Scheduling Utility In Grid And Simulation

The growth of f the technology of computing and network mostly promotes the commercial application and scientific research based on Internet. In many fields, there are more and more requests for the computing capacity and it is difficult for a single computer to undertake calculation task. As the result of the increasing requirements of various applications which need more calculative resources and capacities, the grid computing technology appears. The final goal of grid is to realize resource share and coordinating work in a virtual network environment. The resource management and scheduling mechanism is a part of the Grid core services. An excellent resource management and scheduling strategy can coordinate and schedule the grid resources effectively, it can effectively reduce the total run time and total expense in the grid computing as well, which will implement the purpose of maximizing the preference of the grid. We compared these serveral schedule algorithm that schedules user's Bag of Task applications on a divisable auction-based resource allocation systems. We investigate a divisible auction that is proportionally fair. It is shown that the structure of the auction allows for a quasilinear characterization of a wide variety of broker tasks. We prove that the auction has a unique Nash equilibrium and propose a decentralized bidding strategy.This paper aims at how to manage and schedule the grid resources effectively and reasonably. The major research works are below:(1) Combined with the characteristics of the grid and the actual situation of the market-based, Scheduling model of grid to achieve the purpose of efficient and rational allocation of resources(2) Discussing some basic concepts of mechanism design approach, and proof techniques, and make an example.(3) We had analysis framework of economic models, discuss different schedule algorithms in Nimrod-G. We present an synthesis Deadline and Cost Optimization strategy. Contrast to cost optimal and time optimal, Deadline and Cost Optimization strategy consider more QoS factors. Contrast to C-T strategy, Deadline and Cost Optimization is easy to realized and more dynamic for it can easy adjust weight.(4) For Several different algorithms to achieve the simulation experiments, getting the experimental data, and charts. Then use experimental data to evaluate the effectiveness of different algorithms. It is shows that the design of the algorithm is feasible to achieve our expectation...