Abstract：With the increasing size of OpenFlow flow tables，it is difficult for TCAM resources with limited capacity to meet the storage requirements of OpenFlow flow tables.This paper proposes a discriminative efficient storage algorithm for OpenFlow flow tables by combining TCAM with SRAM.This algorithm first classifies OpenFlow flows into a small number of packet-in-batch ones and a large number of packet-sparse ones，based on packet-in-batch characteristics.Then，we design an identification method of packet-in-batch flows and flow table entry replacement method，and dynamically store packet-in-batch and packet-sparse flow entries respectively in TCAM and SRAM，to solve the storage problem of OpenFlow flow tables and improve performance of flow table lookup.Finally，we evaluate the performance of our proposed discriminative flow table scheme in virtue of real network traffic traces by experiments.Experimental results indicate that our proposed algorithm performs better than currently popular discriminative algorithm of elephant/mouse flows in terms of TCAM hit rates，enhanced dynamic adaptability of OpenFlow flow table storage and can effectively improve the lookup performance of OpenFlow flow tables，thereby meeting large-scale flow table storage requirements.
2019, Vol. 40 
