Research On Virtual Resource Management Technology In Mobile Cloud Computing

As an extension of cloud computing on the mobile Internet,mobile cloud computing is a combination of mobile Internet and cloud computing.It provides mobile users with services such as the storage and computing on the cloud platform,which greatly improves the performance of mobile applications.However,cloud computing and mobile Internet belong to different suppliers which brings a great challenge to the resource scheduling of mobile cloud computing during the process of resource management because of the lack of uniform deployment and management.Therefore,it has become an urgent problem that how to manage and schedule mobile cloud resources reasonably and improve the resource allocation efficiency while ensuring the quality of experience for mobile users.This thesis mainly studies the problem of load prediction and resource scheduling of mobile cloud computing.The main research work of the thesis is as follows:(1)According to the characteristics of the randomness and instability of the load sequence of host in the mobile cloud computing center,we analyze the load data of the host deeply,and propose a load prediction model of the host based on the improved wavelet support vector machine.The wavelet support vector machine has the characteristic that it can perform regression analysis on non-linear data,which can be used by the model to analyze and model the load sequences of the host.Particle swarm optimization algorithm is also used to select the optimal model parameters and the Google cloud computing data set is used to validate the effectiveness of the proposed algorithm.(2)Based on the deep analysis of application tasks in mobile cloud computing and the results of the load prediction of the host,a dynamic resource scheduling algorithm based on the Stackelberg game is presented.In order to improve the efficiency of resource scheduling in the cloud,a two-stage Stackelberg game model with multiple leaders and multiple followers is proposed.At the same time,the utility functions of leaders and followers in the model are designed.The proposed algorithm is verified by simulation experiment.Finally,a virtual resource scheduling and management system of mobile cloud computing is designed,including a load prediction module of the host and a dynamic resource scheduling module.The experimental results show that the models and algorithms designed in this thesis can predict the load according to the historical data of the task load and schedule the cloud resource dynamically in the simulation environment so as to improve the resource allocation efficiency,and shorten user waiting time in the mobile cloud computing.