Analysis of the IP telephony security issues using automatic neural network classifier

The complexity and robustness of the currently used IP telephony infrastructures requires systematic monitoring of the network against security threats, where captured incidents are often used as a source of information for further research and analytical optimization of the monitored topologies. This article presents an analysis of IP telephony security issues which were clasiffied using neural networks and it also describes a novel design of the clasiffication itself. To capture individual incidents a distributed honeypot network with the autonomous probes has been developed too and the interconnection of the proposed mechanisms leads into a comprehensive system of frequent attacks detection. These security threats are currently used by attackers to penetrate an IP telephony network elements, and their effective detection can reduce a growing number of abused users. First two chapters describe used technologies and the current state of the art. The following part defines the architecture of the proposed system and the fourth passage is then formed by the analysis of the classified data results obtained. These are then evaluated in the final discussion and conclusion.