Harvesting unique characteristics in packet sequences for effective application classification

摘要: Network traffic classification is critical to both network management and security. Identifying application at the flow level with signature matching has been widely used as most efficient method due its reliability robustness. However, increasing number of applications their frequent updates, we have constantly regenerate signatures, which resource intensive time consuming. To address this issue, propose explore unique characteristics in packet sequences discovered two types sequence signatures. We introduce our design implementation an automated packet-sequence construction (APSC) system, based on association rule mining data clustering technologies. This system can not only automatically generate traditional signatures from individual payloads but also construct new or features sequences, even for encrypted flows. best knowledge, first practical that supports construction. Our experimental results show proposed high quality a variety limited overhead.