A Workflow Scheduling Deadline-Based Heuristic for Energy Optimization in Cloud

Abstract

This article addresses the scheduling of heterogeneous scientific workflows while minimizing the energy consumption of the cloud provider, by introducing a deadline sensitive algorithm. Scheduling in a cloud environment is a difficult optimization problem. Usually, work around the scheduling of scientific workflows focuses on public clouds where infrastructure management is an unknown black box. Thus, many works offer scheduling algorithms designed to select the best set of virtual machines over time, so that the cost to the end user is minimized. This article presents a new HEFTbased algorithm that takes into account users deadlines to minimize the number of machines used by the cloud provider. The results show the real benefits of using our algorithm for reducing the energy consumption of the cloud provider.

Online Multi-User Workflow Scheduling Algorithm for Fairness and Energy Optimization

“…Those experiments were performed on the Seduce platform [20], a scientific testbed integrated to Grid'5000, that monitor the electrical consumption of the nodes described in Table II. This evaluation section shows comparisons between NEARDEADLINE, NEARDEADLINERATIO, HEFT [5], and HEFT DEADLINE [9]. Both HEFT and HEFT DEADLINE are presented in the Section II.…”

“…The IC-PCPD2 algorithm (Iaas Cloud Partial Critical Path with Deadlines Distribution) which was originally dedicated to Grid infrastructure in [8], uses the cheapest resources in priority while trying to respect the deadline of the workflow. In our previous work [9], we have presented the HEFT deadline algorithm that schedules workflows for multiple users in a Cloud infrastructure, while respecting the deadline of each workflows. The objective of this algorithm was to minimize the energy consumption of the infrastructure, by minimizing the number of machines used to execute a given workload.…”

“…The algorithms presented in [3], [7], [9] have the same limitation: they do not take uncertainty into account and perform a static schedule that has to be respected. In addition, all the algorithms considering public Cloud environment, such as [3], [7], are not designed for multi-users scheduling problems.…”

“…Such capacity constraints can be modeled as a classical bin packing problem. A more formal definition of this problem has been presented in our previous work [9] and is not detailed in this document for sake of space. As over-provisioning is not allowed in this work the deteroration of VM computing speed is assumed to be low [14], and considered to be 5% of the host node speed.…”

“…The fairness is represented by the respect of the deadline of each workflow, and thus this objective can be define as the minimization of the sum of deadline violations. The deadline violation of a workflow is defined by Equation (8) and the associated objective by Equation (9).…”