Research On Resource Management Technology Of Grid Computing With Multiple Qos

Grid computing era has been getting closer, more and more research on ithas been done. For the analysis of the current grid technology, it’s easy to seethat service-oriented technology will become an inevitable trend of the gridcomputing’s future direction. As a key part of the grid system, resourcemanagement system has also become a top priority to the development of gridtechnology. At the same time, the core issue of the resource managementsystem, how to choose or improve grid scheduling algorithm for solving thegrid task scheduling problem, has become a hot and difficult problem ofacademic research.The essence of grid task scheduling problem is NP-complete problem,combined with its own characteristics of the tasks and resource, making theproblem more difficult than the general combinatorial optimization problem.Most current research is focused on solving the optimal span of one-dimensional QoS constraints of grid task scheduling problem, and it does notconsider the complexity of multi-class scheduling objective optimizationresults. Therefore, it do not have a high practical value.This paper come up with a multi-class QoS grid task scheduling by studyinga variety of grid resource management model, the grid task scheduling and QoS parameters modeling. At the same time, the discussion of the commonlyused types of the field of grid task scheduling algorithms has been made, theant colony algorithm has been chose. In the application process, drawing onsolving the transportation problem of urban disaster relief, developing thetraditional QoS utility function as expectations of inspiration factor, and selectthe MMAS algorithm. The minimum span, high quality of service and multi-objective on load balancing of the grid task scheduling have been made by theuse of optimal span of the objective function and QoS window.In order to verify the validity of the model and algorithm, the paper has alsomade some related simulation experiments. The comparison with differentalgorithms has been made on solving the same grid task scheduling problem.The experimental results show that both the time span, the performance ofload balancing have been improved compared with the traditional algorithm.Providing a reference for the actual engineering application of grid computing.