» ISP News » 

UPDATE A Quick Update on BT’s G.fast and FOD2 Ultrafast Broadband Trial

Thursday, September 17th, 2015 (8:46 am) - Score 4,239

At the end of last month BTOpenreach began connecting customers to their large-scale 6-9 month trial of 500Mbps capable NGA2 G.fast broadband technology and 1Gbps Fibre-on-Demand (FOD2) in the market town of Huntingdon (Cambridgeshire). Now we have the first official progress update.

The update (thanks Carl), which stems from a joint presentation given by both Ian Boothman (Openreach) and Dawn Devonald (TalkTalk) to the UK Network Operators’ Forum (UKNOF), doesn’t reveal much that we didn’t already know (see here for details) or couldn’t have guessed. Never the less there were a few bits of useful information.

As a quick recap, G.fast works in a roughly similar way to the current 80Mbps capable hybrid-fibre FTTC (VDSL2) technology that dominates the UK market (ISPs often market this as “fibre broadband“), although it requires significantly more spectrum (G.fast 106MHz+ vs FTTC 17MHz) and thus operates best over a much shorter run of copper cable (ideally less than 350 metres).

The technology’s Nodes / Distribution Points (DP) could be installed inside / alongside existing street cabinets, as well as underground (i.e. similar to existing NGA Aggregation Nodes) or on top of local telegraph poles. All of these nodes would be fed by a fibre optic cable, with the remaining G.fast link running over your existing copper line.

BT intends to trial G.fast with around 2,000 premises (availability only – actual uptake will be in the hundreds) in Huntingdon and then the same again in Gosforth (Newcastle), the latter of which should begin this month. On top of that a smaller “technical trial” will take place in Swansea (Wales). Hardware from ADTRAN, Alcatel-Lucent and Huawei is being tested.

BT are also testing a 1Gbps Fibre-to-the-Premises (FTTP) based upgrade to their business focused and long-stalled Fibre-on-Demand (FoD2) service, which is trying to find a more cost effective approach to deployment (note: final product pricing won’t see much change and will remain very high). Here’s a useful map to show where the current trial is in operation.

Huntingdon G.fast deployment trial map

The Trial Update (Huntingdon)

Firstly, the latest update appears to confirm the trial’s exact product specification, which includes some clarity on upload speeds and the “prioritised rate” (these were previously unclear).

NGA2 Technology Types

1. NGA2 G.Fast
* Up to 330 Mbit/s downstream
* Up to 30 to 50Mbit/s upstream
* Downstream prioritised rate 80Mbit/s
* Multicast enabled

2. NGA2 Fibre on Demand (FOD)
* Up to 1Gbit/s downstream
* Up to 100Mbits/s upstream
* Downstream prioritised rate 100Mbit/s
* Multicast enabled

Apparently the initial G.fast results have been “very promising” and “in line with our lab modelling“, with “almost” all of the customers that have been connected receiving 300Mbps+ speeds (most are said to be capable of receiving an even higher rate). “These results give us confidence that G.fast is an excellent technical solution to deliver ultrafast speeds to most homes,” said the report.

However they also confirm that most of the above customers are “fairly close” to the DP, which would of course result in the best speeds. The trials will also test over longer distances too, although there’s no feedback for those yet. It’s further noted that the FOD2 trial hasn’t yet begun and will follow “later in the year“.

One other point to make concerns G.fast’s upload rate, which the progress update highlights as “up to 30 to 50Mbps“. However the UKNOF speech itself stated that this information is out-of-date and in fact they’re currently only testing with 50Mbps.

The hope here, as BT has stated before, is to begin a commercial G.fast roll-out to “most homes” from 2016/17 (assuming Ofcom doesn’t throw a spanner in the works by splitting Openreach). This would be a 10-year long deployment, with top service speeds starting at 300Mbps and rising to 500Mbps towards the end of that period. We suspect this would at most reach 60-70% of UK premises. A further trial update is expected next month.

UPDATE 10:47am

Anybody with an interest in the technical side of the G.fast trial might also like to check out the related SIN 518 (Issue 1 – August 2015) document, which confirms the 50Mbps upload speed in writing.

Add to Diigo
Tags: , , , , ,
Mark Jackson
By Mark Jackson
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 Twitter, , Facebook and Linkedin.
Leave a Comment
45 Responses
  1. Avatar alusufferer

    so it seems i’ve been confusing FTTdp and G.fast.

    if this equipment is installed within the Cabinet it will do nothing for my town as it’s the cables from the cabinet to the homes that are the issue. i’d assumed the g.fast equipment would be run out to the dp under the street. From this, is just looks like it will really boost speeds for those close to the cab then tail off extremely quickly, and probably a lot quicker still on aluminium lines.

    have i got it wrong or is there any hope for us?

    • Avatar MikeW

      Yes. Or no.

      Ian confirmed that the DPU could be placed at the cabinet, or at the DP, or anywhere inbetween. The trials are to help confirm the “rate vs reach” results, as an input to the plans.

      Right now, the DPU is sized for 16 lines, which limits its use at the cabinet. Ones sized for 48 or 96 lines might come in the future.

    • Avatar MikeW

      Just so you can get an idea of what the trial is doing…

      See the large group of purple dots in the centre of the map Mark published?

      That estate seems to be covered by 3 cabinets. However, each purple dot appears to represent an individual FTTdp DPU, each of 16 lines.

      That looks like there are a lot more DPU’s than “just” alongside each cabinet.

    • Avatar alusufferer

      MikeW, thanks, that makes sense.

      Hopefully openreach target some problem areas as part of this next phase

    • Avatar Steve Jones

      I’m afraid OpenReach will target areas based on where it makes the most commercial sense. There is no incentive under the regulatory regime for any company to deal with problem spots.

  2. Avatar MikeW

    More updates:

    As I posted yesterday, it looks like they have 16 connections so far. All 300+/50, but all relatively close to the DPU. “Max attainable” downstreams being seen as 400-500.

    Longer distances will be deployed; “rate vs reach” looks promising.

    The G.fast upstream was chosen to be “up to 50Mbps” rather than “up to 30Mbps”.

    Nodes are currently 16 lines, and so far seem to be being placed underground – worries about access via ladder reducing the likelihood of a pole-top location, even though the report pictures on of those. Forward-fed power, with some above-ground power pillars. Reverse power not mature enough yet.

    3 vendors; 2 chipsets.

    Huntingdon seems to be more underground infrastructure, while Gosforth is overhead.

    These two sites are CP-focussed trials, while Swansea remains a BT-specific “technical trial” of “special cases”.

    Overlaid cabling, to prevent ANFP problems.

    Some of the FoD2 trials will focus on providing fibre connection from the G.fast node, rather than the AN. Target is to go back to a single appointment, not two separate ones.

    • @MikeW What do you make of the 2nd copper pair for the G.Fast DPU function? Is the 16 port unit the same for RN and G.FAST DPU modes?

      In G.Fast DPU mode how far away can the 16 port DSlAM be?

    • Avatar Ignition

      NGA – DPU is the term Openreach are using for the G.fast DSLAM. They are currently deploying them underground, co-locating them in underground DPs, hence Distribution Point Unit.

      When they have higher capacity kit than 16 lines they can then look at placing them elsewhere, but there’s no distance between DPU and DSLAM as DPU = G.fast DSLAM.

    • @ignition – The graph with the second pstn suggested DPU – might have been distrubuted power unit, or reverse powered varient.

      That does not look as if it is being tested. Is that the case? If so, where is the power source – the adjacent cab, carried from the HOP or local FTTC cab?

      The power issues are likely to dicate the utility of this DSLAM in a rural environment.

    • @Ignition forgot to say thank you.

    • Avatar Ignition

      No worries. There are some power pedestals feeding the DPUs, I believe both power from cabinets delivered over twisted pair and power from nearby AC feeds are being used.

    • Avatar MikeW


      Does @Ignition’s post answer your questions?

      Even knowing that the DPU is the term for the G.fast DSLAM, and that the power pillar is a separate, centralised point, there is still a valid question: what does it take to get power from the pillar out to the DPU? What limitations are there on distance?

      TBH, I’m not really sure of the answers here – and there was nothing extra added in this presentation.

      There are a lot of presentations for powering a DPU, but most focus is on the new feature of reverse power (from the CPE), and few concentrate on the older “remote power” fed from a central location (which I usually call “forward power” to distinguish from “reverse power”).

      Generally, the concept of “remote power” appears to require spare copper pairs, and cannot use pairs that are currently in use for a phone service. The number of pairs will depend on total power and distance, and (as ever) early trial DPU’s are unlikely to be optimised in their use of power.


      The last BT presentation on G.Fast gave details for the Heathersett technical trial, showing that the final power connection to the DPU (forward-fed from a central pillar) as being done with “10/20pr power cable”, no length details. Power to multiple DPU’s was by “50 pr power cable”, so the trial units will go through a lot of “spare” copper.

      Obviously BT are installing an overlay copper network for this, because these trials don’t meet ANFP limitations in the real network. There is obviously plenty of “spare” when you do that…

    • @MikeW @Ignition, the immediate question on power is answered so thank you. I do want wish to ask a follow on.
      Does this mean, or could it mean the Fibre DP is, (unless there are 16 cusomers clustered together), some 350m from the edge of the network? If so, through ECC charges it kills FTTP! Can Openreach at all costs resist such a request from BT Goup should such a request be made?

      G.Fast like the FTTC cabs are great as long as you recognise the limitations and not create conditions where the underlying potential of the medium of fibre is denied those who cannot benefit from a DSLAM cab or node on a pole.

    • Avatar Ignition

      Where they aren’t going to be next to existing cabinets the DPUs are either going to be, as the name implies, colocated with existing DPs or deployed in a similar manner to FTTRN where a copper feed is intercepted.

      As far as excess construction goes it’s not an issue. Openreach won’t be building on demand, they’ll build where they see fit just as they do with FTTC. I’m sure if individual operators want to pay for node install they can but that’ll be the extent of their involvement.

      If Openreach decide to build a node and the costs start to rack up they can either abandon the build or suck it up, they can’t charge anyone else directly for building out their network when others didn’t agree to pay for it.

    • Ignition – if G.Fast is presented as a panacea then yes, but it is not and so extending fibre access using FOD kicks in. In these circumstances and they will arise the customer, particularly the first customer or the local authority will be punished with ECC’s.

      It would arise if a dictat that G.FAST is the always the first answer, and where it fails, an expensive FOD would be overlayed from the DPU. I think the choice needs to be preserved where power, distance, population density and condition of copper means a 16 port DSLAM is unlikely to be of value. Maintaining choice translates into retaining the option to support FTT dp – where the customer pays the final drop cost, rather then ECC over much longer distances.

  3. Avatar adslmax Real

    FTTC 80/20 (now) -> G.FAST 300/50 (2025) for my area that I reckon.

    • Avatar JamesM

      Virgin Will be miles ahead by 2025.

      Thanks though ISPr for pointing out that Swansea is in Wales – anyone would does not know this is too stupid to understand this article anyway. There is only 1 Swansea in the UK. Unlike the 4 Newports there are.

  4. Avatar AndyH

    A few corrections…

    It’s Ian Boothman, not Ian Boothsman.

    There is a big difference between the number of homes passed and the number of trialists. There will not be 2,000 trialists in Huntingdon (the aim is for around 25-30% of the homes passed to be on the trial).

    A few other points –

    Openreach issued a consultation last week and are planning a wider pilot across the country to start in the first half of 2016. G.fast will not be offered as a replacement to FTTC, but as an additional product offering (at a premium over FTTC costing). It will be offered across the MPF, WLR and SO (single order) line variants.

    • Not seen a consultation like that, but would make sense.

      The “Boothsman” thing is how UKNOF wrote it on the slides page. Will correct.

    • Avatar AndyH

      It’s on the OR/BTw portals – am sure someone will send it over.

    • Avatar Ignition

      I may well need a more powerful load balancer next year, then 🙂

    • Avatar adslmax Real

      AndyH G.Fast premium cost? How much roughly it will be for G.Fast product 300/50?

    • Avatar AndyH

      No idea – this is why Openreach have issued the consulation. “CPs are invited to comment on how G.fast will be positioned in their portfolio; who their target market is and any insight on propensity to pay for higher bandwidth speeds.”

      Seeing how FTTP 330/30 is £50 a month with BT retail, I doubt it would be cheaper and there is bound to be an installation cost.

    • Avatar MikeW

      That consultation is interesting. It implies that NGA2 (of either variety) may end up being an on-demand product.

      Presumably the installation cost for someone opting for the G.fast variant will have to be significantly lower than that for FoD2, otherwise what would be the point?

      I would agree that the monthly cost would have to mirror the existing infinity 3 and infinity 4 price-points.

      But, but, but….

      Would there be any reason to *not* offer infinity 1 and infinity 2 pricepoints from a G.fast node as well? That is: offer the 40/10 and 80/20 product variants?

      That would allow those on longer FTTC lines to order a G.fast variant to shorten the copper (and increase capability) but still only use what is needed.

    • Avatar DTMark

      They will have a hard time pricing this as a premium product in cabled areas; by then cable will offer 600Mbps to 1Gbps as a standard offering with no additional installation fees.

    • Avatar MikeW

      “by then” …

      It will indeed be something of a problem if VM really can offer 600Mbps as their entry level “by then” (2016 in AndyH’s post)

      Do we expect that?

      By all reports, the next VM upgrade will take entry level to 100 Mbps, and will take them (if the last 2 upgrades guide us) another 2 years to roll out. 300 Mbps will be a premium product there for those 2 years.

      600 Mbps as entry level seems a little way off for now…

    • Avatar Ignition

      VM’s entry level product will be 100Mb by 2016, their top product 300Mb, however whether those have the same terms, conditions, and any traffic management as now is obviously a different matter.

      VM could, in theory, offer 600Mb as their top level product by late 2016. Whether they see any commercial case in doing so is a different matter as it would be expensive.

  5. Avatar William Blunn

    “although it requires significantly more radio spectrum”

    Radio spectrum? What are you talking about?

    G.fast works over wires, not radio.

    “Frequency spectrum” perhaps.

    • Avatar Ignition

      I imagine the man is referring to RF spectrum. Right or wrong xDSL and CATV networks refer to RF extensively as the RF spectrum is where their metallic sections do their thing. I appreciate some might say RF should be used as a synonym for ‘radio’. 🙂

    • Well yes, if we have to be precise, then I could describe it slightly differently, but I tend to assume that more people will understand what I mean by the use of “radio”.

      After all the radio spectrum is the part of the electromagnetic spectrum from 8.3kHz to 3000GHz. Maybe I should just say electromagnetic spectrum to cover everything :).

      Lest we not forget that related signals do sometimes leak out of their copper cables to become wireless, so they’re not entirely one thing or the other.

    • Avatar William Blunn

      Ah yes. The general dumbing down of everything. We have to use “pedagogical facilitations”, because people won’t understand if we say things properly.

      Please stop the world. I want to get off.

    • In this case it’s probably more pedantry than dumbing down 😉 .

    • Avatar Steve Jones

      “although it requires significantly more radio spectrum”

      I think we all knew what he meant…

      nb. DSL and g.fast signals all do propogate as EM waves down a twisted pair so I might make a pedantic case that they are radio by some definitions, but I’m above that sort of thing.

  6. Avatar William Blunn

    “G.fast works in a roughly similar way to the current 80Mbps capable hybrid-fibre FTTC technology that dominates the UK market”

    This is a bit dodgy too.

    G.fast acts to replace the VDSL technology used for the last few hundred metres of an FTTC service.

    Comparing G.fast with FTTC is like comparing apples with oranges.

    The appropriate comparison would be to compare G.fast with VDSL, or perhaps you could compare FTTC/G.fast with FTTC/VDSL.


    “G.fast works in a roughly similar way to the current 80Mbps capable VDSL technology that is used by the hybrid-fibre FTTC/VDSL services which dominate the UK market”

    • Disagree, I did say “roughly similar way” for a reason and then described it, not least to avoid having to over-explain as that quickly becomes laboured. The more terms you add, the more you have to explain, while our goal is to simplify with less jargon.. not more.

      Openreach also tends to market FTTC more as a product than a technology term, they don’t sell a product called “VDSL2” to ISPs but it’s what they’ve chosen to call FTTC the product. If they start to sell G.fast under the FTTC term then I’ll tweak the language, but right now they’re marketing to ISPs as “G.fast”.

      Lest we not forget that the “C” is for cabinet, whereas G.fast doesn’t have to use a street side cabinet.

    • Avatar Ignition

      As of right now FTTC is GEA-FTTC or SOGEA-FTTC, while G.fast is, alongside FoD2, listed as GEA-NGA2.

    • Avatar Steve Jones

      I find that a bizarre comment. G.fast is a hybrid optical/copper system using local nodes in exactly the same way that FTTC/VDSL2 does. The differences are that it will use many more nodes in order to shorten the length of copper loop to the customer and, of course, the detailed differences in modulation and similar issues on the copper side. It is perfectly reasonable to say this works roughly like FTTC.

    • Avatar MikeW

      I’m with the others. The key ingredient is that it remains a hybrid copper/fibre technology. Just that the copper is getting dang short.

      Incidentally, one of the other presentations at UKNOF was on upcoming upgrades to ethernet.

      Of interest to those with installed cat5e or cat6 copper cabling is the potential to upgrade from 1Gb to 2.5Gb or 5Gb, using the existing cables.

      There remains a great interest in pushing already-installed copper to ever-greater speeds, whatever form that copper appears in.

    • Avatar Steve Jones

      Yes, I’m inclined to agree. The copper vs fibre debate gets a bit tiresome (after all, how many people actually use any form of fibre connection in their home? Whilst there was a digital audio optical cable system, it was hardly popular and even 4K video is carried over copper HDMI cables (and, that of course, means proper 50 or 60fps uncompressed RGB; the data rates are huge).

      It doesn’t mean copper doesn’t have issues. It’s vulnerable to water ingression, signals attenuate quickly and so on, but it is convenient and, most importantly, already in situ. That last bit is important.

  7. Avatar adslmax Real

    Interesting in this part:

    Dynamic Line Management
    Dynamic Li
    ne Management (DLM)
    will not be used in the G.fast
    Trials. A
    background process will be used to optimise line

    I guess when G.fast is available to nationwide roll out DLM will be active on G.fast

  8. Avatar Archie Rouse

    Its lovely to see the net speed getting faster and faster… but how about actually doing useful things like sorting out rural areas… im still stuck on a terrible 1.5mbps unstable connection with no upgrade in sight! its 2015 not 2002 this should not be an issue anyore!

    • Avatar Ignition

      I imagine that unless a community project steps in the drive for that has to come from the government rather than BT, Virgin Media or another private company.

  9. Avatar MarkOfThePeak

    I need FoD2 to become commercially available ASAP, it’s our only hope when it comes to broadband.

  10. Avatar Dave

    I live in the trial area in Huntingdon and a lot of work has been done (John Henry Group) on new local infrastructure – cabling, street cabinets etc. Whilst they’ve been doing work on my street i’ve been told i can’t go on the trial (not in area!) – so all a bit random it seems.

    But to my point … given the amount of work they have done for the trial and the perilous state (age) of BT’s copper access network why on earth are they not just looking at putting FTTP everywhere? Surely a better technical and commercial option as the copper will have to go in the long term and they are throwing good money after bad?

Comments RSS Feed

Javascript must be enabled to post (most browsers do this automatically)

Privacy Notice: Please note that news comments are anonymous, which means that we do NOT require you to enter any real personal details to post a message. By clicking to submit a post you agree to storing your comment content, display name, IP, email and / or website details in our database, for as long as the post remains live.

Only the submitted name and comment will be displayed in public, while the rest will be kept private (we will never share this outside of ISPreview, regardless of whether the data is real or fake). This comment system uses submitted IP, email and website address data to spot abuse and spammers. All data is transferred via an encrypted (https secure) session.

NOTE 1: Sometimes your comment might not appear immediately due to site cache (this is cleared every few hours) or it may be caught by automated moderation / anti-spam.

NOTE 2: Comments that break our rules, spam, troll or post via known fake IP/proxy servers may be blocked or removed.
Cheapest Superfast ISPs
  • Hyperoptic £21.00 (*25.00)
    Avg. Speed 50Mbps, Unlimited
    Gift: £50 Shopping Voucher
  • TalkTalk £21.95 (*36.00)
    Avg. Speed 38Mbps, Unlimited
    Gift: None
  • Plusnet £21.99 (*34.98)
    Avg. Speed 36Mbps, Unlimited
    Gift: £75 Cashback
  • SSE £22.00
    Avg. Speed 35Mbps, Unlimited (FUP)
    Gift: None
  • xln telecom £22.74 (*47.94)
    Avg. Speed 66Mbps, Unlimited (FUP)
    Gift: None
Prices inc. Line Rental | View All
The Top 20 Category Tags
  1. BT (2539)
  2. FTTP (2261)
  3. FTTC (1681)
  4. Building Digital UK (1618)
  5. Politics (1448)
  6. Openreach (1433)
  7. Business (1261)
  8. Statistics (1115)
  9. FTTH (1108)
  10. Mobile Broadband (1059)
  11. Fibre Optic (978)
  12. Ofcom Regulation (925)
  13. Wireless Internet (921)
  14. 4G (919)
  15. Virgin Media (871)
  16. EE (606)
  17. Sky Broadband (602)
  18. TalkTalk (587)
  19. Vodafone (535)
  20. 3G (418)
Helpful ISP Guides and Tips

Copyright © 1999 to Present - ISPreview.co.uk - All Rights Reserved - Terms , Privacy and Cookie Policy , Links , Website Rules , Contact