Research On Mobility-aware Task Offloading Algorithm And Protocol For Mobile Edge Computing

Mobile Edge Computing(MEC)has emerged as a promising paradigm to provide pervasive computing and storage services for mobile and big data applications.The mobile edge network(MEN)for computing services has been established due to the deployment of small base stations at the edge of the network.Mobile users are usually able to connect with these small base stations directly and receive fast feedback and short delay services from the MEN.Therefore,mobile users can upload some computation-intensive and delay-sensitive tasks to the currently connected small base station.Then,the mobile edge computing network utilizes hardware resources of small base stations to assist users in handling such tasks.The decision to perform task offloading in this new computing paradigm faces many new challenges such as complex task requirements,high user mobility,diverse applications and services,and limited computing and storage resources of small base stations.Consequently,how to exploit and use new features of the MEN,how to improve the decisions for task assignment which involved in the task offloading process,and how to improve the transmission efficiency are all topics and directions worthy of in-depth research.In this thesis,we carefully study the above-mentioned problems.Specifically,we sort out new features and challenges that emerged from MEC,summarize the state-of-the-art relevant models and related works,analyze and discuss the respective advantages and disadvantages of optimization models,mobility models and wireless bulk data transmission mechanisms and protocols in MEC.Based on this,this thesis mainly designs and evaluates the following three aspects in MEC:(1)a mobile user characteristics sensing service framework and a proactive classification mechanism for offloading tasks,(2)a mobilityaware task offloading algorithm,and(3)a highly-efficient bulk data transfer protocol.Finally,this thesis conducts various experiments to evaluate the performance for all works above in simulation or testbed environment and analyzes the experiment results.According to the analysis,the comparison results show that the work of this thesis can successfully utilize the sensor data to proactively classify the offloading tasks,and obviously reduce the delay of task offloading while maintain a high acceptance rate for offloaded tasks in MEC.Meanwhile,the performance results of the high-efficient bulk data transfer mechanism obtained from both testbed and simulation experiments demonstrate that,compared to the state-of-the-art protocols,the high-efficient bulk data transfer mechanism can greatly enhance the dissemination performance by reducing the dissemination delay by 34.8%,which helps improve the overall efficiency in MEC.