Research On Optimization Of Flow Table State And Rules In Software Defined Networks

Software Defined Networking(SDN)is one of the most representative technologies of the next generation networks,and attracts wide attention from both academic and industry since its appearance.As the core technique of SDN,OpenFlow separates the control plane from the data plane with the purpose of achieving flexible network control.As the control plane,the SDN controller is responsible for managing every unit in the network.On the other hand,the rest units are only able to deal with the data flows with whatever they are told to and achieve fast forwarding.This mechanism is available with the help of Flow Table—the core data structure of SDN.However,as a finely designed data structure with complex attributes,Flow Table is able to provide service for other purposes as well.According to the state-of-the-art researches,we investigate the corresponding prob-lems in both Flow Table storage and benefits.Firstly,we propose two ways to avoid the Flow Table Congestion Problem(FTCP)to improve the network robustness.Then we try to detect and infer the internal states of network function instances in Network Function Virtualization.The work of this paper can be concluded with the following three aspects:1 We discuss and analyze the Flow Table Congestion Problem in SDN,from its original to its disadvantage.It is a significant problem to reduce the number of flow entries needed in the almost full-filled flow tables,and at the same time,to retain the original QoS as much as possible.We propose a mechanism called"Flow Table Reduction Scheme"(FTRS)to efficiently solve FTCP and evaluate the performance of FTRS both via simulation and experiment.The results show that FTRS is able to reduce the number of flow entries by 98%at most of the size of flow table with no influence on network's normal functions.2 FTRS doesn't solve the FTCP perfectly due to the manual-involved feature and over idealized model.As a result,we improve the solution on the base of machine learning method—C4.5.The improved method is manual-free and selt-adapted and confirms its outstanding performance with test-bed experiments.3 The combination of Network Function Virtualization(NFV)and Software De-fined Networking(SDN)possesses a great potential in accommodating dynamic network control via cloning/migration of virtualized NFs and steering of traffic flows.A great challenge is the lack of the proprietary internal NF state infor-mation to the control system(including SDN controller and NFV orchestrator),which may lead to incorrect packet/flow processing at the newly created NF in-stances.In this work,we design a light-weight approach which can function ei-ther independently or as a plug-in to the network control system to reveal the internal NF states.Unlike the previous work,we propose to learn the internal NF states through normal network functions instead of designing extra APIs for certain NFs.Moreover,we propose a feasible way to detect state violations and even correct them automatically.Our approach is tested by experiments,and the results confirm its efficiency and practicability.