Research On Efficient Transmission Technology Of Data Exchanging Network Based On SDN

With rapid development in recent years, internet of vehicles has entered its golden age. Such development will inevitably generate data at an overwhelming scale and demand much higher transmission efficiency of data exchanging network. However, for its inherent constraints, traditional network can hardly meet the requirements of efficient data exchanging of internet of vehicles, such as transmission efficiency, network resource utilization rate, and network scalability and innovation. As a result, the adoption of next generation network as the data exchanging network of Internet of vehicles is rationally proposed.SDN, Software Defined Network, is a model of the next generation network. The SDN network and technologies have become more and more mature after years of development and standardization. OpenFlow protocol, proposed by Open Networking Foundation, transforms traditional vertical closed traditional network into one where control plane and data plane are separate. Such innovative process makes the OpenFlow network much advantageous in transmission efficiency, network resource utilization rate, and network scalability and innovation, compared with traditional network. If adopted as data exchanging network, the OpenFlow model will better guarantee efficient data transmission of internet of vehicles.Accordingly, this dissertation proposes an OpenFlow based Qo S guarantee subsystem. With control plane and data plane separated in the OpenFlow network which senses and controls data plane at its control plane, the subsystem identifies specific flows which will be proccessed to find its optimal Qo S path according to the realtime network status, such as link delay, available bandwidth and packet loss rate, collected from lower data plane. The subsystem uses Floodlight controller as development platform and adds system modules therein to build a subsystem, which consists of five major modules: network status detecting module, flow recognition module, route calculation module, Qo S management module and topology management module. The network status detecting module detects network status and informs the topology management of any changes. The topology manages network topology and provides other modules with a comprehensive topology status view with network status. The flow recognition module classifies data by using SVM(Support Vector Machine) algorithm. Route calculation module calculates a route satisfies with QoS rules by using Tabu search based delay-constrained and minimun cost routing algorithm.The Qo S guarantee subsystem presented in this dissertation has examined on Mininet simulation platform, the results show that the system is capable of collecting network status accurately and allocating network resource efficiently.