Intelligent auditing decision tree generation method of 4A management platform

Abstract

The invention relates to an intelligent audit decision tree generation method in a 4A management platform. The technical points of the method are that according to audit material that is input by a user, an audit decision tree is generated automatically by a machine learning module and sent to an audit strategy application module for the 4A management platform to carry out practical audit application; the 4A stands for unified account management, unified authentication management, unified authorization management and unified security audit. The method has the advantages that: an audit strategy is generated automatically; rule entering is not required; the strategy has high traversing performance by being stored in a tree structure; algorithms have a plurality of choices; and man-made strategy modification is supported.

Description

technical field [0001] The invention relates to a method for generating an intelligent audit decision tree in a 4A management platform, which is applicable to the telecommunications industry and the unified security audit in the 4A management platform (4A is unified user account management, unified authentication management, unified authorization management, and unified security audit). . Background technique [0002] With the development of business networks of major telecom operators, the number of internal users continues to increase, the scale of the network expands rapidly, and security issues continue to appear. However, each business network system maintains a set of user information data, manages the accounts and passwords in the system, and audits the operator's operation behavior in the system in the form of logs in isolation, which is far from meeting the needs of its own business development. International business needs. [0003] The proposal of the 4A solutio...

AI Technical Summary

Problems solved by technology

[0010] The audit object of the unified security audit subsystem of the 4A management platform is syslog, snmp, snmp trap, ftp, webservice, ODBC / JDBC and other logs from authentication, authorization, network audit and comprehensive platform, and its data volume can reach at least 5000 per second Items - 7000 items. For the processing of such massive data, most software products currently implement complex audit policy sets through customization and manually enter the system for application. This not only requires a lot of work, but also the actual effect is not good: sequential traversal The performance of the policy set is unacceptable for the audit system used to process massive data, the complex audit logic structure is messy, the versatility is poor, and the portability is poor, etc.

[0012] (1) The generation of 4A comprehensive audit strategy has a large workload, and because the whole process relies too much on people, there are great security risks

[0013] (2) It is not easy to enter complex logic into the audit policy set

[0014] (3) The logical structure of the audit strategy realized by customization is relatively messy, the versatility is poor, and the audit strategy is not easy to optimize

[0015] (4) The traversal efficiency of policy traversal stored in the form of rule sets is low, and it is not suitable for audit systems that process massive data

[0016] (5) It is difficult to establish a common audit strategy template for different business systems, which makes the portability of the 4A comprehensive audit system worse

[0017] (6) If the business system changes, the audit strategy will not be easy to optimize

Images