A4 Refereed article in a conference publication
Energy Aware Consolidation Algorithm based on K-nearest Neighbor Regression for Cloud Data Centers
Authors: Fahimeh Farahnakian, Tapio Pahikkala, Pasi Liljeberg, Juha Plosila
Publication year: 2013
Book title : 6th IEEE/ACM International Conference on Utility and Cloud Computing
ISBN: 978-0-7695-5152-4
Abstract
As energy consumption of ICT infrastructures has increased considerably in the recent years, the research community and companies are working on energy-aware resource management strategies. In order to reduce energy cost, we propose a dynamic virtual machine consolidation algorithm to minimize the number of active physical servers on a data center. However, the reliable quality of service that defines via Service Level Agreement (SLA) between data centers and their users is essential for Cloud environment. Therefore, the reduction of SLA violation level and energy cost are considered as two objectives in this paper. The proposed dynamic consolidation method uses the k-nearest neighbor regression algorithm to predict the amount of future CPU usage. Based on prediction utilization, the consolidation method decides (i) when a host becomes over-utilized (ii) when a host becomes under-utilized (ii) which host will not be over-utilized by allocating a virtual machine. Experimental results on the real workload traces from more than a thousand PlanetLab virtual machines show that the proposed technique minimizes power consumption and maintains required performance levels.
As energy consumption of ICT infrastructures has increased considerably in the recent years, the research community and companies are working on energy-aware resource management strategies. In order to reduce energy cost, we propose a dynamic virtual machine consolidation algorithm to minimize the number of active physical servers on a data center. However, the reliable quality of service that defines via Service Level Agreement (SLA) between data centers and their users is essential for Cloud environment. Therefore, the reduction of SLA violation level and energy cost are considered as two objectives in this paper. The proposed dynamic consolidation method uses the k-nearest neighbor regression algorithm to predict the amount of future CPU usage. Based on prediction utilization, the consolidation method decides (i) when a host becomes over-utilized (ii) when a host becomes under-utilized (ii) which host will not be over-utilized by allocating a virtual machine. Experimental results on the real workload traces from more than a thousand PlanetLab virtual machines show that the proposed technique minimizes power consumption and maintains required performance levels.
UTU Research Portal