AntiWorm NPU-based Parallel Bloom Filters for TCP/IP Content Processing in Giga-Ethernet LAN

TCP/IP protocol suite carries most application data in Internet. TCP flow retrieval has more security meanings than the IP packet payload. Hence, monitoring the TCP flow has more strength than only monitoring the IP packet payload in the AntiWorm system. The main idea of this paper is to use the flexibility and high performance of Network Processors to scan TCP flow for locating worm’s binary codes, and cut off their propagation. A stateful TCP flow inspection engine is implemented based on IXP Network Processor, which can monitor about 512K flows. The performance issues about IXP Network Processors are evaluated and collected, and an analysis is made for further optimizing the system performance. The system is also demonstrated and proved by using the Internet traces and real assaults of Worms. Software Package TCPScanner 1.0 is also given as a software release of the research.