A multi-person collaborative penetration test data sharing method and system

By aggregating and numbering data in a shared database in real time, and combining image recognition and deep learning training with distributed cloud nodes and a central management platform, the security and efficiency issues of external personnel identification in the shared database are solved, achieving flexible identification and comprehensive security protection.

CN118586008BActive Publication Date: 2024-11-12XINCHAO RUISHI INFORMATION TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411066774.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-06
Publication Date
2024-11-12
Estimated Expiration
2044-08-06

AI Technical Summary

Technical Problem

Existing shared databases lack security in identifying external personnel, have a single and inefficient identification mode, and are difficult to effectively distinguish between internal employees and external personnel, causing inconvenience to internal employees.

Method used

A multi-person collaborative penetration testing data sharing method is adopted. Penetration testing data is aggregated to a shared database in real time, numbered and divided, and image recognition and deep learning training are performed using distributed cloud nodes and a central management platform. Combined with partition monitoring and video surveillance, judgment algorithms and reaction models are used to make security judgments, and a cache cloud and big data repository are set up for information management and identification.

Benefits of technology

It enables flexible identification of outsiders, improves identification speed and security, ensures convenience for internal personnel, and provides comprehensive security protection for the shared database.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118586008B_ABST
    Figure CN118586008B_ABST
Patent Text Reader

Abstract

The present invention discloses a multi-person collaborative penetration test data sharing method and system, which relates to the field of computer technology, including: a data generation terminal performs face collection preprocessing on internal personnel of a shared database, the data generation terminal uploads information to a distributed cloud node, the distributed cloud node uploads data to a central management platform, and the central management platform stores the data in a big data repository; the central management platform determines the safety of personnel according to a determination algorithm, temporarily stores the information of the personnel determined to be safe in a cache cloud, and the central management platform allows the personnel stored in the cache cloud to enter the shared database, but does not allow the personnel stored in the cache cloud to enter the shared database. By setting a central management platform, a video monitoring platform, a distributed cloud node and a data generation terminal, external personnel are identified, the security management capability of the shared database for external personnel is improved, and convenience is brought to the internal personnel of the shared database.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a multi-person collaborative penetration test data sharing method and system. Background Art

[0002] Penetration testing is an authorized simulated attack on a computer system to evaluate its security. It is a targeted security test of the application system from the attacker's perspective. It reduces risks by actively discovering security risks and vulnerabilities in the application system. It is a mechanism to prove that network defense is operating normally according to the expected plan. Penetration testing is an essential stage in the software security development life cycle. In order to enhance its penetration efficiency, penetration testing chooses to share penetration data.

[0003] The existing shared database may have security issues for the entry of outsiders. Although the existing intelligent shared database can automatically identify the entry of outsiders, the identification mode is single and will exclude security personnel who are not in this shared database, causing inconvenience to internal employees. In addition, the algorithm efficiency is not high and it is necessary to traverse the entire stored data to identify the source of the personnel. The algorithm efficiency needs to be improved. Summary of the invention

[0004] In order to solve the above technical problems, a multi-person collaborative penetration testing data sharing method and system are provided. This technical solution solves the problems raised in the above background technology.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A multi-person collaborative penetration test data sharing method, characterized by comprising:

[0007] Aggregate penetration test data to a shared database in real time and divide the shared database into numbers;

[0008] The data generation terminal performs face collection preprocessing on the internal personnel of the shared database, and the data generation terminal uploads the information to the distributed cloud node, and the distributed cloud node uploads the data to the central management platform, and the central management platform stores the data in the big data repository, and the big data repository completes the storage of the images of the internal personnel of the shared database;

[0009] The central management platform determines the safety of personnel according to the judgment algorithm, and temporarily stores the information of the personnel determined to be safe in the cache cloud. The central management platform allows the personnel stored in the cache cloud to enter the shared database, but does not allow the personnel stored in the cache cloud to enter the shared database;

[0010] The central management platform uses simulated data for deep learning training and forms a response model;

[0011] Different monitoring platforms are used to carry out zone monitoring in different areas of the shared database. The personnel images or video information generated by the monitoring are transmitted to the data generation terminal, which transmits the information to the distributed cloud node, which transmits the information to the central management platform. The central management platform calls the internal personnel images of the shared database stored in the big data repository and the data temporarily stored in the cache cloud, adopts a reaction model, makes decision instructions, transmits them to the distributed cloud node, which is then transmitted to the data generation terminal by the distributed cloud node, and the data generation terminal feeds back the instructions to the monitoring platforms in different areas of the shared database, and the monitoring platform responds according to the instructions.

[0012] The central management platform stores long-term data in the big data repository, and the central management platform stores short-term data in the cache cloud;

[0013] The central management platform collects warning information and stores it in the warning module;

[0014] Carry out zone monitoring for different areas of the shared database, and use a video monitoring platform to monitor the key areas and non-confidential areas of the shared database;

[0015] The video surveillance platform obtains facial image information and human behavior video information. The image recognition platform calls the data in the big data repository to compare the information obtained by the video surveillance platform. When the comparison obtains the information of a person outside the big data repository, the video recording platform is activated to record the person and transmit the video to the big data repository. The central management platform uses the judgment algorithm to determine whether the person is a safe person or a dangerous person.

[0016] The entrance of the shared database captures the image of the person. If the image of the person is stored in the big data storage repository, it is released. If the image of the person does not belong to the big data storage repository, the entrance of the shared database does not release it, and the alarm linkage console sends an alarm to the shared database security terminal;

[0017] The internal communication station is used to match and communicate between outsiders and staff in the shared database. The information verification station is used to verify the safety of outsiders. The information verification station updates the corresponding information in the cache cloud, and the cache cloud lists the person as a temporary safe person.

[0018] When the big data repository information is called, the data in the big data repository is automatically matched to generate a directed graph, and the data in the big data repository is called according to the directed graph.

[0019] Preferably, the determination algorithm specifically includes:

[0020] Count the number of times a person enters a shared database;

[0021] Based on the behavior of each person entering the shared database, the person is scored according to the shared database rules, and the behavior score of each person entering the shared database is obtained. If one of the shared database rules is met, one point will be added; if one of the shared database rules is violated, one point will be deducted. ;

[0022] Determine whether the behavior score of the person entering the shared database each time is greater than a first preset value. If so, mark the person as a person to be verified; if not, determine the person as a dangerous person;

[0023] For the person to be verified, the person's assessment score is calculated through a comprehensive scoring formula based on the number of times the person enters the shared database and the behavioral score of each entry into the shared database;

[0024] When judging whether the personnel assessment score is greater than a second preset value, if so, the person is judged as a safe person, and if not, the person is judged as a dangerous person;

[0025] The comprehensive scoring formula is:

[0026] ; Among them, A is the personnel evaluation score, Score each entry into the shared database, where n is the number of entries into the shared database.

[0027] Preferably, the central management platform uses simulated data for deep learning training, specifically comprising the following steps:

[0028] Generate a flow diagram, which includes ,in, Store response command information, Store various first-level instructions, Store various secondary instructions, Stores common data mode information, Pointing to the corresponding Pointing to the corresponding , thus forming The rationality of the formed directed chain is verified by using simulated data for monitoring. The central management platform receives the data and matches the corresponding There is a corresponding directed chain , the central management platform executes The first level instructions and The secondary instructions in the directed chain are determined according to the rationality of the actual effect of the instructions. The rationality of If it is unreasonable, then the first-level instruction or the second-level instruction must be unreasonable. If the first-level instruction is unreasonable, change Directed If the secondary instruction is unreasonable, change Directed , for the new The directed chain continues to be verified until all directed chains generate instructions that meet the actual requirements for the simulated data. The central management platform saves all directed chain information to form a reaction model.

[0029] According to a multi-person collaborative penetration test data sharing method according to claim 3, it is characterized in that the real-time aggregation of the penetration test data to the shared database and the numbering of the shared database include the following steps:

[0030] Number the internal personnel of the shared database stored in the big data repository. ;

[0031] The multiple blocks divided into the shared database are numbered as follows: ;

[0032] Number the shared databases. .

[0033] Preferably, capturing the image of a person at the entrance of the shared database comprises the following steps:

[0034] The entry of the shared database extracts the image information of the person;

[0035] The entry of the shared database transmits information to the distributed cloud nodes;

[0036] Distributed cloud nodes transmit information to the central management platform;

[0037] The central management platform calls the big data repository to compare whether the person's image information has been stored. If it has been stored, the person is an internal person in the shared database. If it has not been stored, the person is an outsider. The entrance to the shared database prohibits people from entering the shared database.

[0038] Preferably, the temporarily stored data in the cache cloud specifically includes:

[0039] Clean up information immediately, within the first preset time;

[0040] Temporarily store information for a short period of time and clear it within a second preset time;

[0041] Long-term temporary storage of information, which will be cleared within a third preset time;

[0042] The third preset time is greater than the second preset time, and the second preset time is greater than the first preset time.

[0043] Preferably, the automatic matching of data in the big data repository to generate a directed graph comprises the following steps:

[0044] Statistics on internal personnel in establishing a shared database Frequently appearing areas, which are blocks in the shared database and shared database , establish a mapping, Map to and , mapping all to of Composition Collection , mapping all to of Composition Collection , forming a shared database of blocks and shared database A directed graph V.

[0045] Preferably, calling the data in the big data storage repository according to the directed graph comprises the following steps:

[0046] Determine the location of the person;

[0047] If the person appears in each block in the shared database , then the information calling module transmits instructions, and the cloud-edge collaboration module calls In , the cloud-edge collaboration module compares personnel and image information to determine whether the person is an internal member of the shared database;

[0048] If the person's location is a shared database , then the information calling module transmits instructions, and the cloud-edge collaboration module calls In , the cloud-edge collaboration module compares personnel and image information to determine whether the person is an internal member of the shared database.

[0049] A multi-person collaborative penetration test data sharing system, used to implement the multi-person collaborative penetration test data sharing method, comprising:

[0050] The cloud-edge collaboration module is composed of a central management platform, a big data repository, distributed cloud nodes, data generation terminals, a cache cloud and a cache cloud. The cloud-edge collaboration module performs large-scale overall data analysis, deep learning training, and autonomously generates commands for feedback data;

[0051] An early warning module, which interacts with the cloud-edge collaboration module. The cloud-edge collaboration module issues instructions to control the early warning module to issue an early warning, issue an early warning to personnel, and expel the personnel from the shared database or the shared database;

[0052] Image monitoring module, which is composed of a video monitoring platform, a video recording platform and an image recognition platform. The image monitoring module is equipped with cards at key locations, which can call up stored employee photos for confirmation and identify visitors through image scanning and comparison;

[0053] A control module, which is arranged in the shared database and is composed of an entrance to the shared database and an alarm linkage control console. The control module can provide a complete security control function;

[0054] An information matching auxiliary module, which is composed of an internal communication station and an information verification station. The information matching auxiliary module is used to assist the cloud-edge collaboration module in its operation and improve the functions of the cloud-edge collaboration module.

[0055] An information calling module is used for calling information in a big data repository, and the information calling module enhances the calling efficiency of the cloud-edge collaboration module.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. By setting up image monitoring module, control module, information matching auxiliary module and cache cloud, it is possible to flexibly identify outsiders and allow some personnel to enter the shared database, bringing convenience to the internal personnel of the shared database.

[0058] 2. By setting up the information call module, you can identify people who appear in the area based on the regular activity range of people in the shared database. When identifying, you only need to call the data of people who appear in the area, which reduces the calculation amount of the program algorithm and improves the identification speed.

[0059] 3. By setting up cloud-edge collaboration modules, image monitoring modules, control modules and early warning modules, the entire shared database can be protected in an integrated manner, and both visible and invisible areas can be protected to ensure the security of the shared database. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a flow chart of the multi-person collaborative penetration test data sharing method of the present invention;

[0061] Figure 2 It is a schematic diagram of the flow chart of the determination algorithm of the present invention;

[0062] Figure 3 A schematic diagram of a process of aggregating penetration test data into a shared database in real time and dividing the shared database into numbers according to the present invention;

[0063] Figure 4 A schematic diagram of a process for capturing images of persons at the entry of a shared database of the present invention;

[0064] Figure 5 The figure is a schematic diagram of the process of temporarily storing data in the cache cloud of the present invention. DETAILED DESCRIPTION

[0065] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0066] Reference Figure 1 As shown, a multi-person collaborative penetration test data sharing method includes:

[0067] The penetration test data is aggregated in real time to a shared database, and the shared database is numbered and divided. The cloud-edge collaboration module uses the information matching auxiliary module and the information calling module to call internal information based on the information fed back by the image monitoring module and the control module, and issues an alarm through the early warning module to complete the security protection of the shared database.

[0068] The cloud-edge collaboration module consists of a central management platform, a big data repository, distributed cloud nodes, data generation terminals, and a cache cloud. The data generation terminals perform face collection preprocessing on people inside the shared database. The data generation terminals upload information to the distributed cloud nodes. The distributed cloud nodes upload data to the central management platform. The central management platform stores data in the big data repository. The big data repository completes the storage of images of people inside the shared database.

[0069] The central management platform determines the safety of personnel according to the judgment algorithm, and temporarily stores the information of the personnel determined to be safe in the cache cloud. The central management platform allows the personnel stored in the cache cloud to enter the shared database, but does not allow the personnel stored in the cache cloud to enter the shared database;

[0070] The central management platform uses simulated data for deep learning training and forms a response model;

[0071] Different monitoring platforms are used to carry out zone monitoring in different areas of the shared database. The personnel images or video information generated by the monitoring are transmitted to the data generation terminal, which transmits the information to the distributed cloud node, which transmits the information to the central management platform. The central management platform calls the internal personnel images of the shared database stored in the big data repository and the data temporarily stored in the cache cloud, adopts a reaction model, makes decision instructions, transmits them to the distributed cloud node, which is then transmitted to the data generation terminal by the distributed cloud node, and the data generation terminal feeds back the instructions to the monitoring platforms in different areas of the shared database, and the monitoring platform responds according to the instructions.

[0072] The central management platform stores long-term data in the big data repository, and the central management platform stores short-term data in the cache cloud;

[0073] The central management platform collects warning information and stores it in the warning module;

[0074] The cache cloud temporary storage data is classified into three levels: level one is immediate cleaning information, which is cleaned within the first preset time, which is less than ten minutes; level two is short-term temporary storage information, which is cleaned within the second preset time, which is between 5 hours and 6 hours; level three is long-term temporary storage information, which is cleaned within the third preset time, which is 24 hours;

[0075] The cache cloud classifies information according to its importance, and then determines its existence time in the cache cloud. Unimportant information is deleted immediately, and relatively important information has its existence time increased. The evaluation criteria are pre-entered by the shared database staff.

[0076] The early warning module exchanges information with the cloud-edge collaboration module. The cloud-edge collaboration module issues instructions to control the early warning module to issue early warnings. The early warning module can warn and expel suspicious persons through broadcasting. If the warning is ineffective, the shared database security system will send relevant personnel to handle the matter.

[0077] The image monitoring module consists of a video monitoring platform, an image recording platform and an image recognition platform. The video monitoring platform is set up in the key areas of the shared database, important facilities of the shared database and the non-confidential areas of the shared database. The video monitoring platform obtains facial image information and human behavior video information. The image recognition platform calls the cloud-edge collaboration module data to compare the information obtained by the video monitoring platform. When the information of people outside the cloud-edge collaboration module is obtained by comparison, the image recording platform is activated to record the person and transmit the video to the cloud-edge collaboration module. The central management platform compresses the video and stores it in the big data repository. The cloud-edge collaboration module uses a judgment algorithm to determine whether the person is a safe person or a dangerous person.

[0078] When the image recognition platform calls the cloud-edge collaboration module data, it uses the information matching auxiliary module. According to the location where the image recognition platform data is obtained, the information matching auxiliary module retrieves the internal personnel information of the shared database that appears at that location for comparison by the cloud-edge collaboration module. If it is not an internal personnel of the shared database, it is an outsider.

[0079] The control module is set in the shared database, and mainly controls personnel entering the shared database. The control module consists of the entrance of the shared database and the alarm linkage console. The entrance of the shared database releases the personnel stored in the cloud-edge collaboration module. When the entrance of the shared database identifies personnel outside the cloud-edge collaboration module, the alarm linkage console sends an alarm to the shared database security terminal, and the shared database security expels outsiders out of the shared database.

[0080] During identification, when the control module calls the cloud-edge collaboration module data, it uses the information matching auxiliary module. According to the location where the control module data is obtained, the information matching auxiliary module retrieves the internal personnel information of the shared database that appears at that location for comparison by the cloud-edge collaboration module. If it is not an internal personnel of the shared database, it is an outsider.

[0081] The information matching auxiliary module consists of an internal communication station and an information verification station. The internal communication station is used for matching and communicating between outsiders and staff in the shared database, and the information verification station is used to verify the safety of outsiders. Through the internal communication station and the information verification station, the safety of outsiders can be coordinated and verified by internal personnel in the shared database, and the corresponding information can be updated in the cache cloud in the cloud-edge collaboration module. The cache cloud lists the person as a temporary safe person. After one day, the cache cloud deletes the person's information. During this period, the person can enter the guarantor's location;

[0082] The information calling module generates a directed graph for the data in the cloud-edge collaboration module. The information calling module calls the data in the cloud-edge collaboration module according to the directed graph. The information calling module assists the above modules in making data calls.

[0083] The decision algorithms used by the cloud-edge collaboration module include:

[0084] Count the number of times a person enters a shared database;

[0085] Based on the behavior of each person entering the shared database, the person is scored according to the shared database rules, and the behavior score of each person entering the shared database is obtained. If one of the shared database rules is met, one point will be added; if one of the shared database rules is violated, one point will be deducted. ;

[0086] Determine whether the behavior score of the person entering the shared database each time is greater than a first preset value. If so, mark the person as a person to be verified; if not, determine the person as a dangerous person;

[0087] For the person to be verified, the person's assessment score is calculated through a comprehensive scoring formula based on the number of times the person enters the shared database and the behavioral score of each entry into the shared database;

[0088] When judging whether the personnel assessment score is greater than a second preset value, if so, the person is judged as a safe person, and if not, the person is judged as a dangerous person;

[0089] The comprehensive scoring formula is:

[0090] ; Among them, A is the personnel evaluation score, Score each entry into the shared database, where n is the number of times the shared database is entered;

[0091] That is, when a person's performance complies with some of the shared database rules, even if he is not an internal member of the shared database, he is still allowed to appear in areas outside the shared database. When a single malicious incident occurs, the score is lower than the first preset value, and he is directly scored as unqualified, and is prohibited from entering the shared database in the future, thereby ensuring the security of the shared database and flexible identification.

[0092] The central management platform uses simulated data for deep learning training, which specifically includes the following steps:

[0093] Generate a flow diagram, which includes ,in, Store response command information, Store various first-level instructions, Store various secondary instructions, Stores common data mode information, Pointing to the corresponding Pointing to the corresponding , thus forming The rationality of the formed directed chain is verified by using simulated data for monitoring. The central management platform receives the data and matches the corresponding There is a corresponding directed chain , the central management platform executes The first level instructions and The secondary instructions in the directed chain are determined according to the rationality of the actual effect of the instructions. The rationality of If it is unreasonable, then the first-level instruction or the second-level instruction must be unreasonable. If the first-level instruction is unreasonable, change Directed If the secondary instruction is unreasonable, change Directed , for the new The directed chains continue to be verified until all directed chains generate instructions that meet the actual requirements for the simulated data. The central management platform saves all directed chain information to form a reaction model.

[0094] During training, Can correspond to multiple , It can also correspond to multiple , forming multiple A directed chain of , which will lead to , and then point to , thus obtaining a series of operations, according to the rationality of the operations, Corresponding multiple Filter and limit, narrow range, yes Corresponding multiple Filter and limit, narrow range, so that it is possible to re-establish A directed chain of The directed chain is more reasonable, so as to complete the training, and the improvement can also be , , , , On the basis of , and train again in the above manner.

[0095] The penetration test data is aggregated in real time to a shared database, and the shared database is numbered and divided, including the following steps:

[0096] Number the internal personnel of the shared database stored in the big data repository. ;

[0097] The multiple blocks divided into the shared database are numbered as follows: ;

[0098] Number the shared databases. .

[0099] The entry of the shared database to capture images of people includes the following steps:

[0100] The entry of the shared database extracts the image information of the person;

[0101] The entry of the shared database transmits information to the distributed cloud nodes;

[0102] Distributed cloud nodes transmit information to the central management platform;

[0103] The central management platform calls the big data repository to compare whether the person's image information has been stored. If it has been stored, the person is an internal person in the shared database. If it has not been stored, the person is an outsider. The entrance to the shared database prohibits people from entering the shared database.

[0104] The information calling module generates a directed graph for the data in the cloud-edge collaboration module, including the following steps:

[0105] Statistics on internal personnel in establishing a shared database Frequently appearing areas, which are blocks in the shared database and shared database , establish a mapping, Map to and , mapping all to of Composition Collection , mapping all to of Composition Collection , forming a shared database of blocks and shared database Therefore, when calling data comparison, you only need to call the corresponding and , because people inside the shared database must appear and Therefore, and Search, if the person is not and If the personnel in the shared database are not inside the shared database, this comparison does not require calling all the data, which can reduce the pressure on the server and speed up the identification.

[0106] The information calling module calls the data in the cloud-edge collaboration module according to the directed graph, including the following steps:

[0107] Determine the location of the person;

[0108] If the person appears in each block in the shared database , then the information calling module transmits instructions, and the cloud-edge collaboration module calls In , the cloud-edge collaboration module compares personnel and image information to determine whether the person is an internal member of the shared database;

[0109] If the person's location is a shared database , then the information calling module transmits instructions, and the cloud-edge collaboration module calls In , the cloud-edge collaboration module compares personnel and image information to determine whether the person is an internal member of the shared database.

[0110] Furthermore, based on the same inventive concept as the above-mentioned multi-person collaborative penetration test data sharing method, the present invention proposes a multi-person collaborative penetration test data sharing system, comprising:

[0111] The cloud-edge collaboration module performs large-scale overall data analysis, deep learning training, and autonomously generates commands for feedback data;

[0112] Early warning module: According to the instructions of the cloud-edge collaboration module, the early warning module issues early warnings to personnel and expels them from the shared database or the shared database;

[0113] Image monitoring module: The image monitoring module is set up at key locations to call up stored employee photos for confirmation and identify visitors through image scanning and comparison;

[0114] Control module: the control module can provide complete security control functions;

[0115] Information matching auxiliary module: The information matching auxiliary module is used to assist the cloud-edge collaboration module and improve the functions of the cloud-edge collaboration module;

[0116] Information calling module,The information calling module enhances the calling efficiency of the cloud-edge collaboration module.

[0117] Furthermore, the present solution also proposes a storage medium on which a computer-readable program is stored, and when the computer-readable program is called, the above-mentioned multi-person collaborative penetration test data sharing method is executed.

[0118] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0119] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A multi-person collaborative penetration test data sharing method, It is characterized in that include: Aggregate penetration test data to a shared database in real time and divide the shared database into numbers; The big data repository completes the storage of images of people within the shared database; The central management platform determines personnel safety based on the judgment algorithm; The central management platform uses simulated data for deep learning training and forms a response model; Different monitoring platforms are used to carry out zone monitoring in different areas of the shared database. The personnel images or video information generated by the monitoring are transmitted to the data generation terminal, which transmits the information to the distributed cloud node, which transmits the information to the central management platform. The central management platform calls the internal personnel images of the shared database stored in the big data repository and the data temporarily stored in the cache cloud, adopts a reaction model, makes decision instructions, transmits them to the distributed cloud node, which is then transmitted to the data generation terminal by the distributed cloud node, and the data generation terminal feeds back the instructions to the monitoring platforms in different areas of the shared database, and the monitoring platform responds according to the instructions. The central management platform stores long-term data in the big data repository, and the central management platform stores short-term data in the cache cloud; The central management platform collects warning information and stores it in the warning module; Carry out zone monitoring for different areas of the shared database, and use a video monitoring platform to monitor the key areas and non-confidential areas of the shared database; The video surveillance platform obtains facial image information and human behavior video information. The image recognition platform calls the data in the big data repository to compare the information obtained by the video surveillance platform. When the comparison obtains the information of a person outside the big data repository, the video recording platform is activated to record the person and transmit the video to the big data repository. The central management platform uses the judgment algorithm to determine whether the person is a safe person or a dangerous person. The entrance of the shared database captures the image of the person. If the image of the person is stored in the big data storage repository, it is released. If the image of the person does not belong to the big data storage repository, the entrance of the shared database does not release it, and the alarm linkage console sends an alarm to the shared database security terminal; The internal communication station is used to match and communicate between outsiders and staff in the shared database. The information verification station is used to verify the safety of outsiders. The information verification station updates the corresponding information in the cache cloud, and the cache cloud lists the person as a temporary safe person. When the information in the big data repository is called, the data in the big data repository is automatically matched to generate a directed graph, and the data in the big data repository is called according to the directed graph; The step of aggregating the penetration test data in real time to a shared database and dividing the shared database by numbers comprises the following steps: Number the internal personnel of the shared database stored in the big data repository, numbered b p ; The multiple blocks divided into the shared database are numbered c q ; Number the shared database, numbered d s ; The automatic matching of data in the big data repository to generate a directed graph comprises the following steps: Statistics establishment of shared database internal personnel b p Frequently appearing areas, the areas are blocks in the shared database c f and shared database j , establish a mapping, b p Map to c f and d j , mapping all to c f b p Composition set C f , mapping all to d j b p Composition set D j , forming a shared database of blocks c f and shared database j A directed graph V of The calling of data in the big data repository according to the directed graph comprises the following steps: Determine the location of the person; If the location of the person is the blocks c in the shared database f , then the information calling module transmits instructions, and the cloud-edge collaboration module calls C f b p , the cloud-edge collaboration module compares personnel and b p image information to determine whether the person is an internal member of the shared database; If the person's location is a shared database d j , then the information calling module transmits instructions, and the cloud-edge collaboration module calls D j b p , the cloud-edge collaboration module compares personnel and b p image information to determine whether the person is an internal member of the shared database.

2. A multi-person collaborative penetration test data sharing method according to claim 1, It is characterized in that The big data repository completes the storage of personnel images in the shared database, including the following steps: The data generation terminal performs face collection preprocessing on the internal personnel of the shared database, and the data generation terminal uploads the information to the distributed cloud node. The distributed cloud node uploads the data to the central management platform. The central management platform stores the data in the big data repository, and the big data repository completes the image storage of the internal personnel of the shared database.

3. A multi-person collaborative penetration test data sharing method according to claim 2, It is characterized in that The determination algorithm specifically includes: Count the number of times a person enters a shared database; Based on the behavior of each person entering the shared database, the person is scored according to the shared database rules, and the behavior score a of each person entering the shared database is obtained. m If one of the shared database rules is met, one point will be added; if one of the shared database rules is violated, one point will be deducted. The cumulative score is a m ; Determine whether the behavior score of the person entering the shared database each time is greater than a first preset value. If so, mark the person as a person to be verified; if not, determine the person as a dangerous person; For the person to be verified, the person's assessment score is calculated through a comprehensive scoring formula based on the number of times the person enters the shared database and the behavioral score of each entry into the shared database; When judging whether the personnel assessment score is greater than a second preset value, if so, the person is judged as a safe person, and if not, the person is judged as a dangerous person; The comprehensive scoring formula is: Among them, A is the personnel evaluation score, a m Score each entry into the shared database, where n is the number of entries into the shared database.

4. A multi-person collaborative penetration test data sharing method according to claim 3, It is characterized in that The central management platform determines the safety of personnel according to the determination algorithm, including the following steps: The information of those judged as safe personnel is temporarily stored in the cache cloud. The central management platform allows the personnel stored in the cache cloud to enter the shared database, but does not allow the personnel stored in the cache cloud to enter the shared database.

5. A multi-person collaborative penetration test data sharing method according to claim 4, It is characterized in that The central management platform uses simulated data for deep learning training, which specifically includes the following steps: Generate a flow diagram, which includes v 1 , v 2 …v 100 ,u 1 ,u 2 …u 100 , w 1 , w 2 …w 100 , where u 1 ,u 2 …u 100 , w 1 , w 2 …w 100 Store response command information, u 1 ,u 2 …u 100 Store various first-level instructions, w 1 , w 2 …w 100 Store various secondary instructions, v 1 , v 2 …v 100 Stores common data mode information, u i Point to the corresponding w l , v k Points to the corresponding u i , thus forming v k →u i →w l The rationality of the formed directed chain is verified by using simulated data for monitoring. The central management platform receives the data and matches the corresponding v according to the data pattern. k , v k There is a corresponding directed chain v k →u i →w l The central management platform executes u i The first level instruction and w l The secondary instructions in the directed chain v are determined according to the rationality of the actual effect of the instructions. k →u i →w l The rationality of k →u i →w l If it is unreasonable, then the first-level instruction or the second-level instruction must be unreasonable. If the first-level instruction is unreasonable, change v k Pointing to u i , if the secondary instruction is unreasonable, change u i Pointing to w l , for the new v k →u i →w l The directed chain continues to be verified until all directed chains generate instructions that meet the actual requirements for the simulated data. The central management platform saves all directed chain information to form a reaction model.

6. A multi-person collaborative penetration test data sharing method according to claim 5, It is characterized in that The entry of the shared database to capture the image of a person comprises the following steps: The entry of the shared database extracts the image information of the person; The entry of the shared database transmits information to the distributed cloud nodes; Distributed cloud nodes transmit information to the central management platform; The central management platform calls the big data repository to compare whether the person's image information has been stored. If it has been stored, the person is an internal person in the shared database. If it has not been stored, the person is an outsider. The entrance to the shared database prohibits people from entering the shared database.

7. A multi-person collaborative penetration test data sharing method according to claim 6, It is characterized in that The temporary data in the cache cloud specifically includes: Clean up information immediately, within the first preset time; Temporarily store information for a short period of time and clear it within a second preset time; Long-term temporary storage of information, which will be cleared within a third preset time; The third preset time is greater than the second preset time, and the second preset time is greater than the first preset time.

8. A multi-person collaborative penetration test data sharing system, used to implement the multi-person collaborative penetration test data sharing method as described in any one of claims 1 to 7, It is characterized in that include: The cloud-edge collaboration module consists of a central management platform, a big data repository, distributed cloud nodes, data generation terminals, and a cache cloud. The cloud-edge collaboration module performs large-scale overall data analysis, deep learning training, and autonomously generates commands for feedback data. An early warning module, which interacts with the cloud-edge collaboration module. The cloud-edge collaboration module issues instructions to control the early warning module to issue an early warning, issue an early warning to personnel, and expel the personnel from the shared database or the shared database; Image monitoring module, which is composed of a video monitoring platform, a video recording platform and an image recognition platform. The image monitoring module is equipped with cards at key locations, which can call up stored employee photos for confirmation and identify visitors through image scanning and comparison; A control module, which is arranged in the shared database and is composed of an entrance to the shared database and an alarm linkage control console. The control module can provide a complete security control function; An information matching auxiliary module, which is composed of an internal communication station and an information verification station. The information matching auxiliary module is used to assist the cloud-edge collaboration module in its operation and improve the functions of the cloud-edge collaboration module. An information calling module is used for calling information in a big data repository, and the information calling module enhances the calling efficiency of the cloud-edge collaboration module.