An ambitious vision of the future of the Internet would include network virtualization. This new paradigm consists in obtain and share resources of the network infrastructure among several virtual network providers, whose goal is to improve the utilization of resources of the physical networks. However, few studies have dealt with practical challenges that appears in virtual network design, such as embedding problems, scheduling resources, mobility, privacy etc. In this work, one of the problems of interest is the virtual network embedding problem, where the objective is to map virtual networks in to physical networks, in order to use the resources efficiently, satisfying QoS requirements, in terms of bandwidth, processing capacity and memory, minimizing the rejection rate to requests of virtual networks. Since the embedding problem is NP-hard, to deal with its computational difficulty requires unusual approaches. Another problem of interest, which is also NP-hard, is the virtual network scheduling problem, where the goal is to meet all requests of Virtual Networks, in the shorter possible time. Our objectives in this work is to investigate the state of the art in embedding and scheduling of virtual networks, and to propose heuristics and exact algorithms to these problems, in order to obtain good solutions for these problems.
News published in Agência FAPESP Newsletter about the scholarship: