Strategies for Rescheduling Tightly-Coupled Parallel Applications in Multi-Cluster Grids

Abstract

As computational Grids are increasingly used for executing long running multi-phase parallel applications, it is important to develop efficient rescheduling frameworks that adapt application execution in response to resource and application dynamics. In this paper, three strategies or algorithms have been developed for deciding when and where to reschedule parallel applications that execute on multi-cluster Grids. The algorithms derive rescheduling plans that consist of potential points in application execution for rescheduling and schedules of resources for application execution between two consecutive rescheduling points. Using large number of simulations, it is shown that the rescheduling plans developed by the algorithms can lead to large decrease in application execution times when compared to executions without rescheduling on dynamic Grid resources. The rescheduling plans generated by the algorithms are also shown to be competitive when This work is supported by Department of Science and Technology, India. project ref. no. compared to the near-optimal plans generated by brute-force methods. Of the algorithms, genetic algorithm yielded the most efficient rescheduling plans with 9-12% smaller average execution times than the other algorithms.

Experimental Performance Analysis of Job Scheduling Algorithms on Computational Grid using Real Workload Traces

“…A work presented in [19] developed three scheduling algorithms to schedule and reschedule the jobs on grid resources by considering dynamics of resources and applications. The three algorithms follow the incremental, divide and conquer and genetic algorithm principle.…”

“…The three algorithms follow the incremental, divide and conquer and genetic algorithm principle. The study [19] suggests continuous monitoring of tasks in execution and resources executing the scheduled tasks, if the performance is degraded rescheduling of tasks will be performed.…”

Process Migration: Controlling Application and Resource Dynamics by Combining Computation, Communication and Memory Metrics

“…Consequentemente, o reescalonamento emerge como uma alternativa para remapear processos de maneira eficiente em Grids. Caso contrário, os benefícios de usar múltiplos processadores pode ser nulo [13]. Basicamente, as principais questões no tratamento do reescalonamento são: (i) como ele é oferecido para os desenvolvedores e; (ii) quais métricas serão usadas para compor a noc ¸ão de carga.…”