Research On The Principle Of Data Center Service Resource Scheduling In Edge Computing

With the development of edge computing technology,more and more applications in terminal devices can be offloaded to the data center which has characteristics of high distribution at the edge of the network for service and computation.The edge data center is distributed at the edge of the network.Compared with the traditional cloud data center,the edge data center is more close to users and terminal devices in the geographic location,which can reduce the time delay of application service requests and bring higher quality of service to users.However,the edge data center is affected by many aspects such as geographical location,infrastructure and so on,it is less than the cloud data center on the scale of construction and the physical equipment.The limited resource capacity limits the computing and storage capacity of the edge data center,and the edge data center can not completely replace the cloud data center.An optimal method is to combine the advantages of the Cloud and Edge to provide services together.In the interconnection of all scenarios,many new applications emerged.These applications are widely used and have stringent requirements on delay.The combination of edge computing and cloud computing provides a new service model for users,but it will face many challenges when computing massive time sensitive applications.The first problem is how to provide the effective deployment and management in data center to improve the utilization rate of resources,reduce the resource cost and realize the load balance.The second problem is task offloading,offloading tasks to edge or cloud server can effectively improve the execution rate,reduce the energy consumption of user equipment,application.How to choice the place for running applications and how to tradeoff the energy consumption and delay in the system,these problems need to be studied and resolved.Aiming at the problem of resource deployment in edge computing,we Analyse of multiple factors affecting the deployment of resources in a multi-time-sensitive-application scenario.We Abstract resources of different applications into different virtual machines for deployment and management in combination with virtual technology,and then we formulate a MILP(Mixed-integer linear programming)model to solve the problem of resource deployment with the goal of minimize the amount of resources.Finally,the optimal solution of resource deployment strategy is given by using IBM WebSphere ILOG CPLEX tool.According to the problem of task offloading in edge computing,this paper give a detailed analysis in components of energy consumption and delay when applications are offloaded to cloud and edge server,or running locally.Through the modeling and analysis of the problem shows that the problem is a NP-Hard problem.And a heuristic algorithm of minimum delay and energy consumption in Local-Edge-Cloud system(LEC-MINDE)is presented to solve the task offloading problem.We normalize delay and energy consumption to system cost and use the tradeoff factor to represent the proportion of each part.The result shows that the algorithm is an approximate optimal algorithm,which can effectively reduce the system cost.