HyViDE

Abstract

Cloud computing technologies have matured enough that the service providers are compelled to migrate their services to virtualized infrastructure in cloud data centers. However, moving the computation and network to shared physical infrastructure poses a multitude of questions, both for service providers and for data center owners. In this work, we propose HyViDE - a framework for optimal placement of multiple virtual data center networks on a physical data center network. HyViDE preselects a subset of virtual data center network requests and uses a hybrid strategy for embedding them on the physical data center. Coordinated static and dynamic embedding algorithms are used in this hybrid framework to minimize the rejection of requests and fulfill QoS demands of the embedded networks. HyViDE can employ suitable static and dynamic strategies to meet the objectives of data center owners and customers. Experimental evaluation of our algorithms on HyViDE shows that, the acceptance rate is high with faster servicing of requests.