Following the broadband money

BT makes the case for fibre to the home

with 45 comments

BT has inadvertently made the case for a national fibre to the premises roll-out.

Today it announced it was delivering 10Gbps to a Cornish engineering company over that company’s 330Mbps fibre link between its office and the BT exchange in Truro.

BT claims this is the fastest “retail” broadband link in the world, but it’s not connected directly to the public internet. Nothing on the internet presently works at that speed, it claims. Indeed, the computer and networking kit on either end of the link run slower than the fibre link.

The BT press release quotes Ranulf Scarbrough, programme director for the Cornwall superfast broadband programme, saying,“What is exciting about this trial is that these hyper-fast speeds have been obtained over the exactly the same fibre that carries BT’s fibre broadband services today. All we are doing is changing the electronics at either end.”

Scarbrough goes on to say, “This trial shows we are thinking and ready for the future even though there are no current plans to deploy this technology. A lot of this project is about future proofing – making sure that it’s not just the fastest speeds today but that we can continue to be at the cutting edge for five, 10, 20 years.”

Case closed.


Written by Br0kenTeleph0n3

2012/11/22 at 23:19

45 Responses

Subscribe to comments with RSS.

  1. Actually a few weeks old this bit of news, just seems the PR has gone online now.

    The usual raft of mainstream tech mags also covered it around that time.

    A criticism of FTTP via GPON has been the speed limits, so this trial was a way of showing that speeds well beyond Gigabit are possible using XGPON.

    Where does BT claim this is the worlds fastest RETAIL link, cannot see anything about retail in the press release. 10GigE optics are not cheap, and the price has not dropped fast enough to ensure widespread adoption in data centres – don’t have link but one mag was talking about that this week.

    The neat bit was that this was done over the same link as the existing 330 Mbps fibre service.

    Andrew Ferguson

    2012/11/22 at 23:41

    • BT doesn’t say “retail” explicilty, but what else would you call such a service if it were available? As to running it on the 330Mbps fibre, was this concurrent? If so, that sounds like some fancy opto-electronics. As to 10Gbps, this is not a data centre environment, so different economics apply. Besides, you will have seen JANET’s new baseline speed will be 100Gbps, with a planned upgrade to 400Mbps. It appears 10Gbps is table stakes for carriers for now.


      2012/11/23 at 01:55

      • So you are adding the word retail? If so why? If the service were available I would call it a wholesale service, as until a retail provider offers it, it is just that.

        Fancy opto-electronics, or just that small thing called WDM and diffrent frequencies of light.

        Different economics? So optics that run at 10GigE are magically cheap if they are outside a data centre? 10GigE is common enough and Sky fibre backbone runs at 700Gbps (if pre coffee brain is working). I said nothing about the technical difficulty, just that compared to Gig optics the price difference is still very large.

        Andrew Ferguson

        2012/11/23 at 09:59

      • You are being premature. BT says its testing the tech. No current plans to launch it as a product. Do you expect Openreach to offer this as dark fibre?


        2012/11/23 at 10:20

      • You started with the word RETAIL, so it is you who is being premature.

        Stop trying to put words into other peoples mouths.

        Andrew Ferguson

        2012/11/23 at 14:46

  2. It would be interesting to see how the fibre was integrated with the other FTTC cables.

    Merrow Drover

    2012/11/23 at 01:03

    • Judging from the release, it looks like the fibre is the constant, and just the electronics at either end changed. If BT is reading this, perhaps one of its engineers would be kind enough to explain, and give us pictures 😉


      2012/11/23 at 01:59

      • That is exactly what happened – same fibre and different electronics

        Andrew Ferguson

        2012/11/23 at 09:59

      • Apologies for the slow response but have only just come across this – happy to answer any questions on the XGPON trial.

        It doesn’t use WDM, it uses XGPON aka NGPON1 or 10GPON. This is the next evolution up from GPON – which is essentially TDM on a single wavelength. Its the first instance of commercially available kit on XGPON1. XGPON2 is on the way (this is WDM) – this will do 40G; NGPON3 is probably 100G and a couple of years after that (likely to be OFDM). The standards body (FSAN) is co-chaired by a BT colleague as it happens.

        There certainly are faster optical standards out there than 10G but the key thing about this technology is it is designed to deliver broadband economics rather than core or long haul economics. So Janet may be adopting 100G but I doubt thats at a consumer broadband price point !

        For the trial we have introduced an XGPON headend alongside the existing GPON headend in Truro exchange (which is already delivering FTTP service). It delivers into the same 32-way split PONs we have been deploying in Cornwall for FTTP but runs on a different wave to GPON. So it shows that upgrading the electronics on either end of the PON is straightforward as a future upgrade for faster speed.

        The technology delivers a 10Gbps aggregate across users on the PON, so is capable of delivering future 10G end user retail services. The 10G would be shared but I suspect as you get to these faster speeds that usage would be quite spikey as large data objects are moved across the net, compared to todays streaming traffic which tends to have a more consistent bandwidth.

        The trial uses the 10G from the end users to Truro. From the headend to the internet is achieved on a 330Mbps circuit. We may do a 10G backhaul but its really the access network that is the challenge and thats what we wanted to evaluate in the trial. We have some kit (content servers etc) running 5-10G traffic from the exchange over the 10G links.

        So to be clear, it is a proof of concept trial at this stage. The technology works very well but you would need a lot of users on these sorts of speeds to achieve broadband price points (ie. in order to share out the headend costs). That demand is not there today but thats not to say it won’t be in 5 or 10 years time.

        It terms of the request for pictures, even better we have a little film of the trial on youtube at

        Happy to answer any other questions …

        best regards – Ranulf

        Ranulf Scarbrough

        2013/01/06 at 22:08

      • Many thanks Ranulf, and season’s greetings to you. Glad you found us. Here’s hoping the tests go well – from what you say, it certainly looks that way.
        A couple of points, if I may: you say BT’s main interest was in the access network. Have the tests indicated whether and how much remedial work might be needed to upgrade the copper links to ensure a consistent 10G service to end users, and if so, what can you tell us about that and the relative cost of replacing the copper with fibre?
        Secondly, you say you would need “a lot of users” to achieve certain broadband price points. Is it currently possible to provide a demand curve for users or speeds against costs for XGPON?
        Finally, I’m sure you are aware of the Bell Labs/Alcatel Lucent report, “Video shakes up the IP edge” (you may need to register) on the impending need to switch the total network from centralised to a distributed delivery system mainly to cope with the explosion of video traffic and users’ desire for personalised or enhanced video on demand experiences. How does that fit with the thinking in the XGPON trials, or do you see a different scenario or time-scale emerging?


        2013/01/07 at 21:57

  3. This has not ‘inadvertently’ made the case for a national FTTC rollout. It’s a technology that enables a 10G link, nothing more.

    And it would be a ‘wholesale’ link for all ISPs to use?


    2012/11/23 at 09:05

    • It appears to be a point to point link. Perhaps Openreach could be persuaded to offer it as a wholesale product, but I wouldn’t hold my breath. That’s why the UK needs to insist that Openreach offers a dark fibre product.


      2012/11/23 at 10:10

      • Why wouldn’t they offer it as a product? What wholesale products do BT have?


        2012/11/23 at 10:30

  4. I think its interesting that any sort of speed in the future can be delivered over fibre just by altering the lights at the ends. This proves the case for using it in areas that can’t be served by phone lines, rather than wasting millions trying to make the phone lines deliver what currently passes for ‘superfast’. If we start at the edges and work inwards then everyone can have a fit for purpose connection now, and into the future. If we put the money into making those near cabs and exchanges go faster then the pot will be empty when the time comes for everyone to be on fibre, and this country will be borked.


    2012/11/23 at 09:10

    • @ Chris,

      It might not be unreasonable to say we are already borked and that there seems little sign of a significant and desperately need change. You could start by looking at the response to the expensive HoL report. Heads in the sand comes to mind!

      Surrey Hills

      2012/11/23 at 09:43

      • To a big extent once .gov projects are underway changing direction is very difficult.

        If we did – then the planning/state aid/procurrement stages mean that first FTTH via new project would probably be delivered in 2016, by which time the 10% of UK on under 2 Mbps would have moved home to nearest big town.

        Andrew Ferguson

        2012/11/23 at 10:07

    • Why the edges? Why not the middle?

      Remember the edges will take longer to roll the fibre to as distances are greater. Did b4rn network link all the houses at the far end before building the core network?

      If we do those at the edges the money pot will also be empty too. Just because you switch to a full fibre roll-out nationally it does not make the money pot endless. Lets scale the size of B4rn – £1000000 for around 1,000 properties, 26,000,000 properties to cover so £26,000,000,000 not unlike the estimates bandied around is it.

      Perhaps the answer is to install fibre to perhaps 80,000 points around the country, and at this place allow people to blow their own fibre to it, or pay the commercial operator to do it for them. Though in areas with pavements/roads I can see self dig being difficult, up on the fells/moors of course it is easy as just go across a field – if you can get permission.

      Andrew Ferguson

      2012/11/23 at 10:05

      • I don’t think the B4RN costs scale to the rest of the country. It is digging fibre to homes in some very remote parts of the country, which is not the same as digging it to the cities, where the duct infrastructure is already in place. The cost of the civil engineering work in towns should be a fraction of what it costs B4RN – just rip out the copper, replace it with fibre, and sell the copper to defray the cost.


        2012/11/23 at 10:16

      • And deal with copper cables buried in the ground, insufficient space to add fibre, blocked ducts, can’t turn off service for a few days etc…

        Yes, easy. Like to design and project manage?


        2012/11/23 at 10:21

    • ‘Trying to make’. Please explain for us all.

      Fibre speed capability does not prove any case.


      2012/11/23 at 10:24

  5. The elephant in the room is the lack of willingness for any of the current telco operators to open the flood gates on “unlimited bandwidth”. Bandwidth limited only by what you’re prepared to pay YOURSELF for the optics on either end of the shiny bit of glass. An entire economy of overcharging and funny fees for transit goes out the door when SMBs and customers with access to dark fibre get hold of it.

    As long as this denial remains, artificially restricting the dark fibre market, the further and further behind our economy will fall in the online market place. The telcos are doing this to us, supported by our government’s obsession with the incumbent. There are a few bright sparks who realise the next market to get into is dark fibre, but they are very much the exception to a cartel-like norm.


    2012/11/23 at 12:02

    • Well said Rob, perhaps C&W will revolutionise the market then….?

      Guy Jarvis

      2012/12/03 at 13:09

      • C&W have had 25 years to do it…


        2012/12/03 at 13:12

  6. For rural, establishing c10-14 Handover points per County (out of 80-100 exchanges) and pushing that fibre as deep as the available money will allow along 500-700 paths to Cabinets and Exchange Line Only exchanges per county is not rocket science. The underlying multi-Gigabit capability is in place for the market to divy up that capability on an ability to pay. The focus for rural must be managing the c£1.4-5bn to achieve VFM where >90% should be achieveable in most places but not I assume the Highlands where 4G/FWA ought to be used for 10Mbps edge of network.

    The challenge or question now must be whether the current Ofcom WLA review should be used to push on to create or allow a fibre switchover activity in urban areas, particularly for mulit-tennanted premises, and those LAs needed to get ride of satellite dishes from mulit-tenanted housing blocks. In its current form Ofcom is not minded to do so which I think is an error. They should not be shutting down efforts (chapter6 of WLA review) to develop SLU, PIA, dark fibre until at least all options for competition in urban areas are exhausted.

    NGA for All

    2012/11/26 at 10:59

    • Need to define exactly what a ‘handover point’ is and who pays who for installation and ongoing costs.

      The likes of Sky have one dish solutions for blocks of flats/apartments. Also needs a product to view and record all the current live channels delivered over broadband.


      2012/11/26 at 21:37

      • The handover point aka DVP aka local POP is simply the nearest geographic point from which high capacity connectivity, with a strong preference for dark fibre, can be interconnected with

        It is the removal/minimisation of excess construction charges and other distance-related charges where the value comes from investing in expanding that middle mile reach

        Guy Jarvis

        2012/12/03 at 13:19

      • And those points can be wherever anyone who is paying for them wants them to be. BT exchanges seem an obvious starting point. And where does the connectivity go back to?


        2012/12/03 at 16:24

  7. You can already buy a 10Gbit optical path from open reach either as a point to point or a wavestream service.

    A good optic will do 10x10Gbit without any problem – just fire off different wavelengths (DWDM) and you can do what you like.

    Dark fiber is something that open reach won’t be in a hurry to offer ever (if at all) if you had a dark path between 2 points that cost you a fixed ammount you would want to run the highest speed/number of wavelengths you could over it. There would be no easy commercial model.

    I am new to following this blog but it makes for interesting reading.

    I should declare that I’ve an interest in the sector as part of my day job, have in the past worked at BT on optical networks


    2012/11/28 at 16:19

    • Welcome to Br0kenTeleph0n3. Let’s hope your expertise can shine a light in some of the dark corners, let alone dark fibres 😉


      2012/11/28 at 18:08

  8. Sorry couldnt work out how to post this further up so re-pasting here:

    “Many thanks Ranulf, and season’s greetings to you. Glad you found us. Here’s hoping the tests go well – from what you say, it certainly looks that way.

    >> Yes – all working very well thanks.

    A couple of points, if I may: you say BT’s main interest was in the access network. Have the tests indicated whether and how much remedial work might be needed to upgrade the copper links to ensure a consistent 10G service to end users, and if so, what can you tell us about that and the relative cost of replacing the copper with fibre?

    >> That work has already been done when we built FTTP/GPON and put the fibres in 6-12 months ago for these customers as part of the Cornwall fibre build. Pretty much the whole business park could get superfast broaband on FTTP from Openreach already before we started the trial. The trail was an upgrade from FTTP-GPON to FTTP-XGPON on the same fibres we had already deployed. Replacing copper with fibre is expensive as you have to replace everthing – its a lot of work and its disruptive too as you have to close roads and the like and – believe it or not – not everyone wants it.

    Secondly, you say you would need “a lot of users” to achieve certain broadband price points. Is it currently possible to provide a demand curve for users or speeds against costs for XGPON?

    >> I dont have a curve and it wouldnt be much use probably. Equipment tends to start expensive when its new and get cheaper – especially when it goes mass market and gives you those nice economies of scale from volume manufacture – so the price in 2013 is probably not that relevant. Similarly, headends are quite big ticket items. If you share the costs between 5 connected customers you need to charge a lot more than if you share it with 10,000. So you get the idea – these technolgies work well when there is a global ground swell of high volumes making the equipment cheaper. Thats not happening today for 10G but it might in the future – if it does, we know what to do and that it works.

    Finally, I’m sure you are aware of the Bell Labs/Alcatel Lucent report, “Video shakes up the IP edge” (you may need to register) on the impending need to switch the total network from centralised to a distributed delivery system mainly to cope with the explosion of video traffic and users’ desire for personalised or enhanced video on demand experiences. How does that fit with the thinking in the XGPON trials, or do you see a different scenario or time-scale emerging?”

    >> Not really my area of expertise but I would say not that relevant. How much data can one pair of eyeballs consume ? Its not that much – say 20Mbps for 4k video ? A big household might have 5 simultaneous viewers. Thats 100Mbps. We can do that today with GPON and VDSL2 is not far off, nor are CDNs. XGPON is more about saying the access network wont be the limit – that just means that something else might be of course. I think these speeds will be about different sorts of applications ie. beyond video. So machine-to-machine may be big. Moving big data sets quickly eg. send all your home movies to the cloud in a couple of minutes to be geo-indexed and face-recognised. I didnt predict all the applications we have today so I doubt I’ll predict the future ones either – I am sure some clever people will though.

    Ranulf Scarbrough

    2013/01/07 at 22:47

    • Hi Ranulf – Many thanks for clearing up those points. Yes, of course FTTP was available earlier from Openreach. I was just wondering whether it might be cheaper to replace the copper with fibre to support XGPON-type speeds. I’m sure the short term case says it’s not, but what about over 10-15 years?

      Given that XGPON is under test, and without committing BT to anything, which do you think would be the right part of BT to bring it to market – Openreach, Wholesale or Retail?


      2013/01/08 at 06:37

      • I’m not sure I follow but if you mean what I think you mean its all about unknowns really. How many people would take what speed of service, when would they take it, what would they pay, what would the equipment cost, what would the deployment cost, what would it cost to operate, how long might people take it for until a faster technology came a long that they might move onto. If you know the answer to those 7-8 questions its a pretty straightforward to determine a yes or no but unfortunately you never do. We now know those answers for ADSL but we had to guess when we made a decision to start to deploying 10+ years ago.

        The XGPON trial is being operated by Openreach.

        Ranulf Scarbrough

        2013/01/08 at 11:12

      • Hi Ranulf – I agree that knowing all the details you mention would make life simpler for us all. I think the bravest person in the world was the one who bought the first telephone. (Who was he going to call?) One thing we do know (by gut feel if nothing else) is that when it comes to digital technology, people’s desire for it is virtually insatiable, that product lifecycles have got shorter with each generation, that convergence is having a cumulative effect on bandwidth consumption, and that Moore’s Law still holds good for the foreseeable future as far as technology costs go. I find it hard to believe that someone somewhere has not earned a PhD for developing the equations that describe these processes. If not, wouldn’t it make sense for BT to sponsor their development? It would save an awful lot of guesswork in the marketing department and allow BT to schedule its investments better.


        2013/01/08 at 14:36

      • Why might it be cheaper to replace copper with fibre? Cheaper than what? The key problem is still the cost of the physical work involved and it’s not replace, it’s add and then remove.

        Is the Deddington fibre only trial continuing?


        2013/01/08 at 15:04

      • Happy new year to you too, Peter. I wasn’t thinking of the “one for one” replacement of copper pairs by fibre pairs so much as the total cost of ownership of a network built to meet the future user demands forecast by network equipment suppliers, notably Cisco, Ericsson and Alcatel-Lucent. Of course, they have their agendas, but I’ve not seen or heard a rebuttal of their scenarios. In fact there was a report today that said 75% of the traffic on Germany’s new LTE services are video, which would support the suppliers’ view. As to the cost of replacing local leads, the current scrap copper price attracts thieves. This suggests BT could either offset the cost of replacing copper with fibre by selling the copper. I recall one somewhat tongue in cheek analysis that suggested the value of BT’s copper network was 30bn more than the company’s market value at the time. Perhaps more sensibly, Ofcom in 2005/6 calculated a “regulatory asset value” (RAV) of 58.51 per line for WLR, and 60.11 for LLU lines. Some might argue that Ofcom at the time introduced a 40-year “holiday” in that it set a 2037/8 deadline for all assets valued under the RAV scheme to be written off and/or converted to a different valuation scheme. This gives it plenty of time and lots of incentive to sweat those copper assets. Openreach’s present prices for new digital WLR and full MPF LLU lines respectively are 120 and 87.41. These figures will be inflated because they include a profit margin; however, if Ofcom has done its work, they should be pretty close to net asset value. Multiplying them by Analysys Mason’s estimateof the number of lines in each category by 2013/4 (11.47m and 6.66m respectively) and adding them values BT’s access network in round figures at about 2bn. (Curiously, the Valuation Office Agency reckons BT’s entire rateable value is around 294m.) But that all begs the question. Would a greenfield fibre network operator duplicate the existing copper network design? Highly unlikely. If you believe Alcatel-Lucent, it would build a decentralised network that puts most intelligence at the edge, closer to the end user, and that would make it cheaper by third to own and operate than most incumbents’ present centralised networks.


        2013/01/08 at 23:05

      • Of course FTTC potentially allows significant amounts of copper to be recovered if voice is also provided from the cabinet. What’s the weight and value of a eg. 2km copper pair?

        Please explain the difference between centralised and decentralised networks and how these are or would be deployed in the UK. In what way is the BT network ‘centralised’ or it realising these savings already on its FTTP deployment?


        2013/01/09 at 23:20

      • Please read documents linked above, particularly the Bell Labs/Alcatel-Lucent white paper, for answers to your questions.


        2013/01/10 at 06:32

      • The document refers to CDNs and earlier you implied that BT should look into them, only to be told that they were very much involved in deploying them.

        The cost savings do not refer to the access network in the ground and I was asking you about where the intelligence in the network of the UK incumbent is situated. Note that the document says ‘Alcatel-Lucent is ideally positioned to help service providers through this necessary transition.’ Like many other suppliers?


        2013/01/10 at 15:57

      • BT indeed claims it is “deploying CDNs” and even that it has developed its own CDN technology. From the comment, it seems BT is trying every possible version of “deploying CDNs”. One wonders how long it will be before it, like AT&T and Orange, abandons its in-house system for Akamai, or perhaps another. Indeed, where is the intelligence in BT?


        2013/01/10 at 17:03

  9. Sorry for coming to this discussion a bit late…

    Like GPON, XGPON is a shared technology. It’s not point-to-point. The standard allows for up to 128 drops on a single segment, so you could be talking about a time-slot of 1/128th of 10Gbps per user – i.e. 78Mbps – pretty much the same as you’d get with VDSL now, if you’re close to the cabinet.

    Running 10G-PON on a different wavelength from 1G-PON means the two can coincide, and whilst it’s nice to see it working “in the field”, it isn’t that much advanced than any wavelength multiplexing technology – we’ve had lensed-CWDM products for nearly twenty years now.

    In terms of wholesale. the reason BT is obsessed with GPON is because it doesn’t lend itself to unbundling individual subscriber local loops (i.e. there aren’t any, they are all on a single shared loop)

    Instead, their wholesale proposition is at a higher protocol layer – giving BT a stranglehold on the access technology and sets a floor on the wholesale price i.e. guaranteed profit for BT.

    BT are great at blowing their own trumpet but I don’t think I’ll be hanging out the BT bunting just because they’ve tested some hardware vendor’s gear when what they are really doing is creating a diversion so that behind the scenes they can further their monopoly on our vital infrastructure.

    The sooner Openreach is split from BT, broken up or taken back into public ownership the better.

    Robert Robinson

    2013/03/18 at 00:42

    • Do you really want to see a telecomms company in public ownership?


      2013/03/18 at 10:53

      • Do you really want health care to be run by politicians?


        2013/03/18 at 11:19

      • There are good arguments for municipally owned (or controlled) passive infrastructure, yes.

        Stokab (Stockholm), Amsterdam CityNet, Singapore OpenNet, Chattanooga …


        Robert Robinson

        2013/03/18 at 11:24

      • Can you ever see that happening in the UK? Would taking Openreach into public ownership also involve taking over the local ends of VM and other telcos as well.

        What would the cost be of nationalising Openreach and what’s the advantage compared with regulation?


        2013/03/18 at 12:07

  10. The problem is regulation isn’t working. Whether you think Ofcom is BT’s puppet or just a chocolate teapot, we don’t have a proper competitive infrastructure and the country is falling behind in comparison to other countries in the availability of the latest broadband technology.

    We need a way for small, nimble, new-entrants to offer services to local areas over a common “dark” infrastructure (fibres, ducts, poles) on a fair and equal footing to the big player(s).

    To be fair I did say “split, broken-up OR renationalised” – there isn’t a simple solution here, but it should be clear that doing nothing (and letting BT run the country) isn’t working in the public interest.

    None of this is helped by the blame culture in government – what is perceived as “safe choices” are the only ones considered, hence the adulation of big companies. “No one got fired for choosing BT”, right? Look also at the rise of these too-big-to-fail mega service providers (G4S, Capita, Atos etc.) which is the club BT want to be part of.


    Robert Robinson

    2013/03/18 at 12:29

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: