Maximizing access to IT services on resilient optical grids

An optical grid network provides high speed communications for large scale applications and services may require, for (very) limited time periods, an ultra-high bit rate network services at the order of the transmission capacity of the network infrastructures. In the context of resilient optical grids, we investigate how to maximize the grade of services for given transport capacities, while maximizing the protection level, i.e., against single link vs. single link and node (including server node) failures. In this paper, we present a large scale optimization model, solved with the help of Column Generation (CG) techniques. The model allows the exploration of different protection schemes: protection against single link failures, or single link and node failures, including or not the single failures of the server nodes. Numerical results are presented on some European and German network topologies. Results show that, for given transport capacities, guaranteeing protection against single link and node failures, or only against single link failures does not affect significantly the grade of services. Results also show that the link dimensioning which is proposed is fairly efficient in terms of bandwidth utilization.