The Effect of Laxity of Real-Time Workflow Applications on the Performance of Elastic Cloud Resources

An increasing number of workflow applications with varying deadlines are processed on cloud platforms. One of the main characteristics of such environments is the elasticity of the virtualized resources, which allows them to be dynamically scaled in or out. Consequently, it is imperative to examine in this context how the laxity (slack time) of the workload affects performance. To this end, in this paper we investigate the effect of laxity of real-time workflow jobs on the performance of elastic cloud resources. We formulate the problem and evaluate through simulation the impact of laxity on the deadline miss ratio, energy consumption and monetary cost associated with the resources, considering workloads with different slack times. The simulation results provide useful insights into how each performance metric is affected by the laxity of the real-time workflow jobs.