Research On OpenFlow Scheduling Strategy In Software Defined Network

The emergence of Software Defined Network(SDN)provides a solution for the bottleneck of traditional network.The biggest difference between them is that SDN decouples control plane and data plane.SDN connects applications with network devices by SDN protocols.The controller,a logically centralized entity,has a good command of the whole network and makes OpenFlow routing flexible and programmable.Open Flow,as the standard protocol of SDN,transmits information through flow tables.Each flow table is composed of lots of flow entries,which correspond to the flows in the network.OpenFlow network has a large number of flows,so it is very important to schedule flows reasonably and efficiently and provide excellent service for network users.When designing OpenFlow scheduling algorithm,the fair principles should be concerned.Utility Function is a comprehensive function and is used in this paper's subsequent studies.It is found that OpenFlow scheduling in data plane mainly studies the scheduling strategy of congestion in the network.The pursuit of excellent user experience drives the need to take users' demands into the consideration.Therefore,in this paper,the user center network(UCN)is added to the original architecture of SDN,which is used to extract the parameters with user attributes.In the conventional resource allocation,to meet the needs of users,the resources will be used as much as possible.When there are new users,there will be little resource left,which seriously affects the satisfaction of the entire network.Because network users are in the state of competitive,game theory is introduced as the scheduling algorithm.It is divided into two kinds.In non-cooperative game framework,the nash equilibrium point exits and it can be proved to be pareto-optimal in some condition.In cooperative game framework,the optimization problem of system utility function is solved.Finally,the distributed algorithm suitable for OpenFlow network is obtained.The simulation proves the stability of the algorithm and shows that the cooperative gaming scheduling scheme has higher user group satisfaction degree.