TCAM-Based Packet Classification Using Multi-stage Scheme

Abstract

As the number of network services increases, the scale and complexity of network also arise. Software-Defined Networking (SDN) is a new network architecture during the past few years. OpenFlow is a celebrated protocol for SDN. It has two characteristics we mostly concerned. First, it provides many match fields to fulfill different network strategies. Second, rule table partition is allowable. However, these two characteristics may result in considerable memory space and prolonged lookup time. In this paper, we propose a TCAM-based packet classification method to mitigate the problem mentioned above. Our method includes multiple stages, where only some fields are compared in a stage. We also propose a refinement to further reduce memory accesses. Our scheme achieves different levels of improvements for different classifier as the experimental results show.