An optimal strategy for FTTH deployment at the scale of a country

International audience To follow the emergence of new broadband services, key actors in telecommunications are pushing towards the migration from copper to fiber-based access networks. Statistics released by the FTTH Council Europe in February 2018 [1] show that the majority of Middle-East countries have less than 25% FTTH/B network coverage. Many of these countries aim at reaching high FTTH coverage rates in upcoming years. Nevertheless, deploying FTTH everywhere requires huge efforts and costs, and is often difficult to achieve at once on large territories. Sometimes, hybrid technologies, like Fiber-To-The-Node/Curb/Building, provide alternatives offering performances close to that of FTTH. Such technologies are generally less expensive to deploy and often provide a faster return over investment for areas with existing copper infrastructures. For Greenfield areas, the state usually promotes and finances FTTH construction works. All this makes large-scale fiber deployment a complex task that must consider not only financial constraints, but also technical and sometimes political restrictions. Operators are always looking for efficient methods to optimize their deployment decisions. The challenge lies in building an optimal strategy, for hundreds of areas within a country, to progressively migrate from copper-based to high-speed networks using hybrid and/or FTTH technologies. This strategy consists of scheduling fiber deployment on a set of candidate areas over a time horizon, while identifying a suitable technology for each one. The deployment plan must respect the investment budget and the annual projects quota. We propose an innovative approach, developed in partnership with Tunisie-Telecom, which relies on an economic model that optimizes the ROI over a time horizon. It calculates a fiber deployment plan and estimates for each area the required investment, the list of necessary equipment/civil works and the expected revenue. The problem is formulated using Linear-Programming while integrating technical and financial constraints, a cost model and a revenue model. As input, we find the areas to be deployed, the population to be connected and information about each area's topology. The cabling cost of a candidate area depends on the chosen technology, its population, geography and urbanization. It considers existing infrastructures and includes construction costs for Greenfield cases. The revenue of a candidate area is estimated from its population and the penetration rates of the different proposed commercial offers. These penetration rates result from the marketing strategy and vary with the socioeconomic category of the area.QoS Design integrated the developed approach in a newtool. Itwas tested on a pilot project, with Tunisie-Telecom,including seven candidate areas. In our tests, we noticed that connecting subscribers with FTTHinstead of FTTCin areas with existing copper infrastructures costs, on average,three to four times as much.For Greenfield areas, FTTH and FTTCcosts are generally comparable since there are no copper ducts, and FTTH deployment is often chosen. Moreover, areas with dense populationsand significant consumption profilesare generally promoted in the early years of the planning horizon. Variousresults have been concluded for different scenarios. The proposed deployment plan has been validated and the tool will soon be adoptedby Tunisie-Telecom.