Design And Implementation Multi-level Flow Table Optimizer Based On OpenFlow Network

With the growth of the network size, the more and more complex of network architecture. Traditional networks cannot meet today's business growth rate and demand for flexible management of the network, and people urgently need a new architecture and innovative development of the next generation of Internet technology.Software defined network is put forward in this background. And the control and data plane are decoupled in Software defined network, which provide a new solution for researches of future Internet technologies. However, the flow table's scalability is poor and thus an OpenFlow switch can suffer from limited flow table size. This thesis studies on the optimization techniques of flow table.This thesis proposes a multi-level flow table optimizer based on OpenFlow network to solve the problem of storage efficiency and forwarding performance of OpenFlow vSwitch. The thesis mainly works on the following issues:1) Analyze the OpenFlow multi-level flow table forwarding technology and flow table features. At the same time, analyze the existing multi-level flow table optimization scheme to understand the shortcomings of the existing OpenFlow network multi-level flow table processing, laying the foundation for the design of multi-stage flow table optimizer.2) Optimize the storage space. The method of set-covering is used to reduce thesparse degree of the sub-flow table, so as to reduce the storage space needed to store the table entries.3) Optimize the forwarding performance. Based on the idea of the decomposition algorithm, different packet classification algorithms are designed for the fields with different matching types in the subflow table. The exact-match field is queried by an exact-match hashtable,and the prefix match field is matched by a segment hash. The ID value of each field bit mapping, and then cross-cascade bit string to establish rules hash table.4) The multi-level flow table optimizer scheme proposed in this thesis is experimentally verified. Based on the experimental scenario, we test the multi-level flow table optimizer, and compares the performances of the multi-level flow table optimizer and the unoptimized scheme in terms of the compression ratio and the forwarding delay.