Optimization of survivable networks with simultaneous unicast and anycast flows

This work focuses on the problem of survivable network design with simultaneous unicast and anycast flows. The motivation for this problem is twofold. First, in recent years we can observe an increasing popularity of network services using anycast flows (e.g. Content Delivery Networks (CDN), Domain Name Service (DNS), peer-to-peer (P2P) systems. Second, current users require reliable network services, therefore network operators are obliged to develop methods to improve the network survivability. We formulate an Integer Programming problem of joint optimization of network capacity, working and backup connections for both unicast and anycast flows. The goal is to minimize the network cost required to protect the network against failures using the single backup path approach. To facilitate the high computation complexity caused by the NP-completeness of the considered problem, we develop new cut inequalities. We evaluate these cuts by making a comparison using various experiment scenarios. Moreover, we propose and evaluate through experiments a new heuristic algorithm.