Computation Offloading And Wireless Resource Management In Edge Cloud Systems

With the advent of the era of Internet of things,the number of smart terminal equipment shows exponential growth.With more and more data intensive computing applications or services are extended to mobile terminals,the demands of mobile users for network services are growing.At the historic moment,the concept of mobile cloud computing is arisen,which fully combines the advantages of cloud computing technology and mobile Internet.Through the wireless network,the terminal device can offload(migrate)computation or other applications requirements that are difficult to handle to remote Cloud in the mobile cloud computing system.Then,the remote Cloud utilizes its virtualized unlimited computing and storage resources to provide appropriate services for terminals.However,in the case of facing instantaneous massive service requests,the mobile cloud computing system may generate uncertain high latency,which would greatly reduce the users' satisfaction.Taking the differences of available resources in smart terminal devices into consideration this thesis mainly studies the selection of collaborators or offloading computation collaboration group in the mobile cloud computing,providing edge-computing services with low latency and low energy consumption for those smart terminal equipment who own poor-resources.Main content of this thesis can be summarized in the following:1)Based on the mobile cloud computing system,the thesis considers that fog computing has the characteristics of providing real-time application processing services,combining with the enhanced D2D communication to reduce the load pressure and network coverage.The D2D communications refer to direct communications between users with adjacent communication distance.Then,the thesis designs an efficient computation offloading mechanism based on D2D communication between user nodes,and proposes an iterative computation offloading algorithm to optimize the joint computation offloading and power allocation problem.Furthermore,a matching based suboptimal computation offloading algorithm is also proposed to reduce the computation complexity.The simulation results show that the proposed scheme can reduce the energy consumption of the whole system while meeting the personalized delay needs of the terminal equipment.2)Based on fog computing system,this thesis constructs efficient user group and takes advantage of NOMA technology in improving the spectrum efficiency and supporting the simultaneous transmission of multi-users.In addition,the beamforming technology is also utilized to enhance the spatial diversity gain.Then,the thesis designs a cooperative computation offloading scheme as well as an efficient user pairing and resource allocation algorithm.The simulation results show that the proposed scheme can not only make more efficient use of local resources but also can improve the energy efficiency of fog computing system.