Research On Grid Reource Locating And Task Scheduling

Grid computing, emerging as a new paradigm for next-generation computing, enables the aggregation of geographically distributed heterogeneous resources for solving large-scale problems in science, engineering, and commerce. The resources in the Grid are heterogeneous and geographically distributed. Availability, usage and cost policies vary depending on the particular user, time, priorities and goals. The management of resources and application scheduling in such a large-scale distributed environment is a complex task.Due to the existing problem and default of resource location and task scheduling in Grid resource management,the dissertation apply data clustering theory on resource locating field which is motivated by the development of Grid and P2P, and present a query pattern of Grid location based on clustering . This pattern can improve the effectiveness of resource locating on metrics of query overhead and maintenance. Inspired by the self-adaptive, self-learning, self-organization, parallel and distributed characteristics in immune system, a fuzzy immune scheduling algorithm is presented based on the applying of immune theory on task scheduling of Grid. Simulation results show that these methods are feasible and efficient. So, our research on resource locating and task scheduling in resource sharing environment is a meaningful exploring.The main contributions of the dissertation are summarized as follows:1. This paper comprehensively survey the basis Grid knowledge, such as Grid notion, classfication, architecture, and usage schema.2. The present Grid resource management technologies are introduced, meanwhile the notion and the functionality of resource management are discussed briefly. Our analysis is focus on resource location field. In order to overcome the shortage of existing resource locating in sharing environment, the alignment of Grid and P2P must be the future researching works in Grid community.3. A clustering based Grid discovery schema is proposed for sloving the issue of resource locating in this dissertation. The idea behind this schema is that distributed index data can be effectively organised by clustring. In this schema , index data are categorized by resource type firstly, and then a varite of index data are deployed into the class spaces ring organized by Chord protocol. The time cost for determining class...