Packet classification using rule caching

Abstract

Data rates on Internet links keep increasing with the deployment of optical technology. Packets coming into highspeed networks need to be classified quickly. Different packet classification schemes have been developed but they require a number of memory accesses as classification is complex and memory is slow. We follow the approach of providing support with fast memory, as cache, in computer systems, to support packet classification schemes. Here, we propose a scheme based on memory cache to support packet classification. The scheme not only makes use of faster and smaller memories but also reduces the number of memory accesses to perform packet classification. It can make the performance of the adopted classification scheme independent of the number of connection flows. We present various packet-classification caching schemes for performing classification and provide the cache hit ratio results for various traffic models generated with Classbench.