security of data operation, it is necessary to select a more convenient security strategy to improve the user experience. For example, both security strategy A and security strategy B meet the needs of protecting data security, but the encryption of security strategy A is more complex and the operation occupies more resources. In this case, selecting security strategy B will reduce interference with edge devices and improve the user experience.

[0135] An intelligent edge dynamic security optimization system, by applying an intelligent edge dynamic security optimization method as described above, includes:

[0136] The leakage probability module obtains the usage environment of the edge computing device, evaluates the probability of data leakage based on the usage environment and records it as the data leakage probability.

[0137] The data necessity module obtains the data accessed by the edge computing device and obtains the data necessity degree based on the usage environment analysis.

[0138] The data sensitivity module extracts the data features of the data and analyzes the sensitivity of the data according to the data features to obtain the data sensitivity.

[0139] The data security module comprehensively analyzes the security level of the data and obtains the data security level by combining the probability of data leakage, data necessity and data sensitivity.

[0140] The security policy module obtains all security policies, evaluates the security and convenience of the security policies, and obtains the policy security and policy convenience.

[0141] The policy matching module analyzes the policy matching degree based on the policy security and policy convenience, combined with the data security, and selects the security policy with the highest policy matching degree to run.

[0142] The security optimization module uses Docker container technology to deploy lightweight AI models, trains AI models according to the selection method of security policy operation, and realizes dynamic security optimization of intelligent edge through AI models.

[0143] In actual use, the probability of data leakage is evaluated according to the use environment of edge computing devices, the necessity of data access is confirmed, and the security level of the data is analyzed in combination with the data sensitivity of the accessed data. According to the data security level, a security policy that meets the requirements and is safe and convenient is selected for operation. At the same time, through AI training, automatic and high-speed intelligent edge dynamic security optimization is achieved, which improves the security and convenience of data.

[0144] The implementation principle of this system is as follows: The leakage probability module is used to obtain the use environment of the edge computing device, evaluate the probability of data leakage according to the use location and device use information, and record it as the data leakage probability. The data necessity module is used to obtain the data accessed by the edge computing device, and analyze the necessity of accessing the data according to the services supported by the access data and the services that can be handled at the use location to obtain the data necessity. The data sensitivity module is used to extract the data features of the data, and analyze the sensitivity of the data according to whether the data features are unique to the user holding the privacy features and non-privacy features to obtain the data sensitivity. The data security module is used to comprehensively analyze the security of the data by combining the data leakage probability, data necessity and data sensitivity to obtain the data security. The security policy module is used to obtain all security policies, evaluate the security of the security policy by the complexity of data encryption, response speed and attack success rate, and obtain the policy convenience according to whether the security policy is manual or automatic. The policy matching module analyzes the policy matching according to the policy security and policy convenience, combined with the data security, and selects the security policy with the highest policy matching to run. The security optimization module uses Docker container technology to deploy lightweight AI models, trains AI models according to the selection method of security policy operation, and realizes dynamic security optimization of intelligent edge through AI models.

[0145] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for dynamic security optimization of intelligent edge, characterized in that: The following steps are involved: Obtain the usage environment of the edge computing device, evaluate the probability of data leakage based on the usage environment and record it as the data leakage probability; Obtain the data accessed by the edge computing device, and analyze the necessity of accessing the data based on the usage environment to obtain the data necessity; Extract data features of the data, and analyze the sensitivity of the data according to the data features to obtain data sensitivity; Combine the probability of data leakage, data necessity and data sensitivity to comprehensively analyze the security of the data and obtain the data security level; Obtain all security policies, evaluate the security and convenience of the security policies, and obtain the policy security and policy convenience; According to the policy security and convenience, combined with data security, the policy matching is analyzed and the security policy with the highest policy matching is selected for operation; Use Docker container technology to deploy lightweight AI models, train AI models based on the selection method of security policy operation, and achieve dynamic security optimization of intelligent edge through AI models; The step of obtaining the use environment of the edge computing device, evaluating the probability of data leakage according to the use environment and recording it as the data leakage probability is specifically as follows: Obtaining a usage environment of the edge computing device, wherein the usage environment includes a usage location and device usage information; Obtain historical location data of edge computing devices, and extract common locations based on the historical location data; Determine whether the usage location is a common location. If the usage location is not a common location, obtain the usage habits of the user holding the edge computing device, analyze the usage habits, and obtain the data leakage probability based on the usage habits and record it as the first leakage probability; If the usage location is a common location, obtaining a preset data leakage probability as a first leakage probability; Analyze the device usage information to obtain a second leakage probability, calculate the average of the first leakage probability and the second leakage probability, and obtain the data leakage probability; The step of determining whether the usage location is a common location, and if the usage location is not a common location, obtaining the usage habits of the user holding the edge computing device, and obtaining the data leakage probability according to the usage habits and recording it as the first leakage probability, is specifically as follows: Get the average crowd density SM of the usage location; The usage habits include the usage height SG, the font size SZ, and the usage duration SS of the edge computing device; According to the first leakage correlation function The first leakage probability SY is calculated, where , is the scaling factor and is greater than 0.

2. The method for dynamic security optimization of intelligent edge according to claim 1, characterized in that: The step of obtaining the second leakage probability according to the device usage information analysis is specifically: The device usage information includes physical usage information and network usage information; The physical usage information includes real-time temperature AW, real-time humidity AS and electromagnetic intensity AD; Obtain the standard temperature BW and standard humidity BS used by edge computing devices; According to the damage correlation function The equipment damage probability AL is calculated, where , , is the scale factor and is greater than 0; The network security degree AA of the network used by the edge computing device is obtained based on the network usage information; According to the second leakage correlation function The second leakage probability BE is calculated, where , is the scaling factor and is greater than 0.

3. The method for dynamic security optimization of intelligent edge according to claim 2, characterized in that: The step of obtaining the data accessed by the edge computing device and analyzing the necessity of accessing the data according to the usage environment to obtain the data necessity degree is specifically as follows: Obtain the data accessed by the edge computing device and record it as access data, and extract the user privacy data in the access data; Find the usage business associated with the user's privacy data and form a standard business library; Find all the businesses that have been handled using location history and form a real-time business database; The intersection of the business libraries is formed based on the standard business library and the real-time business library. If the intersection of the business libraries is 0, it is determined that the data is not needed, and the data necessity DB is 0; If the intersection of the business libraries is not 0, the number of businesses DY in the intersection of the business libraries is counted, and the data necessity DB is obtained according to the number of businesses.

4. The method for dynamic security optimization of intelligent edge according to claim 3, characterized in that: If the intersection of the business libraries is not 0, the number of businesses DY in the intersection of the business libraries is counted, and the step of obtaining the data necessity DB according to the number of businesses is specifically as follows: Record the businesses in the intersection of the business database as pending businesses, and count the ratio of the number of pending businesses among all businesses handled using the location history. , where i is the number of different pending businesses; Obtain the historical processing data of pending business and extract the frequency ratio of access data used in the historical processing data , where i is the number of different pending businesses; According to the data necessity correlation function The data necessity DB is calculated, where i is the number of different pending businesses, and n is the maximum number of different pending businesses. , is the scaling factor and is greater than 0.

5. The method for dynamic security optimization of intelligent edge according to claim 4, characterized in that: The step of extracting data features and analyzing the sensitivity of the data according to the data features to obtain the data sensitivity is specifically as follows: Extract the data features of the access data and determine whether the data features are unique to the holding user. If the data features are unique to the holding user, determine that they are privacy features. If the data feature is not unique to the user holding it, the data feature is judged to be a non-privacy feature; If the data feature is a privacy feature, the risk level of privacy feature leakage is obtained and recorded as data sensitivity; If the data feature is a non-privacy feature, the probability of obtaining the privacy feature through the non-privacy feature is determined and recorded as the privacy probability; The risk level of privacy feature leakage corresponding to the privacy probability is obtained, and the data sensitivity is obtained by multiplying it by the privacy probability.

6. The method for dynamic security optimization of intelligent edge according to claim 5, characterized in that: The steps of obtaining all security policies, evaluating the security and convenience of the security policies, and obtaining the policy security and policy convenience are specifically as follows: Obtain the data encryption complexity FZ, response speed FX and attack success rate FL of the security policy; According to the policy security correlation function The strategy security degree FA is calculated, where , , is the scale factor and is greater than 0; Determine whether the security policy is an automatic operation. If the security policy is an automatic operation, obtain the occupied resources and data processing efficiency of the security policy, set the automatic weight ratio corresponding to the occupied resources and data processing efficiency, and calculate the policy convenience based on the automatic weight ratio; If the security policy is a manual operation, the manual operation time and the number of manual operation steps are obtained, a manual weight ratio corresponding to the manual operation time and the number of manual operation steps is set, and the policy convenience is calculated according to the manual weight ratio.

7. The method for dynamic security optimization of intelligent edge according to claim 6, characterized in that: The steps of analyzing the policy matching degree according to the policy security and policy convenience, combined with the data security, and selecting the security policy with the highest policy matching degree to run are specifically as follows: Match the corresponding policy security level according to the data security level and record it as the standard security level; Screen the security policies whose policy safety reaches the standard safety and record them as candidate security policies; Set the policy weight ratio of the policy security and policy convenience of the security policy to be selected, and calculate the policy matching degree of the security policy to be selected according to the policy security and policy convenience and the corresponding policy weight ratio; The candidate security policy with the highest policy matching degree is screened and run.

8. An intelligent edge dynamic security optimization system, characterized in that: By applying an intelligent edge dynamic security optimization method as described in any one of claims 1 to 7, comprising: The leakage probability module obtains the usage environment of the edge computing device, evaluates the probability of data leakage based on the usage environment and records it as the data leakage probability; The data necessity module obtains the data accessed by the edge computing device and analyzes the necessity of accessing the data according to the usage environment to obtain the data necessity degree; The data sensitivity module extracts the data features of the data and analyzes the sensitivity of the data according to the data features to obtain the data sensitivity; The data security module comprehensively analyzes the security level of data and obtains the data security level by combining the probability of data leakage, data necessity and data sensitivity; The security policy module obtains all security policies, evaluates the security and convenience of the security policies, and obtains the policy security and policy convenience; The policy matching module analyzes the policy matching degree based on the policy security and policy convenience, combined with the data security, and selects the security policy with the highest policy matching degree to run; The security optimization module uses Docker container technology to deploy lightweight AI models, trains AI models based on the selection method of security policy operation, and implements dynamic security optimization of intelligent edge through AI models; The step of obtaining the use environment of the edge computing device, evaluating the probability of data leakage according to the use environment and recording it as the data leakage probability is specifically as follows: Obtaining a usage environment of the edge computing device, wherein the usage environment includes a usage location and device usage information; Obtain historical location data of edge computing devices, and extract common locations based on the historical location data; Determine whether the usage location is a common location. If the usage location is not a common location, obtain the usage habits of the user holding the edge computing device, analyze the usage habits, and obtain the data leakage probability based on the usage habits and record it as the first leakage probability; If the usage location is a common location, obtaining a preset data leakage probability as a first leakage probability; Analyze the device usage information to obtain a second leakage probability, calculate the average of the first leakage probability and the second leakage probability, and obtain the data leakage probability; The step of determining whether the usage location is a common location, and if the usage location is not a common location, obtaining the usage habits of the user holding the edge computing device, and obtaining the data leakage probability according to the usage habits and recording it as the first leakage probability, is specifically as follows: Get the average crowd density SM of the usage location; The usage habits include the usage height SG, the font size SZ, and the usage duration SS of the edge computing device; According to the first leakage correlation function The first leakage probability SY is calculated, where , is the scaling factor and is greater than 0.

Images