Workflow as a Service in the Cloud: Architecture and Scheduling Algorithms

Abstract With more and more workflow systems adopting cloud as their e xecution environment, it becomes increasingly challenging on how to effic iently manage various workflows, v irtual machines (VMs) and workflow e xecution on VM instances. To make the system scalable and easy-to-extend, we design a Workflow as a Service (WFaaS) a rchitecture with independent services. A core part of the architecture is how to efficiently respond continuous workflo w requests from users and schedule their e xecutions in the c loud. Based on diffe rent targets, we p ropose four heuristic workflo w scheduling a lgorith ms for the WFaaS architecture, and analy ze the differences and best usages of the algorith ms in terms of performance, cost and the price/performance ratio via experimental studies.