Research On Grid Resource Managing Technologies Based On Self-organized Virtual Grouping

Features of the grid include the large number of nodes, the heterogeneity among nodes and autonomy of each other. Dynamic and heterogeneous characteristics of large-scale grids make the fundamental problem of resource discovery a great challenge.This thesis has presented a self-organized virtual grouping (SVG) infrastructure of resource managing strategies for grid, while the resource nodes are similar to each other for characteristics in single SVG node. Since the SVG method takes advantage of both centralized and decentralized approaches that were previously developed for Grid/P2P resource discovery, it can minimize the overhead incurred in forming and maintaining groups and maximize resources discovery performance.This thesis has developed and highlighted six SVG algorithms, node joining the grid algorithm, node leaving the grid algorithm, SVG formation and maintenance algorithm, leader replacement algorithm, resource discovery routing algorithm and hybrid multicast algorithm.The thesis has also thoroughly studied on the maximum number of nodes of a SVG and has gotten formulas on variable premises. After laborious research these formulas, I found that the number of nodes of the SVG is decided by supply for the grid application of resources proportion, nodes that have already been constituted, and characteristic of the candidate node and characteristic of the leader.In order to evaluate the SVG-based resource managing strategies and its key technology of the grid system, I carried out four groups of experiments. As expected, the testing results showed that SVG-based resource managing mechanism can effectively resolve the puzzles in the field of grid resource management.