A Proactive Approach to Load Balancing of Workflow Execution in a SaaS Environment

Achieving load balancing is essential to ensure better scalability and higher performance for a SaaS environment, in which requests from different tenants are satisfied concurrently by a single service instance over shared hosting resources. In this paper we discuss load-balancing technologies for a special kind of SaaS infrastructure that provides services to execute workflow processes. First, we explore some key issues in architecting a workflow management system (WfMS) that enables SaaS model and propose our prototype - VINCA4SaaS. Then we propose a proactive load balancing algorithm, by which requests from different tenants are scheduled according to workflow engines´ current capability and the predicted cost of executing the process instances. The experimental results show that the algorithm can distribute workload more fairly when continually executing diverse workflow processes on uniform workflow engines than the formerly presented ones.