All Sections

Copper’s not dead yet: Superfast broadband trials in BT’s G.fast lab

BT has been busy rolling out superfast broadband to millions of premises across the UK. But what happens next? How far can the old copper technology be pushed? 

Regular readers will know that BT is making FTTC (Fibre to the Cabinet) broadband available to millions of homes and businesses across the UK.  

This provides much faster download and upload speeds – up to 80Mbps and 20Mbps – compared to the regular ADSL broadband most of us have been buying for the last fourteen years. 

BT boffins actually dragged a telegraph pole into the lab for G.fast tests
BT boffins actually dragged a telegraph pole into the lab for G.fast tests

As great as these speeds are, it’s unlikely that FTTC will be able to meet the bandwidth needs of future generations. 

Firstly, while FTTC is faster than ADSL, like old school broadband the speeds you actually get still depend on your location. 

As much as the top speed you can get from ADSL broadband is governed by your distance to an exchange, the maximum speeds you’ll get with FTTC are dictated by your proximity to a green street cabinet. 

In other words, while you might be able to get up to 80Mbps on paper, if you’re a few hundred meters away from the cabinet you’re more likely to get something like 50Mbps. This is the reason why ISPs like Sky and TalkTalk, who use BT’s network, advertise their top-tier services as offering ‘up to 76Mbps’ – Ofcom rules stipulate that you can only advertise the top speed of a broadband product provided that as little as ten per cent of your customers can get those speeds. 

So, why isn’t up to 80Mbps enough for the future? 

No holodecks and 3D porn for you

For an example of what this means in non-techie, real-person terms, streaming services like Netflix require customers to have at least 20Mbps of spare bandwidth to watch 4K content. An up to 80Mbps service might suffice today, but what about the future? What about 8K movies? What about all the holodecks and 3D porn the boffins have promised us? 

Because that last mile of an FTTC line is copper, it’s susceptible to interference over longer distances, like ADSL. 

Critics and companies like B4RN, Gigaclear, CityFibre and Hyperoptic either say that BT is wasting its time with FTTC and should be connecting people to FTTP (Fibre to the Premises) and/or they’re getting on with doing exactly that themselves. FTTP is a copper-free, pure fibre connection that guarantees much faster speeds that don’t degrade over long distances. 

BT however argues that it’s just not economical for it to do FTTP everywhere right now. Optical fibre is expensive. And with roughly 29 million properties connected to its Openreach network – B4RN has just over 600 customers, Hyperoptic’s network reaches 75,000 – perhaps BT has got a point. 

While BT is actually connecting some places to FTTP, it’s only available to a small number of customers. BT doesn’t plan to dramatically increase its full fibre footprint anytime soon

Huawei-made test modems sit in the corner while a spectrum analyser on the right measures results
Huawei-made test modems sit in the corner while a spectrum analyser on the right measures results

What is G.fast and why should I care?

If we’re not all getting fibre from BT anytime soon, what are we going to get in the short and medium term? Over the next few years, we may well end up with something called G.fast. 

The way G.fast (pronounced ‘gee dot fast’) works is hard to explain but in a nutshell, it’s a way of dramatically increasing the speeds possible on copper lines. It won’t solve the distance degradation issue that plagues FTTC – because it still uses copper in the last mile – but it does promise to change things dramatically.  

We were invited to Adastral Park, BT’s research and development centre in Ipswich to learn more about G.fast and BT’s ongoing trials. 

BT’s head of access network research Trevor Linney took us on a brief tour of the lab where G.fast is being tested.

The small room was dominated by a truncated BT telegraph pole with ten copper lines, or ‘loops’ as engineers call them, running across to the opposite wall where they trailled down to a range of trial modems. 

It’s here that BT is simulating how G.fast could change things for broadband customers, using copper loops of varying lengths. 

BT is testing out G.fast on a range cable types, data-lazy-sizes and lengths typical of those found in the final part of its access network
BT is testing out G.fast on a range cable types and lengths typical of those found in the final part of its access network

Linney said that in most cases, the copper last mile that connects a customer’s house to distribution points is roughly 60 metres long. Recent test results suggest that BT would be able to deliver download speeds of 700Mbps and uploads of around 200Mbps with G.fast over 66 metres of copper. 

On shorter 19 metre lines, these speeds increased, to nearly 780Mbps down and 230Mbps up. 

“The trial lines were either side of typical lengths we find on our network. Around 20 per cent of lines in the UK have less than 19 metres of copper, whilst 80 per cent have less than 66 metres of copper,” said Linney. 

Again, shorter copper loops equal faster speeds. G.fast won’t change this fact. While the majority of customers should benefit from this rollout, what about those customers who are off the beaten track? How will G.fast benefit people on lines longer than 100 or 200 metres?

“It’s too early to say,” admitted Linney. “That’s why we need to do more tests, but we think G.fast has the potential to improve speeds on loops over 100 metres too.” 

As with FTTC, speeds with G.fast decrease over distance

A sprawling nest of metal frames and cables sheathed in black plastic is situated outside of the lab itself. Linney explained that these are used in the test environment to enable easy connection of new equipment and the changing of test scenarios. Essentially, its a simulation of what’s called the ‘access network’ – the part of Openreach’s network that connects your house to the local telephone exchange. 

“The cables we’ve installed so far range from 10 metres in length to 100 metres, but these can also be linked together in ways that replicate different network topologies, giving us a capability of total loop lengths up to 220 metres,” said Linney.  

Nobody at BT was able to tell us what kinds of speeds people on these longer lines could expect – more testing needs to be done. Linney added that it’s unlikely that every customer out there would benefit from G.fast. 

This metal maze lets BT test out things like crosstalk and how speeds degrade over distance
This metal maze lets BT test out things like crosstalk and how speeds degrade over distance

Dana Tobak, managing director of the striclty FTTP-only ISP Hyperoptic said as much to us in a recent interview, so it’s not a view that’s held only by BT. 

Scales of topographic solutions

To that end, BT is also looking at something called FTTRN (Fibre to the Remote Node). This sees BT attempting to delivering speeds similar to what you’d get with an FTTC line right now (up to 80Mbps) over longer copper lines with smaller cabinets. 

It’s being trialled in locations including east London’s Tech City hub right now, where local businesses and politicians have been vocal about slow broadband speeds. 

G.fast sounds exciting but when is it coming? The ITU (International Telecommunication Union) previously said that G.fast services could be live by 2015. When or even if BT gets round to rolling G.fast services out depends on the results of its experiments – it could be 2015, 2016 or later before such speeds are available to BT customers. 

BT's G.fast trial partner is Huawei, who supplied these trial modems. Potentially, G.fast modems will be built into future Home Hubs
BT’s G.fast trial partner is Huawei, who supplied these trial modems. 

When is G.fast coming?

The BDUK (Broadband Delivery for the UK) programme should deliver superfast broadband to 95 per cent of UK properties by 2017. By then, BT could have worked out exactly how it’s going to use G.fast and who will benefit. But we won’t know for sure until the trials are over. 

What happens after that? Will BT roll out G.fast? Will we finally get pure fibre everywhere for a not-insane price? 

Alcatel-Lucent’s G.fast trial suggests that even faster speeds are possible. BT may even look to using advanced technology instead of opting to replace every last mile of copper with fibre if it works out to be cheaper. 

Though it’s expected that everybody will be on a pure fibre connection in the future, BT isn’t taking a one-size fits-all approach today. 

As everything is very much in the trial phase right now it’s impossible to predict when G.fast will arrive, what speeds we could get and how much it would cost.  

Comments