Research Of Mobile Grid Task Scheduling Based On Resource Reliability

Mobile grid is due to the rapid development of grid computing and mobile computing technology. It is a full inheritor of traditional grid that can provide various reliable grid servers for grid users. Meanwhile, mobile grid can support the mobile devices, which can be treated as grid interface or grid resource. Mobile devices have some common characteristics, such as mobility, limited energy and intermittent network connections and so on, which introduce some problems into mobile grids and especially make the grid resource management to be more complex. The task scheduling that is the core of resource management faces great challenges.The grid task scheduling is actually the process of mapping many tasks to many resources, which should not only satisfy the users, but also ensure the resource utilization. It seems to be a critical and difficult issue to select a good resource for the task in the mobile grid. Because the resources in mobile grid are mostly the mobile devices, they are limited in many aspects. It becomes an important issue that how to weigh the resource reliability. At present, the researches of mobile grid task scheduling mainly focus on mobility, energy consumption and intermittent network connections. The existing scheduling algorithms mostly pay attention to one aspect of resource. In this thesis, we define the resource reliability and propose the mobile grid task scheduling algorithm based on resource reliability, which also involves the user mobility.The leading research and innovation of this thesis are as follow:Firstly, it analyzes the characteristics of mobile grid resource management and the challenging issue in task scheduling. A simple node mobile model is proposed based on the rule of node mobility, which can be used to decide where the user will go. Then the nearby resources will be selected to process the task in order to reducing the network delay.Secondly, it provides a method to compute the disconnection time, so the task completion time can be identified. According to the task implementation, the energy consumption is calculated. The resource reliability is in relevance to both task completion time and resource energy consumption. In this thesis, the tasks are divided into urgent task and normal task, which adopt different methods to calculate the resource reliability.Thirdly, this thesis proposes a mobile grid task scheduling algorithm based on the resource reliability. The proposed algorithm involves both user mobility and resource reliability. It firstly determines the grid area of user according to the mobility model, and then selects the most reliable resource in the user grid for tasks. The algorithm aims to ensure that the tasks will be process successfully.Lastly, the proposed algorithm is simulated in mobile grid environment and is compared with an existing scheduling algorithm. The simulation suggests that the proposed algorithm has good performance and is suitable for the mobile grid task scheduling.