Joint Route Selection And Network Function Placement For Software-Defined Networking

The integration of software-defined networking(SDN)and network function virtualization(NFV)is expected to achieve the convenient and flexible deployment of various sophisticated network functions,and support user applications with guaranteed quality of service(QoS).To meet the application requirements of users,and improve the overall performance of user transmission and network,it is necessary to design a routing strategy for end-to-end transmission of user flows in SDN and implement effective placement and management of network functions.In addition,there may be resource competition problem when multiple user flows are transmitted in SDN,and efficient resource management and allocation are required.This thesis mainly studies joint route selection and network function placement algorithm in SDN,and the details as follows:Firstly,the concepts and related architectures of SDN and NFV are introduced,and an overview on the existing route selection algorithms,network function placement algorithms and joint route selection and network function placement algorithms is presented.For multiple user flows transmission scenario in SDN,joint route selection,network function placement and resource sharing problem is studied.Considering the factors of network resource status,user flow service demand and network resource sharing,the joint optimization problem is modeled as the total end-to-end delay minimization problem.As the optimization problem is an Non-deterministic Polynomial-hard(NP-hard)problem,which cannot be solved directly,it is transformed into three subproblems,i.e.,route selection subproblem,network function placement subproblem and computing resource sharing subproblem.The three subproblems are solved by applying the K-shortest paths algorithm,Kuhn-Munkres(K-M)algorithm and Lagrangian dual method,respectively,and the effectiveness of the proposed algorithm is verified by Matlab simulation.Aiming at the requirements of user flows in network function execution order and traffic changes during user flows transmission in SDN,a joint route selection and network function placement algorithm is proposed.The joint route selection and network function placement problem is modeled as the problem of minimizing total network energy consumption.For the two transmission scenarios with user flows having priority and non-differentiated user flows,the thesis proposes a multi-stage graph based Dijkstra algorithm to optimize the network function placement strategy,and then multiple flows route planning and network resource competition problems are solved by applying the heuristic algorithm to optimize the network total energy consumption.The simulation results show that the proposed algorithm can improve network performance and achieve efficient transmission of user flows.