Research On The Model Of Computing Grid Resource Management Scheduling Based On LDS Algorithm

Grid is a new technology that has been flourishing quickly in recent years. Its occurrence has whipped up the next wave of Internet technology. Computation grid originally aimed to combine supercomputers through Internet so as to solve complex and large-scale computing problems. At present, the aim has evolved into combining various types of computers in different areas in a reasonable way and the highly integrated organic unity can supply powerful computing abilities to common users.Resource management scheduling is a key technology of computing grid. The basic function of resource management scheduler is receiving the request of machines' requirement of resources, assigning given resources to the requester, and scheduling relevant resources in reason in order to start the application presented by the requester. The validity and applicability of the scheduler depend on its scheduling strategy mostly. And the gird characteristics such as high heterogeneity, adaptability, distribution and dynamic decide the complexity of grid resource scheduling. Most existing computing grid resource management scheduling models adopt the strategies whose scheduler drive scheduling methods and ignoring resource agent's character of dynamic, the results of these strategies are often worse than assuming.Aiming at the problem discussed above, a computing grid resource management scheduler based on LDS algorithm is presented in this paper based on current methods and widely study of the theory of computing grid and its resource management scheduling. The scheduler adopts the MDS (meta-computing directory service) of the Globus project as its information collector before scheduling, introduces the LDS algorithm in its scheduling strategy, uses the scheduling method that drove by resource agent and the master and agents correspond during its scheduling process.The paper uses SimGrid to evaluate the model. Experiments prove that model adapts the LDS well and has high performance of enhancing scheduling speed and balancing workload. And it is the samae with the situation of multi-task and multi-agent.