A Decomposition-Based Approach for Scalable Many-Field Packet Classification on Multi-core Processors

摘要: As a kernel function in network routers, packet classification requires the incoming headers to be checked against set of predefined rules. There are two trends for classification: (1) examine large number header fields, and (2) use software-based solutions on multi-core general purpose processors virtual machines. Although has been widely studied, most existing systems target classic 5-field classification; it is not easy scale up their performance with respect fields. In this work, we present decomposition-based approach; supports rule sets consisting our approach, range-tree hashing used search fields input parallel. The partial results from all represented ID sets; they merged efficiently produce final match result. We implement approach evaluate its overall throughput processing latency size varying 1 32 K. Experimental state-of-the-art 16-core platforms show that, an 48 million packets per second 2,000 ns can achieved K set.