Scheduling Workloads with Time-Varying Jobs Type on Distributed Systems

In various scientific domains, a range of complex job types are executed within distributed systems. The nature of these jobs may evolve over time, presenting a dynamic challenge for resource management. A key issue in such environments is the deployment of efficient job scheduling algorithms. investigates the dynamic scheduling of intricate workloads across This study distributed resources. The set of applications under consideration include bag-of-tasks jobs and linear workflows, with job characteristics that change over time. Epoch-based scheduling is utilized, assigning distinct weights to different job assess the types. To performance of this epoch-based scheduling approach, simulation experiments are conducted under different system load conditions.