Hybrid Approach for Resource Provisioning in Cloud Computing

Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effort or service provider interaction. Elasticity of resources is considered as a key characteristic of cloud computing using this key characteristic; internet services are allocated the onlyneeded resources. This allocation of resources however should not be at the expense of the services’ performance. Allocation of resources without degrading performance is called resource provisioning. Resource provisioning does not only support the elasticity of resources, but also enhances cost efficiency and sustainability. The goal of this work is to investigate resource provisioning to increase the percentage of resources utilization without degrading the performance so that the power consumption of the cloud data centers is reduced. To achieve this goal, a hybrid-approach for resource provisioning is developed. In this approach, a list of virtual machines is requested, passed to a selection algorithm, sorting the machines according to their load, compute the threshold of the machines’ load, and combining the high load with low load from two different virtual machines on one super virtual machine. The approach was implemented in a simulator called CloudSim. It was used to run two sets of experiments. The first is to measure the power consumption of the data center as whole and hosts as well. And the second is concerned with the processing times and memory usage. The results have shown that this approach outperforms traditional counterparts in resource provisioning. The results showed that the hybrid approach achieved reduction of (5.85 MW/s) in power consumption compared with the traditional counterparts for the whole data center, as well as reduction of (2.48 MW/s) in power consumption for the hosts.