Job Allocation Strategies with User Run Time Estimates for Online Scheduling in Hierarchical Grids

Abstract

We address non-preemptive nonclairvoyant online scheduling of parallel jobs on a Grid. We consider a Grid scheduling model with two stages. At the first stage, jobs are allocated to a suitable Grid site, while at the second stage, local scheduling is independently applied to each site. We analyze allocation strategies depending on the type and amount of information they require. We conduct a comprehensive performance evaluation study using simulation and demonstrate that our strategies perform well with respect to several metrics that reflect both user-and system-centric goals. Unfortunately, user run time estimates and information on local schedules does not help to significantly improve the outcome of the allocation strategies. When examining the overall Grid performance based on real data, we determined that an appropriate distribution of job processor requirements over the Grid has a higher performance than an allocation of jobs based on user run time estimates and information on local schedules. In general, our experiments showed that rather simple schedulers with minimal information requirements can provide a good performance.

On Efficiency of Multi-job Grid Allocation Based on Statistical Trace Data

“…Ramirez-Alcaraz et al [19] have analyzed different Grid allocation strategies depending on the type and amount of information they require, and they found that information about users' runtime estimate and local schedules does not help to significantly improve the outcome of the allocation strategies. They concluded that quite simple schedulers with minimal information requirements can provide good performance.…”

Energy-Efficient Job Scheduling and Allocation Scheme for Virtual Machines in Private Clouds

“…In this approach [11], [14 ], [15] number of Hosts required two, waiting time reduces , L 3 Backfill with L 2 so consumer fills happy as compare to previous approach and waiting time of lease L 3 reduces, number of hosts required two and conserve more energy as shown in fig.2. …”

“…VM 11 VM 12 EESAS is implement on base of three conditions i) No workload i.e zero workload ii) Minimum workload and, iii) Maximum workload , when minimum workload then start virtual machine as per VM request, if no workload then do not start any VM for conserving energy purpose, maximum workload then apply migration of VM , clone one VM to multiple VM , in this EESAS with easy backfilling FCFS only allocation of job to VM is change.…”

“…In previous study [9], [10] energy-efficient scheduling scheme is same but only job allocation strategy to virtual machine is easy-backfilling with first come first serve [11] , if saw a keen observation on hosts i.e. physical machine …”

ExPERT: Pareto-Efficient Task Replication on Grids and a Cloud

“…Ramírez-Alcaraz et al [44] evaluate scheduling heuristics and optimize a combined objective for parallel jobs, because they assess that computing a Pareto frontier in a GRID environment is too slow. However, approximating the Pareto frontier for the cases we demonstrated here using ExPERT takes only minutes, which we do not consider "too slow" for a BoT that runs for hours.…”