Scheduling bags of tasks and gangs in a distributed system

Scheduling algorithms play an important role in the performance of distributed systems. A scheduling algorithm is more efficient when it is designed to handle complex workloads which may consist of different types of parallel jobs. In this paper we examine the performance of hybrid scheduling algorithms which can handle diverse workloads consisting of a mixture of bags of independent gangs and bags of tasks. Gangs are jobs which consist of a number of interacting tasks which are scheduled to run simultaneously on distinct processors. A simulation model is used to reveal the significant role of the proposed hybrid scheduling algorithms which can handle simultaneously different types of parallel jobs which coexist in the workload.