The G.fast trials that Openreach is running were announced as having
connected the first customers recently and now we can reveal the first speed
test that we have seen across a G.fast connection.
G.fast ultrafast broadband from Huntingdon trial
We have seen speeds like this before from areas where native FTTP from
Openreach is available and people have opted for the fastest service or in some
cases people who got a Fibre on Demand order in before the order book closed,
and as the G.fast product is modelled to match the fastest current Openreach
FTTP product it is no surprise we see similar speeds. The test was tracked back
to a street where Openreach is running its G.fast trials, and in time we might
even see the Gigabit FTTP product appearing in our speed test data too.
It should be noted that we are not expecting every G.fast connection to hit
these sort of speeds if and when it is launched as a product, part of the
trials process will be seeing how the new technology performs over distances in
the real world and also how long it takes engineers to install the G.fast
nodes, its fibre backhaul, power and small copper interconnects and then feed
this into the models to determine which will be the most cost effective to
roll-out.
What is interesting is that the best Wi-Fi routers on the market currently
max out at around 266 Mbps download speeds in the real world (as opposed to
marketing claims on packaging), so with the new wave of ultrafast products
speed demons will need to remember to avoid wireless if the ultimate speed
thrill is their goal.
In the speed battle for advertising Virgin Media will hit the market running
before G.fast hits any volume, but unless Virgin Media boosts its upload speeds
G.fast looks set to be the preferred choice for those wanting more upload
speed, before the ultimate step of symmetric FTTH.
Test your broadband speed
Follow us on X for the latest broadband news 
Was there any mention of cabinet based G.Fast in this trial?
Not so far, housing in a cabinet should be simpler than the pavement chamber solution they are working on in Huntingdon.
Once they do the press tours of funny bits stuck in the ground we have much better answers.
Nice speed for a one-off, though the upload looks a little bursty.
Any idea how the test went latency-wise?
Latency wise – yes a hint in the image
“What is interesting is that the best Wi-Fi routers on the market currently max out at around 266 Mbps download speeds in the real world”
Rubbish. Real speeds can push 800 Mb/s (4-stream) or 600 Mb/s (3-stream). I get well over 400 Mb/s with a 2013 Homehub 5 and Macbook.
Haha. A one-off figure – but I tend to find such figures on speedtesters to be meaningless.
Latency over time, and consistency, would be a much nicer measure.
@WWW
I don’t think g.fast itself will make a significant difference to latency unless a time based interleaving has to be used to deal with impulse noise (hopefully g.inp will render that unnecessary). Beyond that point it’s surely more a matter of the backhaul than anything else.
I should add one point that might affect latency (in theory). g.fast uses TDM rather than FDM for duplexing. The impact that would have on latency would therefore depend on how frequently the system switches between upstream & downstream operation.
All g.fast lines in a bundle will have to be synced in this respect or there would be dreadful x-talk issues.
TheEulerID – the split between upstream and downstream is decided by the G.fast node serving the bundle, there’s no ‘sharing’ of cable bundles between nodes so should be alright.
@dixinormous
I thought it would have been obvious I knew that by implication. What I was referring to was that it wouldn’t be possible to change the upstream/downstream TDM balance between different lines in the same bundle, which would clearly be a desirable, but not feasible.
Regarding upstream capacity on Virgin there is plenty that can be done to increase it and, indeed, work is in progress to do so.
http://www.arrisi.com/data1/content/D3-1_Migration.asp
@Euler
True about the limitation to the downstream/upstream split, plus the impact of TDM on latency.
I wonder whether BT intend to make any kind of variation in the split; I can think of two examples where they might do something:
a) On business parks, they might deliberately go for a 50:50 split
b) When all premises on the node are very close, they could alter the split to allow extra upstream without compromising the ability to hit 300Mbps downstream.
the question is, in the trial are they pushing nodes out 100s of metres away from the cabinet.
chrysalis, I think they are trialling different set-ups, some nodes next to the cab, some further out. I expect the preference being next to the cab.
@WWW
I suspect the TDM intervals are set such that it doesn’t cause a significant increase to latency, but I can’t actually find any reference to it.
As far as special cases go, then there may be some scope in “fibre-to-the-basement” installations to provide higher speeds and a change in upstream/downstream mixes as, presumably, line lengths will be shorter than street deployments. Possibly business parks too, but some of those might have quite variable length lines.
Of course having many variations of service is difficult from both the configuration management and marketing point of view.
I imagine Openreach will be very limited by the need to produce a standardised product for their customers to consume.
It’d be interesting to see if there are any detailed reports on progress at the BB World Forum next month!
A relative has received a letter and email from BT offering 6 months free Ultrafast in the Huntingdon trail area – does anyone know if this is really “free” and what happens at the end of the 6 months (are they charged the going rate for the ultrafast, or do they go back to their existing package – which includes some features not available on the currently sold ADSL2+ packages that they’ve currently got)
They’re a bit hesitant to contact BT about it as at first they weren’t convinced it was even legitimate (not being in to broadband news and not knowing about “ultrafast” or a “trial” even taking place, and not even on FTTC yet) and didn’t want to just get signed up to something and then find out the bad news of a huge bill increase at a later date.
One of the messages (I think the email, not the letter) suggested that the existing connection would be left active, but I’ve not seen the letter/email to be 100% certain on what was said.
there is definitely a trial in the Huntington area. it is free. as to when it ends? plan is 6-9 months. actual end depends how it goes. will it continue to be available as commercial product? too early to say. enjoy it while it lasts and hope!
https://www.youtube.com/watch?v=LLpk2dz6nBQ