E A B C D R F F F F F F F F F F F FP1 FP2 subscriber Cable type 1 Cable type 5 Cable type 4 Cable type 3 Cable type 2 α n houses in a row l = distance between two houses The central office at the center This serves n 2 customers Fig. 1 Triangle Model (TM) Fig. 2 Simplified Street Length (SSL) model Fig. 3 FTTx Designer Framework Geographic Model for Cost Estimation of FTTH Deployment: Overcoming Inaccuracy in Uneven-populated Areas Attila Mitcsenkov 1 , Miroslaw Kantor 2 , Koen Casier 3 , Bart Lannoo 3 , Krzysztof Wajda 2 , Jiajia Chen 4 , Lena Wosinska 4 1 Budapest University of Technology and Economics BME, Magyar tudosok krt. 2., 1111, Budapest, Hungary, mitcsenkov@tmit.bme.hu 2 University of Science and Technology AGH, Al. Mickiewicza 30, 30-059 Krakow, Poland 3 Ghent University - IBBT, Gaston Crommenlaan 8, B-9050 Gent, Belgium 4 The Royal Institute of Technology KTH, School of ICT, Electrum 229, 164 40 Kista, Sweden Abstract A geographic approach is proposed to accurately estimate the cost of FTTH networks. In contrast to the existing geometric models, our model can efficiently avoid inaccurate estimation of the fibre infrastructure cost in the uneven-populated areas. Introduction Fibre to the home (FTTH) has been widely recognized as a future-proof solution for access networks due to its capability to meet the increasing bandwidth demand of the end users. On the other hand, the deployment of FTTH networks is very costly and an accurate estimation of the investment cost is of high importance. Several models have been developed to estimate the deployment cost of FTTH network, e.g. a number of geometric models [1, 2]. The geometric models are used to design the fibre infrastructure based on a set of parameters describing the considered area, e.g. average values for population density, distance between end users and central office (CO), and give an input for cost estimation based on the designed infrastructure. However, these models were optimised for the computing capacity in the 90’s, which was much more restricted than today. Especially when applied to areas with an uneven user population, they suffer from an inaccuracy problem since they consider only the average values. To address this problem, we propose a geographic approach based on the real and detailed geospatial data to design the FTTH outside plant infrastructure in order to accurately estimate the deployment cost. A case study is carried out and it is shown that there is a significant difference between the results obtained by our approach and the geometric models. In this way we are able to quantify the inaccuracy caused by the geometric models. Review of the geometric models Geometric models make an abstraction of the installation region and parameters and have an algorithmic or mathematical approach for calculating the trenching and fibre length. Typically the abstraction assumes a uniform subscriber population density and recursive area structure. In practice, the areas where FTTH networks are deployed are not evenly populated and the fibre trenching is constrained by various local conditions, e.g. parks, railways or highways. This is a reason why the geometric models cannot contribute to the accurate estimation of the deployment cost. In this paper we take two geometric models as examples to compare with our proposed geographic model. Triangle Model (TM): This model is a polygon based model for the access network [1]. Fig. 1 illustrates the model showing polygonal structure surrounding the hub and representing the distribution area. The dispatching boxes FP1 (boxes B, C and D) and distribution cabinets FP2 (points F) are symmetrically located at the gravity centres of the elementary triangles. Length of the cables considered in TM model is calculated using the following formulas: |AB| = |BC| = (R/3)∙cos(α/2) (1) |CD| = |CE| = (R/6)∙√1+8∙sin 2 (α/2) (2) |DF| = R∙(0,132+0,336/n) (3) where n denotes the number of fibre cables leaving the hub. The average distance b between the branching box F and building entrance gives the formula: b=2/3∙√M/(π d) (4) where M denotes the number of potential users per branching box while d is the number of potential users per km 2 . The total trenching length and fibre length can be obtained by summing up these for all triangles. Simplified street length model (SSL): In this model, the potential customer base is uniformly distributed over a squared area (see [2]). One side of the square contains n houses and the distance between two houses is indicated by l. The CO is always situated in the middle of the square. Equations 5 and 6 express length and fibre length respectively. l = n∙(n-1)∙l+(n-1)∙l = (n 2 -1)∙l (5) n-1 F= 4∙l∙∑ [min(i,n-i)∙(n-i)] (6) i=1 Geographic model In this paper we propose to use the real and detailed geospatial data instead of geometric models defined by some average and aggregated parameters. In order to handle the uncertainties and uneven character of the parameters describing the considered service area, Design optimal topology Build network model Process geographic and infrastructureal information CapEx Calculations Business Case Analysis