Research On Low Latency And High Reliable Service Offloading Strategy In 5G Based On Edge Computing

With the development of information and communication technologies,users are increasingly demanding communications and business services.The current 4G network architecture is difficult to meet the requirements of users and applications.Therefore,5G networks have emerged as the times require.The goal of 5G networks is to support different services in the network,where low latency and high reliability services have attracted much attention from all walks of life.In 5G networks,cloud computing is considered to be one of the effective technologies which can efficiently process a large amount of business data and satisfy users' requirements.However,the cloud servers are often deployed far away from the terminal;And with the number of users increasing,the bandwidth of the cloud computing network will be seriously insufficient and the robustness of the cloud computing is poor,so the cloud computing network architecture is difficult to meet the user's requirements for low latency and high reliable service.In order to solve the shortcomings of cloud computing network architecture in supporting 5G services,this paper researches on mobile edge computing(MEC)technology that can make up the cloud computing,and focuses on analyzing the advantages and applications of mobile edge computing technology.Besides,the cloudedge computing architecture,the network architecture's service response latency modeling in the case of equipment failure and no failure,and the computing task offloading strategy in the corresponding model is studied to reduce the service response latency and enhance the service processing reliability.The specific research content of this paper is as follows:In the real-time mobile face recognition scenario,the traditional cloud computing architecture are facing the problem of high response latency in identifying service.To solve the problem,this paper presents a software-defined cloud-edge computing network architecture.In this architecture,Software Defined Network(SDN)technology is combined with MEC technology.The SDN technology can provide flexible and convenient centralized control and management for the network.Meanwhile,the MEC network can process the face recognition services at the edge of the network to reduce the latency.In the cloud-edge computing network architecture,this paper firstly models the service response latency of the face recognition service,and then considers the computing and communication capabilities of the MEC device comprehensively,and proposes the computing offloading strategy based on the fireworks algorithm in multiple MEC device clusters to achieve the goal of further reducing service response latency.The simulation results show that the software-defined cloud-edge computing network architecture based on fireworks algorithm strategy can effectively reduce the service response latency of the face recognition service.When 200 face recognition service requests are processed,the latency performance of the proposed architecture improved 77.8% and 84.2% respectively compared to the latency performance of the single MEC device and the traditional cloud computing architecture.For the problem that cloud computing architecture can hardly support low latency and high reliability services in industrial Internet of Things,this paper proposes a novel cloud-edge computing network architecture based on industrial Internet of Things.In this architecture,the cloud server is introduced into the cluster of multiple MEC devices as a computing node and forms a distributed cloud-edge computing network to process industrial Io T services in parallel.At the same time,this paper models the service response latency of the industrial Internet of Things service in distributed cloud-edge computing networks.And considering that the MEC devices or wireless links may be failure in the hostile environment,the service average response latency model in the failure case is established.In addition,the computing offloading strategy based on Real-Coded Genetic Algorithm for Constrained Optimization Problem(RCGA-CO)is proposed in this paper to optimize the service response delay when the equipment is fault-free.And based on the RCGA-CO algorithm,a reallocation and retransmission computing task offloading strategy is proposed in the failure case to improve the reliability of service processing and optimize the service average response latency.The simulation results show that when 500 Io T service requests are processed,the latency performance of the novel cloud-edge computing network architecture based on the RCGA-CO strategy improved by 84.8%,58.5% and 12.7% compared with the traditional cloud computing architecture,the single MEC device,and the MEC network.Moreover,the proposed reallocation and retransmission computing task offloading strategy could enhance the reliability of the service processing and optimize the service average response latency.