Task Allocation And Computing Offloading In Mobile Edge Computing

With the gradual commercialization of 5G technology,smart devices installed with various novel applications have become indispensable to us.Applications like VR/AR typically utilize cameras and local sensors to perform real-time perception-based operations.these applications have two main characteristics;first,smart devices must perform high-throughput processing of the data streams generated by sensors.Second,the way of extracting useful information from data stream usually needs support from computing intensive service.Due to the above two characteristics and the limited computing resources of intelligent devices,it is difficult to run high computing applications independently in intelligent devices.Therefore,the above problems can be solved by computing offload ing and task allocation technology.Computing offloading generally refers to the migration of high computing tasks to the server with rich computing resources and the task processing is performed by the server.The results of the execution are transmitted back to the user equipment through various communication network technologies.The process is completed in a very short time which meets the requirements of real-time operating system and low delay applications.Task scheduling technology is to split the computing task and put the split task on the server suitable for the task,so as to reduce the overall task execution delay and user energy consumption.In this paper,based on the above two cases,the following work was done:(1)The following work has been done for computational offloading in edge environments.Firstly,the SDSC model is established by using the idea of software defined network and service composition.In this model,the controller is deployed at the edge of the network and is services composed through centralized management.Secondly,low latency service composition is defined as SCO constraint satisfaction problem to meet users' special service requirements.Thirdly,SDSC model takes advantage of SDN's global view of the whole network.Finally,the Qo S constraint function is introduced in MEC environment to design service discovery and task offloading mechanism,so that users can get lower latency and higher quality of service.(2)The following work has been d one for task allocation in edge environments.Firstly,cloud services are introduced in the edge environment.The three-tier network architecture has user layer,edge layer,and cloud layer.Secondly,the application is defined as a partitionable DAG model,where the common nodes in the DAG have low computational resource requirements and the critical nodes have high computational resource requirements.Thirdly,the task and computing resources are constrained by task joint constraints.the resource in resource allocation is set to a fixed value to solve the task allocation problem.because only task allocation is considered.Finally,a heuristic task allocation MECTA algorithm is proposed to solve the user task allocation problem.The heuristic algorithm is used to model the task assignment.Three classic DAG models are simulated and analyzed in MATLAB.The results show that compared with other algorithms,this algorithm can find the best node of task allocation faster and red uce the user task execution energy consumption.