An adaptive random heuristic in virtual networks: Dependability analysis

Network Virtualization has been a central matter for the future of the Internet. Nevertheless, virtualized structures are as failure-prone as their underlying physical infrastructure. Hence, dependability attributes constitute relevant metrics for virtual network allocation decisions. This work proposes and evaluates a greedy solution to mapping virtual networks that is a well-known NP-Hard problem. We take into account a dynamic networking scenario where the time to map requests and availability are important factors. Results show that the proposed heuristic achieves an availability above 99% in the different scenarios simulated.