Research And Design Of Optimizating Strategy For Flowtabl In SDN Switch

With the flourishing development of Internet Technology, our life and work has been changed profoundly. The networking service and bussiness requirements is increasing continuously, that lead to wider networking bandwidth and newer networking architecture are needed. Recently, SDN stand out from plenty of new networking architectures, and obtain consistent affirmation from both academic field and industrial circles.However, confined by the processing power of SDN controller and the storage resource of SDN switch, packet loss and network delay are still serious, many researching institute and companies are taking effort to research the optimizing technology of flowtable. This thesis will describe two kinds of optimizing strategy as follows:(1) Dynamic adaptive timeout algorithm. Timeout decides the living time of entry, and is the initiative evicting method in flowtable. Dynamic adaptive approach firstly predicts required information during next samping period by utilizing the resource in flowtable and flow traffic in current sampling moment. Then modify the value of timeout to guarantee that the rest of entries in flowtable can accommodate enough new arriving flows.(2) Optimizing strategy based on flowtable. There are two different strategies based on flowtable. One is to optimize multilevel flowtable based on FPGA, add a table_manager module in flowtable, increase the expandability and flexibility of multilevel flowtable. The other is to incise and classify the flowtable, devide flow entries into active entries and expired entries. It can improve the matching success rate, and decrease the cost of using LRU evicting method by adding extra cache.