Analysis and detection on long-time quasi-periodic communication behavior of network applications

Internet applications with long-time quasi-periodic (LTQP) communication behavior are always sensitive and critical in traffic monitoring, safety detection, user billing etc. Analysis and detection on such behavior can provide help for the identification of “always online” applications, which is of great significance. This paper first introduces the concept of LTQP communication behavior, then study the inter packet time (IPT) distribution and typical packet group pattern of LTQP communication traffic in detail, finally constructs a cascading structure identification method to stepwise detect LTQP communication behavior, and verifies the performance by experiment.