ARCHITECTURE OF AN INTELLIGENT SYSTEM FOR CONTROLLING ATYPICAL BEHAVIOR OF IP-TELEPHONY USERS IN A CORPORATE NETWORK

Abstract

The article presents the architecture of an intelligent system for controlling atypical behavior of IP-telephony users in a corporate network. The development of such systems is relevant due to the growing number of cyberattacks and internal threats that can lead to serious breaches of security and confidentiality of corporate data. To analyze VoIP content, it is necessary to formalize the method of constructing a basic message in IP telephony. To do this, it is necessary to first convert the VoIP media message into a text representation and thus form a database of already text messages. The proposed system uses machine learning and data analysis methods to detect anomalies in user behavior, which allows for a prompt response to potential threats. The developed AISES-VoIP (Architecture of an Intelligent System for Ensuring the Security of VoIP) system, in comparison with existing solutions, has a powerful subsystem for analytics and reporting with the possibility of intelligent data analysis and the availability of means for adaptive integration into external specialized information systems and the possibility of integrating specialized software modules. The main components of the architecture are described, such as modules for data collection and processing, behavioral analysis, and decision making. Particular attention is paid to the application of artificial intelligence and integration of the system with existing infrastructures, as well as ensuring its scalability and reliability. On the basis of the research, a detailed architecture scheme of the intelligent system is developed. The implementation of natural language processing methods is a key element of the intellectualized system for analyzing atypical behavior of IR telephony users. This allows you to create a more effective system for detecting anomalous traffic and potentially dangerous communications.