Energy-Aware Lease Scheduling in Virtualized Data Centers

TL;DR

This paper proposes energy-aware heuristics for scheduling virtual machine leases in private cloud data centers, balancing energy efficiency with quality of service, and demonstrates significant energy savings through simulation.

Contribution

It introduces novel heuristics for VM lease scheduling that improve energy efficiency while maintaining performance requirements in virtualized data centers.

Findings

Significant reduction in energy consumption compared to FCFS scheduling.

Migration policies further enhance energy savings.

Heuristics effectively balance energy use and QoS.

Abstract

Energy efficiency has become an important measurement of scheduling algorithms in virtualized data centers. One of the challenges of energy-efficient scheduling algorithms, however, is the trade-off between minimizing energy consumption and satisfying quality of service (e.g. performance, resource availability on time for reservation requests). We consider resource needs in the context of virtualized data centers of a private cloud system, which provides resource leases in terms of virtual machines (VMs) for user applications. In this paper, we propose heuristics for scheduling VMs that address the above challenge. On performance evaluation, simulated results have shown a significant reduction on total energy consumption of our proposed algorithms compared with an existing First-Come-First-Serve (FCFS) scheduling algorithm with the same fulfillment of performance requirements. We also discuss the improvement of energy saving when additionally using migration policies to the above mentioned algorithms.