Enhanced Two-Phase Virtual Machine Placement Scheme for Cloud Computing Datacenters

Over the years, Cloud Computing has offered many benefits such as providing services to end-users on demand. However, infrastructure as service providers is faced with the challenge of handling many end-users' requests for virtual resources. In this regard, one of the key challenges of resource allocation is Virtual Machine Placement (VMP) problem. However, the dynamicity and uncertainty of Cloud platform and the unpredictable nature of the end-users' requests have rendered the VMP problem more interesting. Recently, a two-phase Virtual Machine Placement scheme, combining the benefits of both online (dynamic) and offline (static) formulations were presented. The proposed scheme is based on a prediction-based triggering method used to determine when to trigger the VMP reconfiguration phase. However, the existing method leads to a less accurate prediction outcome, therefore, results in less optimal solution from the reconfiguration phase. This work proposes an enhanced two-phase Virtual Machine Placement strategy based on novel triggering method. The proposed triggering method considers Damped trend exponential smoothing method. An experimental evaluation is performed against the previous approach, considering 160 scenarios. The experimental results show that the proposed work achieved a minimum cost-function when compared with the benchmark approach.