Authors: Dragan Ilic, Stephan Jay, Thomas Plückebaum, Peter Stamm
Summary
Cable network operators have multiple options at their disposal to adjust the network to rising bandwidth demands and increasing IP service penetration. Within this study primarily the costs of measures that rely on the extension of fibre networks and a reduction of coaxial segment size are analysed. These network upgrades increase the shared bandwidth available to each customer. Further options, such as the extension of the frequency range used or the migration to more efficient transport technologies, are not analysed. Nevertheless they additionally expand the ability of operators to increase bandwidth in their networks.
Cable network operators can flexibly react to increasing bandwidth demand by re-segmenting customers on the coaxial network and adjusting the allocated up- and downstream domains. This does not even require an extension of the fibre depth or adjustments to the main coaxial network. Such an extension of the fibre network leading to more distributed fibre nodes ("Deep Fibre architecture") allows to reduce coaxial segment size even further and provide higher bandwidths to individual customers.
To estimate the investment for deploying cable networks in Germany at current cost, a modified version of the bottom-up WIK NGA cost model is applied. Approximating the actual coverage in Germany of about 60 % of all homes this results in an investment of roughly 33 billion € which is about 10 % less than WIK's modelling results for an FTTH/P2P network in the same coverage area (both values based on greenfield deployment). This result enables to deduct (by subtraction) the study's main target: the incremental cost of migration from the basic "Full Service DOCSIS" cable architecture to higher performance architectures. The additional investment to migrate from Full Service DOCSIS to Deep Fibre are relatively small. The modelling results show 50 € to 150 € per home passed or 80 € to 240 € per home connected.
The fibre-only "RFoG" architecture (comparable to FTTB) scales relatively poor for wide area deployments. It requires very high investment in CMTS, driven by limited divisibility of domain provisioning for fibre nodes. The economic analysis suggests that RFoG is currently only relevant for large multi-dwelling properties but not for widespread deployment combined with DOCSIS based production.
Analysing the monthly cost per customer on a current cost basis for Full Service DOCSIS and Deep Fibre shows that costs are considerably above monthly revenues in the relevant range of service penetration. Cable networks valued at current cost would therefore not be profitable. Considering replacement investments required in the medium term this should increase the need of cable operators to increase their average revenue per user.
Discussion Paper is available for download.
