Network security: Traffic analysis and countermeasures

As long as information is transmitted through wires or broadcasted in wireless media, shared by different users, network communication cannot be considered completely secure.; Traditionally encryption and encryption-based protocols are used to provide some level of network security. The straight-forward approach to break network security is to crack an encryption schema. Directly cracking modern encryption methods is challenging and in most cases infeasible. This approach is very well studied today. However, encryption alone is not sufficient for secure communication. An adversary can gain information about participants and their communications by observing their traffic.; Traffic analysis of multimedia transmissions is a new area of research in network security. In this dissertation a case of scene detection traffic analysis for live video transmissions is considered. Experimenting with MPEG-4 and libdv based live video transmissions we get network traffic patterns for transmissions of high and low active scenes. Statistical analysis shows that high and low active scene transmissions have different packet size distributions, which can be recognized.; We present statistical traffic analysis scene detection technique. Scene detection was tested on MPEG-4 and libdv coded video streams. Finally we provide an efficient adaptive Low Active Scene camouflaging algorithm, which makes low and high active scene transmissions undistinguishable by statistical traffic analysis.