SDN Based Task Offloading Optimization Strategy For Internet Of Vehicles

With the emergence of more and more on-board applications that can improve driving experience and driving safety,on-board computing devices can no longer rely on their own limited computing resources to complete the processing of these computation-intensive applications.Because vehicles can use on-board communication equipment implementation in high-speed mobile environment of wireless access,and car and car,car and people and cars and roadside infrastructure of the wireless communication link,so vehicles will be excessive computing tasks offloading to computing resources plenty of cloud computing center for processing is a feasible solution.However,the remote deployment of cloud computing centers makes the transmission of computing tasks need to go through the core network.Delaysensitive applications may lead to excessive delay and instability.Multi-Access Edge Computing provides users with required Computing services by deploying MEC platform on the network access side closer to users.Compared with cloud Computing,MEC eliminates the process of packet transmission in the core network.It has the advantages of lower delay,higher bandwidth and higher reliability.However,most of MEC are heterogeneous computing resources.In order to make better use of these heterogeneous edge computing resources,this paper uses the Software Defined Network to assist edge computing.The main research contents and results of this paper are as follows:(1)Firstly,model the problem by establishing the network model,task transmission and computation model,and reliability model.For the computation task generated by on-board application,Three task processing modes,namely Vehicle local computation,Vehicle to Vehicle multicast offload computation and Vehicle to Infrastructure offload computation,are jointly considered.The reliability expressions of vehicle task offloading and time delay calculation are given.(2)the second,in the complex and dynamic car networking environment,to ensure that the time delay sensitive car networking applications of high reliability,introducing and processing mechanism of car networking system,according to the problem model,considering the specific timing constraint,build tasks to maximize the reliability of the transmission and processing of optimization problems.(3)Finally,since the original problem is non-convex and NP-hard,the reliability fault tolerant algorithm based on particle swarm is adopted in this paper to obtain the global suboptimal solution with lower complexity.Finally,through simulation experiments,the influences of different offloading schemes,reprocessing mechanism,V2 V multicast communication and V2 I unicast communication on the system delay and reliability under the scenario of Internet of Vehicles are analyzed,Experimental results show that the proposed method improves the reliability performance of task offloading in Internet of Vehicles scenarios.