Conference Paper

Hollow-core fiber (HCF) technology is experiencing fast progress, moving from a research curiosity closer to a commercially viable product, with small scale deployments being announced over multiple geographies. Earlier deployments of HCF focus on taking advantage of its low latency property in relatively short links. However, the (theoretical) very low loss of HCF and negligible nonlinearities raise the prospect of HCF as an enabler of higher capacity optical networks, motivating its deployment in the demanding regional and long-haul network segments. Given the expected higher cost of HCF, an interim solution to leverage the benefits of this fiber type is to have a mixed deployment of HCF and traditional solid core fiber. In a mesh network, this requires carefully selecting where to place the limited HCF budget as to maximize the improvements in performance. Importantly, this placement decision can be made either at a span- or link-level. The former enables a more granular deployment of HCF, whereas the latter favors deployment simplicity. This paper extends an integer linear programming (ILP) model previously proposed by the authors to consider both span- and link-based HCF placement in a mesh optical network. Detailed simulations over a reference optical transport network provide insight on the efficacy and scalability of the proposed method and the trade-offs between both HCF placement strategies.