Clustering-based many-field packet classification in Software-Defined Networking

摘要: Abstract Packet classification is one of the main core functions networking. With advent Software-Defined Networking, packet has become more challenging by introducing many-field rulesets. In this paper, we propose an algorithm, Clustering-Based Classification (CBPC), which divides ruleset into some clusters using a new hybrid clustering method, based on innovative bit-level view. Those rules that have common wildcard and non-wildcard bit positions are put same cluster. Each cluster uses its to produce keys for hash table insertion query stages. This makes possible use tables without involving difficulties with inserting querying ternary vectors, because our algorithm converts it simple binary operations. fact, ignore portion information in key production tables. It overcomes problem extending wildcards all values. true lost but covered full matching at entries. We two versions CBPC, online offline. The version supports update, important requirement today's algorithms. proposed evaluated compared well-known state-of-the-art algorithms extensive simulations. results show Online-CBPC achieves 197% higher throughput 64% faster update than Tuple Space Search, OpenVSwitch standard while almost amount memory.