Whether it is called Next Generation Access (NGA) or Ultra Fast Broadband (UFB), the same principle applies, the only way to truly achieve a long term sustainable and fully scalable architecture is through the use of fibre optics.
In many territories there is much being made of Fibre Broadband already being available, but this in many circumstances is a classic case of ‘smoke and mirrors’ since what this typically means is an incumbent operator leveraging its existing copper infrastructure and seeking to sweat its existing assets for as long as is absolutely possible. There is truth in the fact that fibre is involved though, since these architectures may be better known as fibre-to-the-cabinet (or FTTC), whereby the reach of fibre extends much more deeply into the network through to the street cabinet level. From the street cabinet, the final premises connection is then via the existing copper telephone wiring, using increasingly complex and sophisticated Digital Subscriber Line ( DSL) techniques (including higher order modulation schemes, compensation/cancellation techniques, pairing and vectoring etc). These enhanced DSL techniques can offer substantial end user data rates of up to 100MBs, though this is rarely achievable since the distance from a street cabinet to a home can still be easily 500m or more, and many factors come into play including attenuation and crosstalk, the quality of the copper (which is sometimes in fact aluminium) and any intervening connection points, and even water-logging of telecoms ducts. To a certain extent these factors can be mitigated with fibre-to-the-pole (FTTP), where an aerial copper infrastructure exists, since the distance to the home is then generally 50m or less, but this then requires a much deeper fibre trunk reach into the network and substantially higher average costs than FTTC since an active unit must be located on each pole.
Whichever way it is looked at though, copper telephone pairs (particularly given their poor twist ratios and cross-talk performance) have a fundamental physical limitation to the data rates that can be reliably transported, and the higher the data rates you wish to pass then the distance you can cover becomes increasingly shorter.
Fibre offers essentially unlimited data throughput, with the data rate being determined by the electronics driving the fibre, and this of course is continuously improving. Once an end-to-end passive fibre infrastructure has been deployed then as time goes on only the equipment driving it may need replacement to cater for increased data rates.
Fibre-to-the-home (FTTH) is then the ultimate solution, delivering a fibre cable from the operator’s main point-of-presence (POP), or distributed mini-POPs, directly into a house, permanently removing the current or future bottle-neck of a copper connection. There are two primary physical/passive layer techniques employed for FTTH:
In the Passive Optical Network (PON) solution, passive splitters are used, which allow one fibre to feed multiple properties and the active electronics layer then time division multiplexes the available data rate among the connections. The passive optical splitters are typically field based and situated in close proximity to the premises. This is a cost effective solution since it minimises the number of fibre cores needed to be carried back to the POP. Historically it has been common to see passive split ratios of 1:32 (32 premises fed by one inbound fibre), however as expectations for future required bandwidth continue to increase then we are now seeing in-field split ratios of 1:8 being deployed in order to give the active electronics a longer asset life. In this case four 1:8 split inbound fibres can be combined at the POP with a 1:4 splitter to achieve effectively a 1:32 split, this can then be reduced to a 1:2 split in the POP at a future time to achieve effectively a 1:16 split, and then finally all splitting at the POP being removed to give the final 1:8 split. Combining or subsequently removing the splitting at the POP is easily achieved and managed. Current mainstream GPON active technology provides 2.5GBs download per port in the POP split among the connections, so over 32 homes this still provides a fully uncontended >75MBs download each (remembering that all internet networks are actually contended at some point). Already technology is available (albeit at a higher price) which can deliver 10GBs per port in the POP, and upcoming mainstream technology will also allow wavelength division multiplexing (WDM) which will further dramatically increase data rates, these new technologies can simply be installed over the same passive network infrastructure when needed.
In the Home Run solution, a separate fibre is taken back to the POP from every individual premises. So then each home can be fed by a separate active electronics port, with no bandwidth sharing (as is the case with PON). Some proponents advocate Home Run since it imposes no limitations, but as always there is a trade-off, since a lot more fibres need to be laid and so overall the passive fibre layer costs (initial CAPEX) tend to increase. This can be mitigated by having more localised mini-POPs (essentially active street cabinets), but these and their necessary power connection also have a cost impact (both CAPEX and OPEX). For active electronic ports, then 100MBs and 1GBs are already standard fare, though the active cost per port is higher than a shared PON system port.
A hybrid of the two models is often considered, where most domestic premises are fed via PON, but sufficient ‘spare fibre’ is allowed at local distribution nodes to patch in a direct home run where required (e.g. for businesses and more critical application needs). Some operators have sophisticated schemes where two or even four fibres are fed to each location, and each can be patched into a separated passive network, selected by PON or home run, allowing multiple services differentiated from multiple service providers. This approach may well become a more common trend, though there is of course a notable initial CAPEX increase aspect that must be justified by the increased flexibility and increased potential ‘sales’ factor.
Given that it is the ultimate solution, why is FTTH not being deployed everywhere already? FTTH requires a complete new core/trunk/distribution fibre network to be deployed, and then a new fibre drop being run into every location, this is a highly capital intensive operation, with long build lead-times. The return-on-investment (ROI) is the key criteria here for a private operator, institutional investor or joint venture, and this is predicated on a number of factors, cost to build and then OPEX, balanced by take-up percentage and average revenue per user (ARPU). Take-up rates have been previously lower and slower to grow than predicted, but more recent initiatives like Google’s FTTH in Kansas are starting to set new trends in stronger take-up, so more privately funded activity is expected and FTTH is very much starting to come of age. Municipal and state fibre broadband projects have had some notable successes, a good example in Europe being Stokab FTTH for the city of Stockholm, which already has 60% of homes in the city able to get 100MBs, and on target to achieve near 100% coverage by 2020. Stokab has grown slowly and steadily over a very long period, and more or less always been profitable too, it is a model that many municipalities could do well to study carefully.
