A static VM placement and hybrid job scheduling model for green data centers

Area of Science:

• Computer Science
• Energy Efficiency in Data Centers
• Cloud Computing

Background:

• Data center energy consumption is a significant concern.
• Reducing physical and virtual machines (VMs) enhances energy efficiency.
• VM migration has associated disadvantages.

Purpose of the Study:

• To propose a static VM placement algorithm to reduce energy consumption.
• To investigate the impact of job scheduling policies on VM requirements and energy use.
• To analyze energy consumption for real-time and non-real-time applications under different scheduling policies.

Main Methods:

• A static VM placement algorithm using the Worst-Fit-Decreasing (WFD) fashion.
• Modeling each VM using an M/M/* queue.
• Analysis of space-shared, time-shared, and hybrid job scheduling policies.
• Evaluation of energy consumption and Service Level Agreement (SLA) violations.

Main Results:

• The hybrid job scheduling policy is more effective than space-shared and time-shared policies in reducing energy consumption and SLA violations.
• The proposed non-migration VM placement method surpasses migration-based algorithms in lowering energy consumption and SLA violations.

Conclusions:

• Static VM placement and hybrid job scheduling offer significant energy savings in data centers.
• Avoiding VM migration can lead to improved energy efficiency and better adherence to Service Level Agreements.