基于TCAM细粒度多级流表存储策略

SDN (Software Defined Networking) is an architecture and technology based on software, its biggest feature is the control plane and date plane of a loosely coupled, supports network state control centralized, to achieve the underlying network infrastructure transparent to upper application. The author focuses on multi-level flow table matching mechanism based on OpenFlow v1.1 protocol. In view of the possible resource shortage in the existing SDN technology, this paper describes a Fine-grained OpenFlow multiple-table pipeline architecture that efficiently stores the flow table in the TCAM memories. The IP address and MAC address for further fine-grained division. we add the single flow table identifier on the multi-level table items, Simulations show that the strategy reduces the TCAM usage for the multiple-table mapping scheme by 14.3%, which is important for scalable designs of OpenFlow data planes.