SAPPM: Self-adaptive Probabilistic Packet Marking for IP Traceback

摘要

IP traceback is an efficient technology to defend the Internet against Denial of Service (DoS) intrusion. Probabilistic packet marking is considered to bo a promising approach to realize IP traceback for floodbased DoS. In this paper, we propose a novel and practical IP traceback approach with self-adaptive probabilistic packet marking called SAPPM, which can conduct path reconstruction with fewer packets and introduce less impact to the performance of network and routers through the traceback process. In particular, SAPPM includes two stages. In the first, stage, it adopts a flexible marking scheme according to the router location and packet unmarked hop length, which can reconstruct the complete path with fewer packets. In the second stage, marking rate is adjusted to the change of some key indicators of routers, which can make routers avoid becoming overloaded. In the end, the feasibility and effectiveness are evaluated by simulations.