A petri net model-based multiple workflow fragment scheduling schemes in cloud

Abstract

In a cloud environment, workflow management system which executes the user requested workflow application within SLA by allocating resources from cloud service providers is emerged. Therefore, multi workflow scheduling problem within the SLA received attentions. Most of research works on workflow scheduling were limited to the single workflow. In addition, research works about multi workflow scheduling considers only the workflow execution priorities or do not copes with user SLA. Therefore we propose a multi-workflow scheduling heuristics that changes the workflow topology into fragments and schedule them onto proper VM. Our heuristics have three steps: workflow fragmentation, fragment scheduling, workflow configu-ration. Workflow fragmentation process recursively identifies the workflow’s critical path by Petri Net model. Then, sub-critical paths in identifying process will be included to the fragments. With this step, we can simpli-fy the complex workflow topology into fragments with sequential tasks. Therefore, we can expect the com-plexity of scheduling process will get lower and resources will be scheduled in an economizing way. Fragment scheduling heuristics schedules fragments onto the VMs in a fast way using the fact that fragments are consisted of sequential tasks. Our heuristics are suitable for the cloud environment and has a low complexity. Therefore it can be applied to the multi-workflow environment. Workflow reconfiguration schemes offers some algorithms that save the resources or the execution time by simply changing the assigned VMs.