An Efficient Conflict Detection Algorithm for Packet Classification

Abstract

Packet classification is essential for supporting advanced network services such as firewall, quality-of-service (QoS), and policy-based routing. The rules used by routers to classify packets are called packet filters. If two or more filters overlap, a conflict occurs and leads to ambiguity in packet classification. In this paper, we propose an algorithm which can efficiently detect and resolve filter conflicts by using tuple based search. The time complexity of the proposed algorithm is O(nW+S), and the space complexity is O(nW), where n is the number of filters, W is the number of bits in a header field, and S is the number of conflicts. We use the synthetic filter databases generated by ClassBench to evaluate the proposed algorithm. Simulation results show that the proposed algorithm can achieve better performance than the plane-sweep algorithm both in time and space, particularly for databases with larger numbers of conflicts.