Many‐field packet classification with decomposition and reinforcement learning

Scalable packet classification is a key requirement to support scalable network applications like firewalls, intrusion detection, and differentiated services. With ever increasing in the line-rate core networks, it becomes great challenge design solution using hand-tuned heuristics approaches. In this paper, we present learning-based engine by building an efficient data structure for different ruleset with many fields. Our method consists of decomposition fields into subsets separate decision trees on those deep reinforcement learning procedure. To decompose given ruleset, consider grouping metrics standard deviation individual introduce novel metric called diversity index (DI). We examine schemes construct each scheme compare results. The results show that SD 11.5% faster than DI metrics, 25% random 2 40% 1. Furthermore, our selection can be applied varying rulesets due its independence.