Efficient Packet Classification with Learned Bloom Filter in Software-Defined Networking

Abstract

The development of emerging network technologies represented by Software-Defined Networking (SDN) has made traditional routing technologies that are based on a single IP address domain unable to meet the increasing demand for network services and security. Packet classification is a technique to differentiate multi-domain network traffic in a fine-grained manner using packet header fields. Packet classification requires to operate efficiently to avoid it becoming a bottleneck in the packet routing process. In this paper, we propose a learned Bloom filter (LBF)-based packet classification algorithm that combines the RNN learned model with a support Bloom filter (SBF) to improve the classification accuracy. Specifically, the learned model is trained with the positive and negative sets, which is used as a pre-filter to identify the two sets. For the filter outcomes with a negative result from the learned model, SBF is constructed to perform the second filtration. To ensure efficiency of RNN and SBF, we carefully select key features to be used in RNN and SBF, which can also maintain efficient search in the final stage of rule matching. We perform simulation with various datasets showing the performance advantages of our proposed algorithm over existing solutions in terms of memory usage and classification accuracy.