An Enhanced Algorithm for Cloud Scheduling with Setup Cost

Abstract

The study aimed at investigating the problem of online task scheduling of identifying job of MapReduce on cloud computing infrastructures. It was proposed that the virtualized cloud computing setup comprised machines that host multiple identical virtual machines (VMs) under pay-as-you-go charging and booting a VM requires a constant setup time. The problem occurs when the VMs turning off after finishing processing a task and running it again for another. The goal is to limit the delay resulted from setting up the VMs, and to minimize the idle cost (VM does not find any task to possess) when VM is continuing in an active state without processing tasks. It was a constant number of VMs in an activation state continuity, and the cost of idle was distributed over existing tasks where the cost should be less than the cost of setting up one VM. The researchers' algorithm limited the delay resulted from setting up the VMs, the cost resulted from the continuing VMs in an active state without processing tasks was distributed to the current tasks fairly. The known cases was discussed where the duration of each task was known upon its arrival.