AUTO RESOURCE MANAGEMENT TO ENHANCE RELIABILITY AND ENERGY CONSUMPTION IN HETEROGENEOUS CLOUD COMPUTING

Moataz H. Khalil^{1, 2}, Mohamed Azab² , Ashraf Elsayed³, Walaa Sheta^{1, 2} Mahmoud Gabr³ and Adel S. Elmaghraby^1,2

¹CECS Department, University of Louisville, Kentucky, USA.

²The City of Scientific Research and Technology Applications, Egypt.

³Department of Mathematics & Computer Science, Faculty of Science,

Alexandria University, Alexandria, Egypt.

Abstract 

A classic information processing has been replaced by cloud computing in more studies where cloud computing becomes more popular and growing than other computing models. Cloud computing works for providing on-demand services for users. Reliability and energy consumption are two hot challenges and tradeoffs problem in the cloud computing environment that requires accurate attention and research. This paper proposes an Auto Resource Management (ARM) scheme to enhance reliability by reducing the Service Level Agreement (SLA) violation and reduce energy consumed by cloud computing servers. In this context, the ARM consists of three compounds, they are static/dynamic threshold, virtual machine selection policy, and short prediction resource utilization method. The Minimum Utilization Non-Negative (MUN) virtual machine selection policy and Rate of Change (RoC) dynamic threshold present in this paper. Also, a method of choosing a value as the static threshold is proposed. To improve ARM performance, the paper proposes a Short Prediction Resource Utilization (SPRU) that aims to improve the process of decision making by including the resources utilization of future time and the current time. The output results show that SPRU enhanced the decision-making process for managing cloud computing resources and reduced energy consumption and the SLA violation. The proposed scheme tested under real workload data over the CloudSim simulator

Keywords 

Cloud computing, Service Level Agreement, Reliability, energy consumption, Virtual machine migration, Resource management, utilization resource prediction.
                       

                                                  Full Text