A Dynamic Flow Table Management Method Based on Real-time Traffic Monitoring
Jinyuan Zhang,Xuanbo Huang,Jian Li,Kaiping Xue,Qibin Sun,Jun Lu
DOI: https://doi.org/10.1109/hpsr54439.2022.9831366
2022-01-01
Abstract:In Software-Defined Networking (SDN), the controllers implement flexible and scalability networking policies by installing different flow rules. Each rule matches a specific class of flows, instructs the switches to execute actions, and then expires when they finish their tasks. OpenFlow introduces the timeout mechanism to manage these flow rules. However, finding a reasonable timeout value becomes a difficult problem for the network managers. When a relatively small timeout value is given to an elephant flow, the rule expires early, introducing extra cost for the controller and long latency for the matching flow, respectively. On the contrary, a large timeout value for a mice flow makes a rule occupy the switch memory too long, wasting the caching memory and causing the flow table prone to overflow. Therefore, it is necessary to allocate appropriate timeouts for different flows dynamically. In this paper, we achieve this goal with real-time traffic monitoring and heuristic algorithms. By considering different network loads and designing corresponding dynamic timeout algorithms for different scenarios, we make full use of the advantages of SDN to improve the utilization rate of the switch memory and save the controller resources. Further, we implement our scheme in a simulation SDN platform and evaluate the algorithms with the public datasets. Experiments show that our scheme has low control overhead and is memory efficient compared with current mechanisms.