Towards Feasible and Effective Load Sharing in a Heterogeneous Computational Grid

Abstract

Abstract.A grid has to provide strong incentive for participating sites to join and stay in it. Participating sites are concerned with the performance improvement brought by the gird for the jobs of their own local user communities. Feasible and effective load sharing is key to fulfilling such a concern. This paper explores the load-sharing policies concerning feasibility and heterogeneity on computational grids. Several job scheduling and processor allocation policies are proposed and evaluated through a series of simulations using workloads derived from publicly available trace data. The simulation results indicate that the proposed job scheduling and processor allocation policies are feasible and effective in achieving performance improvement on a heterogeneous computational grid.

Improving Processor Allocation in Heterogeneous Computing Grid through Considering Both Speed Heterogeneity and Resource Fragmentation

“…This scheme works fine in a homogeneous grid. However, in a heterogeneous grid with computing speed differences among participating sites, the best-fit method may not perform well since it does not consider the speed heterogeneity [9]. In such an environment another processor allocation method called fastest-first has been proposed [9].…”

“…However, in a heterogeneous grid with computing speed differences among participating sites, the best-fit method may not perform well since it does not consider the speed heterogeneity [9]. In such an environment another processor allocation method called fastest-first has been proposed [9]. The fastest-first method focuses on speed heterogeneity in a heterogeneous grid and allocates a job to the fastest one among all the sites which can accommodate the job.…”

“…The proposed intelligent allocation method is inspired by the adaptive allocation strategy presented in [9] which makes allocation decision based on a calculation of which policy can further accommodate more jobs for immediate execution. The improvement in the intelligent allocation method is to take the speed difference into account.…”

“…We compared the proposed Intelligent (IT) method with the adaptive (AD) [9], best-fit (BF) [5], and fastest-first (FF) [9] methods. In order to evaluate the performance of the proposed method on various workload conditions, we conducted a series of experiments by varying three adjustable parameters listed in Table III.…”

“…For the best-fit method a particular site is chosen for a job on which the job will leave the least number of free processors if it is allocated to that site. Although the best-fit method can effectively alleviate the resource fragmentation problem, it cannot achieve good performance in a heterogeneous grid as shown in [9]. This is because in a heterogeneous grid resource fragmentation is not the sole factor that affects the overall system performance.…”