Optimizing satellite and core networks for a global quantum network

Quantum key distribution (QKD) promises information theoretic security. However, the exponential decay of the secure key in optical fibers leads to limitations in long distance QKD distribution across fibers, which is necessary for global quantum networks (QNs). Satellite QKD can be used to generate keys over long distances bypassing fiber limitations and is thus a promising approach for global QNs. In this paper, we construct mixed integer linear program (MILP) models to investigate how to best connect the core fiber network to ground stations to minimize the overall network cost. We design one MILP that can provide a quantitative value for the number of satellites needed for a given configuration and another one to optimize the allocation of the core network nodes to ground stations to minimize the overall network cost. We use these models to investigate different strategies to allocate satellites to ground stations during a satellite overpass, showing that allocating satellites based on the expected transmission requirements can reduce the number of satellites needed in a network by up to 40% compared to randomly allocating the satellites to ground stations. Furthermore, we use these models to investigate securing the data center traffic in two networks, one local European network and one global network, and show that costs in the optimal configuration can be up to 40% cheaper than simply connecting core network sites to their geographically closest ground station.