Research Area: Cloud Computing
The increasing cost in power consumption in data centers, and the corresponding environmental threats have raised a growing demand in energy-efficient computing. Despite its importance, little work was done on introducing models to manage the consumption efficiently. With the growing use of Cloud Computing, this issue becomes very crucial. In a Cloud Computing, the services run in a data center on a set of clusters that are managed by the Cloud computing environment. The services are provided in the form of a Software as a Service (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (IaaS). The amount of energy consumed by the underutilized and overloaded computing systems may be substantial. Therefore, there is a need for scheduling algorithms to take into account the power consumption of the Cloud for energy-efficient resource utilization. On the other hand, Cloud computing is seen as crucial for high performance computing; for instance for the purpose of Big Data processing, and that should not be much compromised for the sake of reducing energy consumption. In this work, we derive an energy-aware tasks scheduling (EATS) model, which divides and schedules a big data in the Cloud. The main goal of EATS is to increase the application efficiency and reduce the energy consumption of the underlying resources. The power consumption of a computing server was measured under different working load conditions. Experiments show that the ratio of energy consumption at peak performance compared to an idle state is 1.3. This shows that resources must be utilized correctly without scarifying performance. The results of the proposed approach are very promising and encouraging. Hence, the adoption of such strategies by the cloud providers result in energy saving for data centers.
Keywords:
Green Computing
Energy Efficiency
Energy Management
Scheduling
Cloud Computing
Performance
Author(s) Name: Leila Ismail, Abbas Fardoun
Journal name: Procedia Computer Science
Conferrence name:
Publisher name: Elsevier
DOI: 10.1016/j.procs.2016.04.178
Volume Information: Volume 83, 2016, Pages 870-877
