Policy-based management approach for energy efficient cloud computing infrastructure
Alhaddadin, Fadi Jamil Shehadeh
MetadataShow full metadata
Cloud computing technology is gaining great popularity in our day due to the utility-oriented information technology services that it offers worldwide. Due to the pay-as-you-go elasticity that cloud computing technology facilitates, hosting pervasive applications has become possible from various ends such as consumer, scientific, and business domains. The technology of Cloud Computing facilitates a computing-as-a-service model where computing resources are made available as a utility service. However, although cloud computing technology returns great benefits in different aspects, data centers consume significant amounts of electricity in order to run, hence they require high operational costs and cause harmful outcomes to the environment such as carbon footprints and emissions. Therefore, we found it valuable to design a new cloud system model which contributes towards reducing the energy consumption of cloud datacenters, with consideration to the quality of service delivered. In this thesis, we have proposed a new cloud system model and architecture that is able to handle and manage a group of virtual machine migration heuristic algorithms proposed by other researchers. The proposed system model, User Profile Aware Policy Switching algorithm (UPAPS) framework has proven ability in managing a group of these heuristic algorithms by employing our proposed architectural components namely: the UPAPS algorithm and the User Service Profile (USP). The UPAPS algorithm together with the USP component contribute to the efficient management of heuristic algorithms and therefore have achieved the desired trade-off between energy consumption and quality of service delivered. The USP component in the UPAPS framework is intended to be the instructive part in the system model which is configured according to users’ requirements. It contains the requirements of the user for other components of the system to work accordingly. The UPAPS component is part of the policy-based management and it functions according to the instruction held in the USP component. The extensive simulation results have shown a great improvement on the cloud infrastructure in terms of power efficiency and resource management using our proposed approach (i.e., UPAPS framework). The validation of our proposed system model has been done through conducting two different cloud scenario studies under both the current system model and the proposed system model. Simulation results for both systems were compared, and the UPAPS framework showed significant reduction in energy consumption in comparison to the current system without violating the quality of service required by the cloud users in both scenarios.