Network Protocols Determination Based on Raw Data Analysis for Security Assesment under Uncertainty

The paper is devoted to issues of the network traffic analysis in conditions of uncertain network protocol specifications. We propose an approach to identify typical structures of network protocols and determine their lexical specifications based on text-inspired methods for the structural analysis of raw data. High heterogeneity, partial lexical uncertainty and use of new, proprietary or modified data transfer protocols in computer networks explain high relevance of the research topic. We present the technique of network traffic analysis and the results of experiments, that confirm the applicability of the proposed approach.