Research On Mobile Device Transfer Strategy In Mobile Grid Computing

Mobile grid computing is the integration of grid computing and pervasive computing technology, the key technologies of which is wireless mobile grid. At present most of the grid structure and algorithm does not take into account the mobile computing environment, because mobile devices have not been seriously considered as a legitimate grid computing resources or interfaces. There are two types of device role in the grid computing environment, the one is a interface which could communicate with grid system so the users can require grid service through the mobile device, complete the task using grid resources, remote monitoring the tasks and get the required results from the grid. The other role is look on mobile devices as the grid computing resources, which not just a recipient of grid services but a participant in grid computing tasks. Therefore, the mobile device into the grid effectively (either as a recipient of grid service also be a grid service provider) which achieve efficient resource sharing with these different functions of mobile devices and access the static grid resources seamlessly become the main research hotspot in recent years.The main research and innovation about this topic is reflected in the following areas:1) Trigger Switch algorithm.Choosing the two typical wireless communication signal flag (signal strength measure and signal interference measure) as the threshold to determine device switching time through analyzing the performance of existing mobile devices and real wireless network, which can achieve active and transparent registration and transfer tasks for the user.2) target selection algorithm.Studying the performance of several selected grid optimal prediction theories by analyzing the existing wireless grid network, getting the better effect for the grid resources through the gray relational analysis theory and computing the grid performance parameters in order to compare the result with the simulation results.3) the seamless transfer algorithm for grid task.By analyzing the characteristics of the grid task, we know that the submitted grid tasks to be divided into some child tasks, and then scheduled on the grid resources by different grid task. Through analyzing the wireless transfer mechanism of the pervasive computing, we get the conclusion that the size of the grid task granularity could be changed and integrating the wireless network and grid computing to achieve uninterruptible processing for the grid tasks.4) the simulation for mobile grid.Through analyzing the traditional grid simulation platform, we know current grid simulation platform always based on a static grid computing, which could not supported the mobile device moving well. This topic main talk about the connectivity layer of grid computing, so choosing the typical network simulation tool NS. Of course making the appropriate changes for some part of the platform is necessary.This thesis is supported by National Natural Science Foundation of China (NSF)under grant (No.60773211, No.60970064),Natural Science Foundation of Hubei Province under Grant (No.2008CDB335),Program for New Century Excellent Talents in University,China under grant (NCET-08-0806),Open Fund of the State Key Laboratory of Software Development Enviroment (No:SKLSDE-2009KF-2-02),Fok Ying-Tong Education Foundation for Young Teachers in Higher Education Institutions of China (No:121067),NSF of Wuhan Municipality (No:201010621207).