ARMM: Adaptive Resource Management Model for Workflow Execution in Clouds

摘要

Cloud offers computational resources as a utility to execute dependent tasks ensemble as an application workflow, where each task has a different resource requirement. Resource management frameworks are required to dynamically provision resources to enable scalability and seamless execution of workflows. In this paper, an adaptive resource management model is presented, which allocates and reschedule the resources based on their usage history and performance metrics. It further makes decisions to adapt workflow tasks to optimize deadline, budget and resource performance. A case study using different workflows is used to describe the model in a simulated environment considering various run time scenarios.