Scheduling Gangs with Different Distributions in Gangs´ Degree of Parallelism in a Multi-Site System

Gang scheduling is considered to be a highly effective task scheduling policy for distributed systems. Gangs are jobs which consist of a number of interacting tasks which are scheduled to run simultaneously on distinct processors. Simulation experiments are conducted to address performance issues which concern the impact of the variability in gangs´ degree of parallelism in the case of implemented migrations. A simulation model consisting of two sites is used to provide results on the performance of the system.