Minimum-cost virtual-topology adaptation for optical WDM mesh networks

This study considers reducing cost to ISPs by allowing them to dynamically lease only the required amount of bandwidth from network operators in order to satisfy their customers´ traffic needs. Network operators can also generate additional revenue from this approach by leasing out the spare bandwidth to time-of-the-day-insensitive applications. We study a virtual-topology adaptation model for the ISP based on minimizing its total dollar cost for running its network. Relative to the flat-rate model (where the ISP´s topology is fixed based on worst-case traffic), our minimum-cost virtual-topology adaptation method can yield significant cost savings of over 21% (based on a US ISP network employing realistic traffic and bandwidth costs). Our optimization approach is based on a mixed integer linear program (MILP) which nominally adds or deletes one link (lightpath) in the ISP´s virtual topology at a time after a certain observation period, which is a few hundreds of seconds.