An IP Telephony Security System based on Methods of Intelligent Traffic Analysis
Andriy Melnyk, Ruslan Shevchuk, Ihor Romanets, Serhiy Voznyak, Oleh Kvasnytsia, Oleksandr B. Yashchyk · 2025
The paper proposes an approach to protecting IP telephony infrastructure from common attacks at the signaling and media traffic levels. The developed system integrates protocol-oriented analysis methods (SIP/SDP/RTP) with modern machine learning algorithms and knowledge-based rules. It enables the detection of both known violations (mass registration, malicious calls, identity spoofing) and new or hidden threats. The system architecture is based on a modular principle, comprising components for data collection, feature extraction, intelligent analysis, and automated response. Experimental results on test datasets confirm high accuracy and explainability of the decisions. The presented approach is of practical significance for telecommunications operators and corporate networks, as it combines performance, flexibility, and compliance with VoIP standards.