Cable length minimisation in long-reach-PON planning for sparsely populated areas

Tackling the digital divide has become one of the main goals of many governments around the world, and a main target for the European Commission. Fibre-to-the-premises (FTTP) is the only access technology that has the potential to totally eliminate the digital divide since, within access distances, fibre capacity can be considered distance independent. In practice, however the higher cost means that many rural areas might not be served by FTTP services, further widening the capacity gap between rural and urban access. However, while fibre roll out is certainly less cost-effective in rural than urban areas, a proper deployment strategy that minimises cable installation costs can increase the network profitability and thus the rural FTTP coverage. In this paper, we propose a novel agglomerative clustering algorithm. The algorithm together with ILP and heuristic models, attempt to minimise the total fibre cable required to connect sparse rural users to a drop PON splitter located up to several kilometres away from the user. Our results show that proper deployment strategies can increase the use of large last-stage splitter, thus increasing sharing of fibre among users, while still minimising the overall cable length. In addition, we show that our heuristic is able to obtain similar results as the ILP (i.e. within an 8% margin), but can reduce the computation time by about 6 times, for the scenarios we have considered.