G.Fast Broadband Technology and FTTP – What are we waiting for?

BT Openreach have had broadband technology capable of 330Mbps plus available for years now. Why is no one using it?

You have to give credit where it is due. BT Openreach have done quite a job getting the most out of a telephone cabling infrastructure that was designed decades ago. BT engineers on some exchanges are working with an almost unbelievably creaky set of cabling put in the ground before the advent of colour TV. Aluminium wiring, paper sheathing, and barely waterproof ducting – you name it, this is not cable that any self-respecting data-comms engineer would touch with a barge pole these days, but, it is what it is. That is what there is and that is what BTOR get to use.

The fact that BT have managed to roll-out 80Mbps broadband to the majority of premises in the UK over this junk is not to be sniffed at. And they are aiming to go much further. BT Openreach’s work on G.Fast is close to roll-out and this will potentially offer another step change over current FTTC speeds. G.Fast will run over copper for the last mile, just like FTTC, but using very clever compression and error correction algorithms, be able to offer up to 330Mbps and then 1Gbps on a copper pair. And this roll out is only months away now.

There are 2 ways to see this. The first is to say “yeah, big deal, Virgin have that type of speed already”. The other is to recognise that this is a major change in how consumers, and more so, businesses will be able to get connectivity. BT copper is everywhere. Virgin is not. BT has a huge ecosystem of partners providing services to all types of customer across the spectrum. Virgin is monolithic and its business cable services are roundly seen as sub-par and have not been widely adopted.

Before we move onto the prospects for G.Fast, let’s also talk about FTTP. This is Fibre to the Premises. A pretty ambitious proposal to push fibre out all the way from the exchange to the client’s premises to delivery broadband services. So there are few things to say here. First off, yes, this is a great idea. It totally nails the cable legacy issue forever and is the most future-proof way of developing the network, not just for the next 3-5 years but for the foreseeable future. But, wow, is this a big ask. Delivering new fibre to premises in the UK is a mission. Just look at how hard it is to get leased lines installed. Now it is true to say that BTOR are going at this is a completely different way to installing fibre for Ethernet leased lines, but it is still a massive undertaking – both very slow and hugely costly. But, where it is available, then it is potentially terrific and BTOR are at least doing their job as a near-monopoly of last mile cabling and ploughing lots of time and effort into this, so good luck to them.

So, that’s it right? We are surely all set for another revolution in our broadband experience. Goodbye 80Mbps FTTC, Hello 330Mbps G.Fast or beyond!

Well, not just yet I’m afraid. The thing is that although the last mile is of course a big factor here, it is not the only one. It’s necessary but not sufficient.

The missing piece is the carrier networks and their bandwidth-charging model. ISPs are buying their services in the main from BT Wholesale or TalkTalk, with a few other players along for the ride. And ISPs don’t just pay for the “tail” (the last mile delivery back to the ISP’s network). They pay as much or more for any given connection for the bandwidth used on that connection, in an aggregated feed usually. And this bandwidth does not come cheap. In fact the price for 100Mbps of available aggregate bandwidth will set an ISP back anything from £1000 to £5000 a month. So if you are an ISP, you are managing your network fine and you have enough spare bandwidth on your network to accommodate a sensible amount of burst on your FTTC and ADSL users. All is ticking along. Now, you want to enable a G.Fast user on the network, with access to 330Mbps of bandwidth for just one user on his tail. And he [DH2] expects to see something like that number when they do a speed test. Just having that bandwidth available is going to cost you something in the order of £600 a month. But how much can you sell this for? Well, maybe £80-100 a month if you are lucky. So it is a huge money pit. Add to which the fact that customers prepared to pay, 2 or 3 times the price of an FTTC connection are likely to want to use that bandwidth on a regular basis.

Now sure you get some aggregation effects as you take on more, but the bottom line is that in ISP-land, the current wholesale bandwidth models are not set up for these bandwidth numbers. They are really a legacy of the days of ADSL. In those days the range of bandwidths in use was much smaller and the whole thing was much more manageable. If ISPs tried to pass on the carrier bandwidth costs for these services it would make them unaffordable in most cases, and so they remain a very niche offering, and where you can find them, they are really pricey and they come with various caveats about the available bandwidth.

So, in the end this is down to the wholesale carriers – the likes of BT Wholesale and TalkTalk Business. Once they decide that they want to see adoption of these services, and they provide approachable bandwidth charges to ISPs, then I am sure they will take off really quickly. Sure they won’t be symmetrical and sure they won’t be as reliable as Ethernet leased lines, but they will be a fraction of the price.  And just as the introduction of FTTC took a big bite out of Ethernet market share, G.Fast and a FTTP will do as well, maybe even more so, and businesses will get used to having 300Mbps or even a Gig on tap.

Until then, watch this space and consider bonded FTTC as an interim solution for the medium term.