Openreach Deploy G.INP to Boost UK ECI FTTC Broadband Lines
Openreach (BT) has finally succeeded in deploying ReTransmission (ReTx / G.INP) across their ECI (Ribbon) based UK estate of Fibre-to-the-Cabinet (FTTC / VDSL2) broadband ISP cabinets and customer lines, which comes after many years of trying to get the technology to work. The result is faster speeds and improved stability.
The news these days is typically dominated by deployments of full fibre (FTTP) broadband lines, which makes it easy to forget that around 15 million lines are still using Openreach’s previous hybrid-fibre (FTTC) technology, and so any developments on this front remain very important.
One such enhancement is called G.INP (ITU G.998.4), which is an error correction solution that was designed to help resolve spikes of Electromagnetic Interference (impulse noise) and thus improve a line’s stability, as well as its performance. The introduction of this technology can, on some VDSL2 lines, even produce a small to modest increase in service speed (downloads).
Advertisement
The technology is nothing new to Openreach, which has already got it to work on their larger estate of Huawei based cabinets. But unfortunately the same cannot be said for those using kit from ECI. Various past attempts have been made to fix the ECI issues and all have run into problems (examples here, here and here).
Suffice to say, deploying G.INP to ECI based hybrid-fibre broadband cabinets has proven to be considerably more complex than was the case with Huawei cabinets, which is due to a mix of issues (e.g. problems with modem / router compatibility within homes and DLM compatibility).
What’s new with G.INP
Last spring (2023) a number of our readers started reporting that G.INP had finally been enabled in their ECI dominated area again, although at the time we opted not to cover this due to the long stop-start history of related trials. Not to mention that this was only occurring in some ECI areas and not all.
However, the good news today is that Openreach has finally identified the specific combination of parameters that had the potential to trigger the issues seen during the previous trials, and they’ve now reconfigured their Dynamic Line Management (DLM) system to avoid this combination.
Advertisement
As we understand it, both ISPs and Openreach then agreed to initially commence a low volume deployment in April 2023, which unlike previous trials appears to have gone without any hiccups. The operator’s initial deployment was then accelerated in June 2023 and in February 2024 it finally reached completion (i.e. the entire ECI / Ribbon estate of cabinets has now been upgraded, including SOGEA ones).
The Results
According to data seen by ISPreview, the average improvement in unconstrained lines (i.e. those not achieving the maximum speed their chosen speed tier allows) is 3Mbps and there has been a 25% increase in the number of lines achieving their “up to” speed (i.e. those achieving the maximum speed their chosen speed tier allows).
Line stability has been improved by the upgrade too, which also means that error protection is now always-on and lines are thus better defended from unwanted noise (interference). The trade-offs with the previous approach meant that error protection was only enabled if a line suffered instability, which in turn meant that lines often had to suffer that period of instability before protection could be applied.
Finally, network latency (response time in milliseconds) has also been significantly improved. For example, the previous type of error protection introduced an 8ms, 16ms or 24ms delay when applied. But ReTransmission has no such inherent impact on latency when applied, so lines previously using the former error protection will see a reduction in latency of 8ms – 24ms.
Advertisement
In summary, it’s great to see that Openreach are still willing to spend the time and money to find ways of squeezing extra bits of performance out of even older technologies, particularly those that have proven to be such tough nuts to crack. The operator has persevered because G.INP offers considerable benefits to the many customers still using an FTTC product from a Ribbon DSLAM / cabinet.
However, it is important to remember that there are a very small number of VDSL2 modems (these are often built into your broadband router) that don’t support G.INP / ReTransmission, so these are excluded, as is already the case on those covered by Huawei’s cabinets. Some more problematic or faulty lines may also see G.INP being disabled.
Mark is a professional technology writer, IT consultant and computer engineer from Dorset (England), he also founded ISPreview in 1999 and enjoys analysing the latest telecoms and broadband developments. Find me on X (Twitter), Mastodon, Facebook and Linkedin.
« CityFibre Start Project Gigabit Broadband Build in Cambridgeshire UK
Latest UK ISP News
- FTTP (5898)
- BT (3615)
- Politics (2672)
- Business (2392)
- Openreach (2381)
- Building Digital UK (2308)
- Mobile Broadband (2095)
- FTTC (2073)
- Statistics (1866)
- 4G (1768)
- Virgin Media (1719)
- Ofcom Regulation (1551)
- Fibre Optic (1451)
- Wireless Internet (1442)
- FTTH (1384)
Brilliant news.
When I saw this, my first thought was that something must’ve gone wrong with the website’s sorting system and now it’s showing an article from 2019 on the front page. Am I the only one?
will be watching with interest. I happen to have two VDSL lines, one on each vendor, and the ECI has always been the worst of the two with a noticeably lower sync speed.
I suppose VDSL is not going away any time soon, so if they can do this quickly/cheaply then why not.
Just as the UK moves to FTTP, BT finally start deploying a fix.
They knew for years that ECI was a major headache yet they continued to deploy the kit. BDUK used Huawei cabinets for those deployments and they have been pretty much rock solid despite the alleged noise around getting rid of it later. ECI kit on FTTP also had issues limiting to 330mbps.
At least this fix can actually benefit a large number of people. I have less truck with this than anyone still pushing G.fast that’s for sure
For “finally deploy” read: “spent years trying to get it to work”, as anyone who’s read the article (and previous ones) can see. I guess we can’t let the facts get in the way of some BT bashing.
I’ve seen ECI cabs in subsidised areas. If it delivered the contractual obligations (which I think was 30Mbps or higher), of course they’d have used it.
@Cognizant
I recently had g.fast installed – what’s wrong with it? I am synced at 160/30 and I get 149.7/24.5 on speedtests and a bit higher on my ISP’s speedtester. (155/26)
Is FTTC now going to give me over 100 before I was lucky to get 67 and that is on ECI kit – genuine question.
@THATisCuckoo
If you can get and make use of G.fast, then that’s great. There isn’t anything inherently wrong with it. My annoyance with G.fast is that it really only benefits those who can get 80Mbps anyway, and then as the distance only increases slightly, the speed drops off a cliff.
FTTC VDSL2 will never give you above 80Mbps, so I’m happy for you that you can get 150Mbps. G.fast has long been seen as a technology to sweat copper for as long as possible though, it is time to move on – and thankfully that is happening…
Ivor, yes they have spent years getting it to work, I agree. The facts are BT selected that troublesome ECI kit and continued to deploy it way after having issues when BDUK contracts used the Huawei kit that just worked. Yes, there was political noise about removing Huawei but this was way after and not happening anyway for FTTC.
Cognizant – only because competition is forcing BT to do something now, else they would still sweat copper and say G.FAST is wonderful and doing a bigger roll out of that.
We might be going to FTTP but FTTC is probably going to be hanging around in some places for at least the next 10 years or so I guess. In my town (Banbury) Openreach have done areas with Poles or underground estates built in the 20 years (with modern ducts), but so far missed out on estates that were done in the early 1970’s and are probably not ducted. It will remain to be seen whether these will ever be viable to upgrade to FTTP
Its been on/off before 2016!….must have cost a fortune in time to resolve.
“I’ve seen ECI cabs in subsidised areas. If it delivered the contractual obligations (which I think was 30Mbps or higher), of course they’d have used it.”
Have you though? Or did you just assume they were subsidised?
Only a tiny handful of the very very earliest publicly funded cabinets were ECI, during the earliest trials.
Some of those aren’t even around now.
Every single publicly funded FTTC cabinet since then has been a Huawei cabinet.
Every BDUK contract went to Huawei.
There aren’t really subsidised “areas”. It’s done on a per cabinet basis, with some lines moved between cabinets as part of the work. Sometimes additional “infill” cabinets were installed, again always Huawei.
None of the ECI cabinets had 30Mb/s contractual targets.
BT, once again closing the gate after the horse has bolted… Still, I’m on VDSL. Jurassic fibre dug and ducted, then failed to deliver active fibre. Virgin’s plumbing in XGS-PON as we type. BT, probably glacially sliding over the finish line sometime in 2028.
meanwhile the actual data shows that OR’s FTTP footprint dwarfs everyone else combined and their build rate makes everyone else look laughable too.
What is it about ISP and mobile coverage that causes people to assume that their own circumstance is the norm?
Sorry, I don’t understand. Do you think they should just give up on improving services for their customer base who can’t access FTTP yet?
“improve” is a very ambiguous word. I see Altnets and NexFibre/VM going to improve FTTP a lot better by offering symmetric services for those that can get it, which is growing all the time. I see BT’s improvements as the minimal they can get away with…
Sure. And Local Authorities should never fill in potholes or resurface or introduce traffic flow or safety measures. They should only build brand new roads. Honestly some of the comments on here are absolutely absurd!
Don’t they also do the minimum work for outrageous cost too????
I was curious why my download connection speed increased by around 10% (after 5 years at the same speed) a while back but the very limited view I get of my line stats hadn’t altered.
Probably your neighbours are all going to FTTP and the crosstalk is reducing as a result
Yip. I wonder what happened to vectoring with higher profile BT – oh, yes too much work for you and as you said for years prior to FTTP, nobody needed over 80mbps (despite most getting around 30-36mbps on average).
Nope, no FTTP available yet – but soon !
Last time my eci line saw it was 2016! on some trial of a trial back then…
@anonymous you are completely obsessed with Openreachs FTTP upload speeds. We understand you don’t like them and the way they do business, but have you been stuck on ADSL 1 meg for years or something? If not, please get over it.
Not if it annoys a fan boy…
That is interesting. I was switched to DV in September. Early December my download speed, (traditionally a stable 58 Mb/s over the last 10 years), suddenly increased to 61 Mb/s and has been stable there ever since. I assumed it was the DV but perhaps not?
Would be nice if Openreach start to apply vectoring in all ECI and Huawei cabinets to avoid crosstalk as I know G.fast have vectoring except most FTTC have vectoring see here: https://www.increasebroadbandspeed.co.uk/vdsl-vectoring-arrives-on-a-rural-fibre-broadband-cabinet
Well G.Inp is a switch, you flick it on (line by line) and it applies the code (yes there is some SERIOUS work involved getting the code to work), but for a handful of highly skilled people.
Vectoring (which has never been supported by ECI full stop) requires addional hardware at each cab, engineer time to set it up, a handful of issues that need resolving as a result of the change etc etc
The 2 are hardly comparable. The cost of vectoring would be astronomically more – I’d rather they spent that time and effort deploying full fibre
The ECI cabinets can’t do system level vectoring.
Board level vectoring (line card) isn’t efficient and wouldn’t work from Openreach’s topology.
Given the limited life span of FTTC does this really make sense
Fttc will probably be around for the next 8-10 years until it finally goes so I guess the answer is yes.
It’s effectively a software update, so low deployment costs for good gains = yes it makes sense.
Given BT’s strategy of wanting to migrate people to FTTH I would assume they will only deploy it to location not in their current FTTH build program
Yes, FTTC will be around for a long time.
How does one find out if G.INP is active on their FTTC connection? We have ~20Mbps download speeds, so I would take any improvement I can get!
You need a CPE with a modem that can access detailed line stats like a Billion 8800NL.
“In summary, it’s great to see that Openreach are still willing to spend the time and money to find ways of squeezing extra bits of performance out of even older technologies, particularly those that have proven to be such tough nuts to crack.”
Frankly as seen from an engineer on these things? G.INP in 2024 when they could have had vectoring, 35b profiles… just sounds like “ECI equipment is rubbish but we need to let it survive a little longer.