Modeling of Survivable Network Design Problems with Simultaneous Unicast and Anycast Flows

Abstract

The paper centers around the problem of survivable network design for simultaneous unicast and anycast flows. This problem arises from the growing popularity of network services using anycast flows (e.g. Domain Name Service (DNS), peer-to-peer (P2P) systems, Content Delivery Networks (CDN)) and the increasing need to provide reliable network services. We formulate four new optimization models of joint optimization of network capacity, working and backup connections for both unicast and anycast flows. The objective is to minimize the network cost necessary to build a survivable network protected against network failures by the backup path method. We consider optimization of capacity for both working and backup flows as well as optimization of only spare capacity. The presented formulation is generic and various failure scenarios can be considered. We show some exemplary results of numerical experiments proving that the use of anycast paradigm can reduce the cost of the survivable network.