Multiple-QoS Constrained Grid Task Scheduling Algorithm

Grid is an environment integrated with calculation and resources, absorbs various computational resources again and translates them to a kind of computing power which can be available everywhere, reliable, standard as well as economic. In the field of grid computing technique, task scheduling is particularly important. How to arrange different tasks to be completed by relevant grid note properly and satisfy the needs of users according to the calculated performance of grid note and the communicated performance between nodes is a significant factor affecting calculated performance of grid.Numerous researches indicate that task scheduling is a problem involving complete NP. The main purpose of this paper is to produce a better scheduling algorithm based on existing scheduling algorithms to increase the efficiency of grid computing and satisfy the needs of users on QoS.In this essay, deep analysis of some static and dynamic scheduling algorithms is presented in the first part. Then I compare the advantages and disadvantages of each other, putting emphasis on the classic Min-Min algorithm. The effect of QoS on scheduling algorithm is considered based on the above analysis and a more reasonable model which can increase the function and practicability of scheduling algorithm is produced.In the last part of my essay, I use GridSim to carry out series of emulation on the improved scheduling algorithm. The result of emulation indicates that the improved scheduling algorithm functions better and satisfies the needs of users on QoS.