Research On Collaborative Resource Management Mechanism In Mobile Edge Cloud Environment

In the environment of mobile cloud computing,mobile devices can offload their computation-intensive tasks to the remote cloud computing data center through the Internet.The tasks will be executed by the resource-rich cloud computing data center and the results will be transferred back to mobile devices.However,these cloud computing data center are typically far from the mobile devices,and the way of migrating computational tasks to the remote cloud can increase the network load and lead to extra data transmission latency,which is a main problem that the traditional mobile cloud computing faced.To solve this problem,the framework of mobile edge computing has been proposed by researchers.By serving users' requests at the network edge,mobile edge computing(MEC)can reduce the service delay,improve energy consumption,and increase users' quality of experience.However,the limited computation capability and resource capacity at the edge clouds become the bottleneck of the performance of MEC.Cooperation with nearby MEC servers or resourceful remote cloud can enhance the performance of MEC.Thus,the way to propose an effective resource collaborative management mechanism to improve the efficiency and user experience of mobile edge networks is of great significance.This dissertation proposes two different collaborative scenarios of mobile edge computing:(1)the edge cloud servers cooperate with other edge servers in the same mobile edge network;(2)the edge cloud servers cooperate and compete with remote cloud servers.For these two different scenarios,an effective mechanism is proposed to optimize the revenue of both user devices and different servers,respectively.The main work of this dissertation includes these parts:(1)For the scenario that the edge cloud servers cooperate with other edge servers in the same MEC network,this dissertation considers all the edge servers in MEC network as peer nodes,and build up the system model based on a third-party broker.By modeling the interactions between multiple edge cloud servers by leveraging non-cooperative games,this dissertation proposes the optimal price strategy under the Nash equilibrium state,and theoretically analyzes the difference of the social welfare achieved by the proposed prediction-based price strategy and the optimal social welfare.Based on the optimal price strategy,this dissertation designs a distributed match algorithm for resource allocation.(2)For the scenario that the edge cloud servers cooperate and compete with remote cloud servers,this dissertation models interactions between multiple cloud servers,and analyze the optimal decisions for users,edge cloud servers,and remote cloud servers in turn by using backward induction method.(3)For the two scenarios above,this dissertation carries out two groups of numerical simulation experiments.By comparing the simulation results and the theoretical results,the system models and analytical conclusions have been validated.