MOEA-based brokering for hybrid Clouds

Hybrid Clouds couple the scalability of public Clouds with the greater control supplied by private ones. Hybrid Cloud Brokers support customers in selecting the most suitable providers' offers, optionally adding the provisioning of dedicated services with higher Quality of Service (QoS) levels. A clear evaluation of the benefits on performance and energy savings brought about by any allocation strategy often breaks into the trade-off between competing goals: maintaining a satisfactory level of QoS without affecting economic benefit and energy costs. Based on a Multiobjective Optimization Problem formulation of four alternative goals, we propose a genetic approach for Cloud Brokering, focusing on allocating resources to application with diverse QoS requirements. The Multi Objective Evolutionary Algorithm (MOEA) approach allows to obtained sets of competing trade-off allocation solutions. We performed an experimentation through an evolutionary-based broker simulator and evaluated the results.