Resource-sharing Policy in Multi-tenant Scientific Workflow as a Service Platform
share
Increasing adoption of scientific workflows in the community has urged for the development of multi-tenant platforms that provides these workflows execution as a service. Workflow as a Service (WaaS) concept has been brought up by researchers to address the future design of Workflow Management Systems (WMS) that can serve a large number of users from a single point of service. This platform differs from a traditional WMS in handling the workload of workflows at runtime. A traditional WMS is usually designed to execute a single workflow in a dedicated process while the WaaS platforms enhance the process by exploiting multiple workflows execution in a resource-sharing environment model. In this paper, we explore a novel resource-sharing policy to improve system utilization and to fulfill various Quality of Service (QoS) requirements from multiple users. We propose an Elastic Budget-constrained resource Provisioning and Scheduling algorithm for Multiple workflows designed for WaaS platforms that is able to reduce the computational overhead by encouraging resource-sharing policy to minimize workflows' makespan while meeting user-defined budget. Our experiments show that the EBPSM algorithm is able to utilize the resource-sharing policy to achieve higher performance in terms of minimizing the makespan compared to the state-of-the-art budget-constraint scheduling algorithm.
READ FULL TEXT