Skip Navigation

Can the UK afford £2.6 billion a year on FTTH roll-out?
Monday 22 February 2016 10:29:49 by Andrew Ferguson

The calls for a nationwide FTTH grow each week and everyone has their own idea on how to achieve it, but there is one part of the UK going for a full FTTH roll-out by 2020 and that is KC in Kingston Upon Hull with its Lightstream service and this makes for an interesting test case to look at what is possible and some of the costs involved.

The ballpark figure for rolling out FTTH in the UK is pretty old but still gets repeated with various mentions in the £20 billion to £30 billion region, but we thought it might be interesting to see what KC has achieved and how this might map to Openreach.

KC (KCom) has we believe 1,700 employees, which is a lot smaller than Openreach which is around 30,000 employees with 21,000 engineering staff but they do serve a smaller footprint. KC has managed to deliver around 45,000 premises passed by FTTH in a four year period, and if the employee to engineering staff ratio is similar to Openreach this will be around 1,100 engineering staff.

Using KC as the model is important, as unlike roll-outs by CityFibre B4RN, Gigaclear and Hyperoptic the incumbent in Hull also has to maintain its existing copper network so may be a better model to work on.

So with an engineering team 19 times bigger at Openreach, in four years working at the same rate they would passed 855,000 premises with FTTH, or if they had started in 2009 we would have 1.5 million FTTH premises passed. Of course to scale this up to a roll-out that matches the VDSL2 footprint of 23 to 24 million premises, it is not a simple multiplier as the number busy dealing with existing copper issues will remain static, so lets assume around half the Openreach staff are involved in the FTTH roll-out and the rest are doing the usual faults and installs. Scaling this up Openreach would need an extra 130,000 staff with an annual wage bill of £2.6 billion to have kept pace (Openreach engineer starting salary is in the £19,000 to £21,000 region, and we have ignored the extra costs of training, fleet vehicles etc for this simple projection).

If Ofcom was to mandate a FTTH roll-out now, of course some of the work for a GPON based FTTH roll-out has already been done as fibres and aggregation nodes supporting GPON FTTH are in every area where VDSL2 is available, but the work of pushing the fibre from these nodes to peoples homes still needs to be done. Lets presume that the nodes represent one third of the work needed for a FTTP deployment, to upgrade those premises to FTTH in the four years before 2020 would still be a mammoth undertaking that looks impossible without a much larger workforce.

With us all as consumers the contribution we make is around £8/month towards our VDSL2 connections, and given the demand for the faster services it is clear that not many would accept price rises that would make a significant dent in the salary bill and there is the problem - namely whoever builds the FTTH network will only see a reasonable return if it is a vertically integrated operation and even then the ROI period will be very long.


Posted by tgon about 1 year ago
KC must be reasonable sure its long term ROI will enjoy its monopoly for many years to come.
Posted by CarlThomas about 1 year ago
KC has exactly no-one who has unbundled its exchanges and I'm not aware of anyone who purchases wholesale access either.

I imagine they are quite confident this will continue.
Posted by cyberdoyle about 1 year ago
The FTTC cabinets will not support FTTH - some only have a couple of feeder fibres that share it out to the old phone lines, they are not active nodes with the equipment ftth needs, they are choke points and stop gaps. New nodes would have to be built to do true FTTH. Even openreach with its massive staff couldn't do it before 2020, because they have wasted too much time (7 years since the digitalbritain report) deploying obsolete technologies. What we need is to leave the FTTC folk alone on their superfast, and get some competition in the areas they don't serve.
Posted by andrew (Favicon staff member) about 1 year ago
@cyberdoyle Correct the cabinets will not support FTTH, that is why they are not used for FTTH, but rather the fibre splitter/aggregation node architecture that has been deployed ready for GPON. Or when you say true FTTH do you mean point to point?

Even if you do the last 5 million as FTTH that would need a workforce four times Openreach have to do the deployment in a 7 year period.
Posted by CarlThomas about 1 year ago
The FTTC cabinets are active nodes. If they were passive they'd have no active electronics which they obviously do.

They can support FTTP with line card swap outs and have more than enough backplane capacity.

Some of the cabinets have WDM multiplexers in them.
Posted by TheEulerID about 1 year ago

Please educate yourself. FTTH on the BT network is fed from the fibre aggregation points, the self-same aggregation points that were put in to support FTTC cabinets. There are many places where there are both FTTC and FTTH fed from the same points. Yes, more GPON nodes would have to be put further out as the fibre network was extended, but it would be on the back of existing investment.
Posted by TheEulerID about 1 year ago

The FTTC cabinets are irrelevant to FTTP extensions. Any FTTP extension would be from the fibre aggregation points, the vast majority of which would have been put in to support FTTC.

What FTTC does (which many people ignore) is enable a lot of properties much faster, with much less labour than FTTP as the really labour intensive bit is all those "final mile" links to each property.
Posted by TheEulerID about 1 year ago
I should add a nice simple analysis by Andrew which is illuminating. It's also work comparing this with the Jersey fibre programme which uses a wholly-owned company called Gigabit Field Force for it's work. That uses a lot of imported labour on zero-hour contracts and has been subject to a lot of disputes. Despite having a well-defined area, all under monopoly control the Jersey Gigabit programme has hit issues on rollout, finance and is not cheap to the retail customers.
Posted by andrew (Favicon staff member) about 1 year ago
And I've ignored the usual pension contributions and other costs of having an employee, which based on old figures probably double the cost.

I'd happily back a roll-out with a 15 year time frame, as will keep me busy until I retire.
Posted by CarlThomas about 1 year ago
I'm aware of the existing architecture, I was merely making the point that the existing cabinets could support FTTP rollout if so desired.

BT may choose to do this in future or they may not.
Posted by chilting about 1 year ago
One obvious question-
Are there enough trained telephone engineers in the UK to undertake such a task and what do you do with the engineers when they have finished?
Posted by andrew (Favicon staff member) about 1 year ago
@chilting Training of new staff would be an additional cost and agencies would start naming a price, law of supply and demand if they had to take up slack.

Hiring agency staff from Europe is an option, but many countries with high FTTH penetration are very small compared to UK.

Posted by TheEulerID about 1 year ago

Nope, and I was chatting to an OR jointer I overheard in a pub a couple of weeks ago. He says (in a complex network) it takes years to become fully proficient and, there's a real shortage with many due to retire soon. His current trainee is a young chap from Portugal who is keen and bright, but will need lots of supervision for a long time. Building a relatively simple new network is one thing. Enhancing one with vast numbers of circuits built to different standards over many decades without a disaster is something else again.
Posted by Somerset about 1 year ago
Much of the installation work is done by contractors. You need to calculate that element.

Chris - 48 fibres into this village. Get some facts.
Posted by WWWombat about 1 year ago
Can we get an idea of the scale of your village? How many premises are there? Or what population?

And how many FTTC cabinets are supported so far? And how many left to add?
Posted by TheEulerID about 1 year ago
It would be certainly interesting to know how many premises in the village.

In the case of FTTP, if those 48 fibres were used with GPON splitters at (say) and average of 16 premises per fibre then that's enough for around 768 properties or, let's say, a population of about 2,200. More with a larger "fan-out". 10G-PON has been trialed, and a number of options for increasing that are under development.
Posted by Somerset about 1 year ago
2 cabinets, 750 properties. 48 fibre cable blown 1.5km to first cab. 5 blockages on the route.
Posted by ivhr about 1 year ago
I know this is hardly a rigorous study but I don't think the numbers stack up

KC or their contractors may not be working flat out and they have reduced economy of scale

I live in an FTTP hotbed - AFAIK they use contractors and can roll it out quickly enough

BT is planning to do G.FAST, which will need the same work to get to the DP as for FTTP anyway

KC by definition is dealing with a urban area, whereas Openreach has more rurality - quicker / less red tape

no mention of the inherent savings once the FTTP is in and copper can be abandoned

If only we started years ago..

Posted by fusen about 1 year ago
Is FTTP really realistic for the UK? The amount of work involved in getting the fibre into everyone's buildings just seems insane. It'd be an absolutely huge initial cost that companies won't want to get involved in.
Posted by WWWombat about 1 year ago
@Euler, @somerset

For a mature brownfield village, I'd dimension
- 10 fibres as breakage spares
- 6 fibres for FTTC
- 8 fibres for G.Fast or business
- 24 fibres for GPON

I would have thought 1 aggregation node would supply the village (from near the first cab?). 7 splitter nodes and 40-50 fibre DP's to be added as needed.
Posted by andrew (Favicon staff member) about 1 year ago
@ivhr never claimed it as a rigorous study, economy of scale helps with cost of hardware, less so with actual civil work and contractors are not free. There may be less cost as you don't have to train people, pensions etc but then I never factored that in, just basic salary. depends on if cabinet based, DP or other location based, but yes more work needs to be done.

KC is a mixture of urban with some fringe rural, as for less red tape depends on if using code powers or going across fields.

Be interested in others envelope calculations of just labour cost.
Posted by ivhr about 1 year ago
@ fusen

Depends on the building. Blocks of flats might be a challenge (but G.FAST will work wonderfully there), and people with directly buried cable are going to have issues, but a lot of people aren't in those scenarios.

In my case, except for new builds or houses built in the recent past (which already have BT FTTP and underground cabling), everything around me is overhead wiring from poles - some of which is already fibred for FTTP. Apart from the challenge of getting fibre to the pole (which is needed to do G.FAST), it shouldn't be too difficult to get it into the home.
Posted by andrew (Favicon staff member) about 1 year ago
The most interesting was that if Openreach had started in 2009 with existing work force, then a foot print of 1.5 million homes passed was possible.

Of course homes passed is a useless stat, so you need to add 3 to 8 hours labour to do the final drop into each property from the final manifold.

Copper recovery would only be possible if a mandatory switch from copper to fibre was in place as areas where rolled out.
Posted by andrew (Favicon staff member) about 1 year ago is not going to make Sky and TalkTalk happy, hence exploring FTTH estimates, rather than a model.
Posted by WWWombat about 1 year ago
I'm also trying to figure out alternate numbers, to sanity check.

However, don't imagine that KC is entirely urban: they cover city, town and surrounding villages.

I think they end up covering around 170,000 premises, where around 115,000 are in the city itself; around 25,000 are in the immediately neighbouring "suburban" villages, and another 30,000 spread around in separate town & villages.

The rollout so far has not particularly concentrated within the city itself. The surrounding villages are a tad richer.
Posted by ivhr about 1 year ago

Perhaps it could be done similarly to the NBN in Australia(before they tore the plan up and are trying to build a cheaper/nastier VDSL network than BT's).

BT installs FTTP to exterior of the property in large scale, then arranges internal installation at homeowner convenience. Customer's ISP arranges migration to new network.

BT then declares that in 18 months after RFS (this is the Australian deadline), copper is dead except for ISDN and other services for which there is no current replacement.

BT would need to find some way of replacing ISDN (SIP to ISDN gw?)/Redcare/etc though

Posted by CarlThomas about 1 year ago
Andrew - I think it's fair to say that PON isn't going to make them happy either. They want to be able to unbundle the fibre so only PtP will do.
Posted by andrew (Favicon staff member) about 1 year ago
The Aus NBN even in previous guise is a poor comparison as was not full FTTH, but perhaps 80% of Aus as target, rest being wireless or satellite. Also it was always with an eye watering cost, even their VDSL2 version is not cheap.
Posted by ivhr about 1 year ago
@andrew - it was 93% FTTP, but (at least before it was neutered) was predicted to deliver a good ROI. Most of Australia lives in urban or suburban areas anyway - relatively few people would be on LTE or satellite

Despite the change to VDSL, that eye watering cost is growing by the day (it's costing more to patch up the copper than predicted) and there's no upgrade path. It's predicted that it will cost more with VDSL (+ G.FAST/HFC) than it would have cost to do FTTP. Are there any parallels to be drawn BT's situation?

I don't see why it's a poor comparison, it's what we should do IMO

Posted by ivhr about 1 year ago

Wouldn't it be fairly prudent to build PtP capability into any large scale FTTP rollout, if BT ever did this?

It'll make leased lines and other services a lot cheaper and quicker to deploy - and you might as well futureproof as much as you can while you're digging up the roads anyway.

Quoting the NBN again, their "fibre distribution hubs" (for the lucky few who got FTTP) have lots of spare fibre, and has been intentionally designed to allow both NBN GPON and PtP services to co exist, with a premises being able to change from one to the other with a simple repatching at the FDH
Posted by Blackmamba about 1 year ago
Hi Broadband Watchers.
In Surrey they are off loading FTTC and hitting the very long lines with Fiber to the (Post Codes addresses) then for the customer to pay the rest if required a cheap way to cover the area.
Posted by chilting about 1 year ago
It would seem to me that any national rollout would need to be over an extended period.
Maybe even 20 years or more?

Companies and especially Governments don't usually think in such long time scales, but in this case it would seem to be the only feasible way forward.
Posted by AndrueC about 1 year ago
FTTPoD would be a reasonable way forward if BT could work out how to offer it at a slightly more sane pricing level. If people need fibre let them pay (a not stupidly high price) for it.
Posted by WWWombat about 1 year ago
NBN v1 was /planned/ to reach 93% with fibre, but was already off the rails cost-wise and coverage-wise when the election brought the MTM model.

I like Simon Hackett's presentation on this front:
Posted by WWWombat about 1 year ago
PtP capacity...

It appears to have been happening, even if it isn't very visible publicly, and the dimensioning certainly isn't visible.

The aggregation nodes are BT's equivalent of NBN's fibre hubs - with splices rather than patch cables.

These nodes seem to come with space allocated for "P2P BAU" fibres, according to the user manual:

That suggest 20% of the fibre in the spine might be allocated to "P2P BAU", which I read as "point to point business as usual", or "build when ordered".
Posted by ivhr about 1 year ago

It was still in its infancy when there was a change of government and change of ideology, though - just as I'd fully expect a large FTTP rollout by BT to hit issues (though maybe less of them than the NBN given that BT should know what to do by now)

(and honestly, I'd rather see a cost blowout with eventual rolls-royce service, than the same blowout with Ladas for all)

Is 20% enough? It might be if BT assumes everyone will use OR GPON and not PtP links for all. Hopefully they've got that right (or would change it for a large scale rollout)
Posted by shamus72 about 1 year ago
Instead of wasting money on High speed trains we should invest that money in High speed Internet IE Fibre to the cabinet at least for 99% of houses.
Posted by Michael_Chare about 1 year ago
What is provided for a full fibre roll out? Is this a connection point on each property? Quite a bit of the cost can be the final connection to a router, this does not have to be done for those that don't want it. Alternatively it can be done without the use of Openreach resources.
Posted by WWWombat about 1 year ago
NBN in its Infancy?
Created 2009; Trial 2010; Live 2011; Election late 2013; Refocussed 2014.

In mid 2013, they'd passed 210k premises, but one third couldn't actually order a service. Just 140k premises.

BT were running a year ahead: trial 2009; live 2010. By mid-2012, the same level of infancy, they'd reached 11 million homes, passing 115k per week.

Obviously it is faster to roll out FTTC, but for NBN to achieve just 1/80th is worse than diabolical.

Infancy is a poor excuse.
Posted by WWWombat about 1 year ago
20% PtP?

I wouldn't expect PtP fibre to ever be a product aimed at the residential market. GPON will go to 10G-PON in preference.
Posted by ivhr about 1 year ago

NBN had the downside of having to do everything from scratch, including working out how they were going to do it (contract it out to Telstra or build and operate?) having to negotiate with Telstra and Optus for pole and duct access, new billing/operational support systems, new core networks, the lot. I would fully expect that even if they had gone VDSL from day 1, they would have a lesser rollout than BT.

BT took the existing 21CN and ran some fibre to a cabinet, in some cases to the premises.

As for PtP, it was suggested that Sky and TT might want PtP fibre. Hence the discussion.
Posted by litesp33d about 1 year ago
Some people might need FTTH. However what should be done is the Universal Services Obligation needs re-writing to take it from 28kbps to 'any UK phone line should be able to get minimum 2 meg'. Most connectivity problems would be fixed by this.
Posted by andrew (Favicon staff member) about 1 year ago
The USO is due to undergo a transition to a 10 Mbps minimum, discussions on the rules in Sprint
Posted by adslmax about 1 year ago
I think Openreach will bring FTTPoD more cheaper when they feel ready after G.Fast!
Posted by TheEulerID about 1 year ago
Incidentally, I think the real question is not whether the UK could afford £2.6bn a year investment. It clearly could as that's less than 0.1% of the £3bn UK GDP. The real question is whether there's a sustainable commercial model for it. OR has an annual turnover of £5bn, of which only about half is wholesale line rental. That really isn't sustainable unless there's a considerable source of extra revenue, at least until the legacy costs of the copper network dropped out (which might be a couple of decades in full).
Posted by TheEulerID about 1 year ago
Also, my back-of-the-fag-packet calculations that I did some time back to work out the workforce size was to assume that about £2bn a year would be needed to cover labour costs (all in, so taxes, pensions, equipment etc.) A realistic cost is going to be of the order of £40K per year, which gives a workforce of about 50,000 (that's all trades - not just network people). Over 10 years, that's about 100m man-days, or around 4 man days per premises.

So a lot smaller workforce than Andrew's calculation many (most?) of which will be contractors.
Posted by andrew (Favicon staff member) about 1 year ago
The latest Vodafone lobbying is saying it costs too much in the UK, so lower prices from Openreach, so as well as wanting Gigabit they all want it rolled at with a revenue stream to the person doing the work of under £8/month.

20 million people paying £6/month is revenue of £1.4billion per year, see the problem, and worse in early years with lower takeup. If you can add voice then easier, but doubt that would go down well.
Posted by TheEulerID about 1 year ago
One way or another the sums don't add up. If people want more investment in the basic network infrastructure then more of the revenue has to go into those wholesale products. A bit of that is happening with GEA-FTTC and GEA-FTTP of course.

If all fixed line BB connections on the BT network were forced onto GEA-FTTP, then (at the lowest current cost), that would mean around £1.8bn extra revenue at the current lowest rate (£82.80 per year) assuming there wasn't a mass exodus to VM. Higher speeds would mean more revenue if the premium could be maintained. It would also kill LLU of course.
Posted by jabuzzard about 1 year ago
The biggest problem with FTTP as it stands is that Openreach won't put the fibre down the same ducting as any existing telephone line to the premise. So you have the crazy situation where you have a property with a perfectly good duct that is shortly to become redundant that could happily take a bit of fibre but no you have to put in new ducting. Work around that and I don't believe it is beyond the whit of man to do so, and you will slash the time and cost of the FTTP deployment.
Posted by jabuzzard about 1 year ago
I will also add that Andrews calculations seem to ignore the fact that many of the people involved in the FTTC rollout are contractors and not BT employers. It would be entirely reasonable to assume the same for a FTTP rollout, so the numbers are all wrong anyway.
Posted by TheEulerID about 1 year ago

In many senses it doesn't actually matter who the people work for. They still have to be paid, whether as direct employees or as contractors although there are long term issues with the former. You can't simply recruit 30,000 people and dump them after 10 years. Whilst my calculation of the workforce size is a lot lower than Andrew's, it's still very substantial and many of the skills are in short supply. Jersey Telecom only have 100 employees on their fibre installation yet have to advertise all through the UK and still have major recruitment problems.
Posted by TheEulerID about 1 year ago
nb. Scale up the Jersey fibre deployment workforce to the size of the UK based on population ratios and it's around the 65,000 mark (albeit Jersey is much less demanding than the UK as it has a very high population density).
Posted by themanstan about 1 year ago

FTTP is far more manpower intensive... so a comparison to FTTC is not applicable.

The reason for the FTTC deployment is that it is cheaper, faster and less manpower intensive than FTTP...
Posted by Kebabselector about 1 year ago
FTTC would be nice.
Posted by andrew (Favicon staff member) about 1 year ago
@TheEulerID The sums are interesting ahead of Thursday I thought this was a nice talking point to show that while we can wish and support full national Gigabit roll-out the scale and cost of the task needs to be seriously considered before heading down the path. The other option is keep workforce same size as now and look at a 20 to 30 plan.
Posted by TheGuv about 1 year ago
With the move to drive prices down I'm not sure how people expect infrastructure companies (whether Openreach, CityFibre etc) to make money.

Openreach need to provide a better/accurate plan on delivering better broadband (so they have nots can have a better idea of timescales). Will the "have nots" accept that improvements will come by more stringent dates?
Posted by TheEulerID about 1 year ago

It certainly is an interesting point, and it's about some time that some of the more blue skies types really did a few calculations, even if they are a bit crude. What has happened in the UK market is pretty speed predictable given the regulatory approach and use of LLU on copper to drive down retail pricing. The big separation debate would not, in itself, change the economics which are really about economic viability. It's the money men that need convincing as without finance things go nowhere.
Posted by DrMikeHuntHurtz about 1 year ago
BT proposed this in the early 90s and Thatcher shot it down.

Perhaps if the government got out of the way, BT could swap copper for fiber.
Posted by TechnoLion about 1 year ago
Is there any country that has done an FTTH rollout on a similar scale to what would be required?

Going with FTTC seems a pretty prudent move for BT. They got to large-scale deployment quickly, and a price low enough to get a significant takeup. Unlikely either would have been possible with FTTP.
Posted by Oddball 12 months ago
Everything is always a transitional stopgap measure to save on costs. I find it shortsighted of BT not to put in the infra structure for FTTP/H when installing the cabs. They have not even implemented vectoring at the cabs nationwide yet (I know they upgraded a portion of it for a small speed bump but we are still a way off vectoring it would seem). BT drag their feet at every opportunity. FTTP should have beem set the groundwork for years ago.
Posted by chrysalis 12 months ago
The answer is a clear yes, 2.6billion is not much money to the uk .gov, as always its whether those in power want it or not.

What I dont like is that people seem to think its a choice of either no FTTP, or nationwide, no cities only option.
Posted by andrew (Favicon staff member) 12 months ago
@oddball - they are putting in the extra infrastructure, but only as far as the aggregation node that is near to the cabinet.

So rather than fibre being in ~5,5000 exchanges, it is native for ~250,000 premises and infrastructure also pushed to 74,000 or so locations ready for further roll-outs be that FTTP or or something else.
Posted by WWWombat 12 months ago
I think .gov has decided it doesn't need to spend on ultrafast coverage; not yet anyway.

I think Ofcom has seen BTs announcement to push on with copper enhancements, and VMs announcements to push on with DOCSIS enhancements, as leaving the door open for a third network that is fttp-based. They're doing their best to set things up to make the proposition inviting.

BT and VM left the door ajar. Ofcom stuck their foot in it. Will someone shove?
Posted by JacktheMac 12 months ago
I don't understand many of the technicalities mentioned above, but I do know that Gigaclear are rolling out FTTP to thousands of properties quickly, cheaply via a brand new network and with the minimum of fuss.

Posted by kmendum 12 months ago
@JacktheMac. Gigaclear have a nice new site to deal with plus an exclusivity deal with the homeowners. There is a single equipment cabin at site, with the 1G circuit to BT or (preferably) another telco for Internet access. No messing with legacy copper, 4 simple piped services to each house.
Posted by WWWombat 12 months ago
Gigaclear aren't doing this especially cheaply - about £1,000 per home passed in their chosen areas, and about £1,500 per home in their BDUK contract areas. Plus £200+ when you want the house actually connected.

Where they choose to go, there is usually nice soft verges to use, and a tenant business alongside nearby backhaul.
Posted by JacktheMac 12 months ago
Kmendum – 1. only exclusive for 12 months. 2. BT effectively mothballed their copper networks in rural area long ago. 3. If GC can do FTTP, why not BT, who have far greater capital reserves ?

WWWombat – GC quote £1600 - still peanuts, I’d happily pay £1k as a one-off to get FTTP. And is ‘nice soft verges’ all that’s stopping BT ?
Posted by chilting 12 months ago
BT are installing FTTP, lots more than Gigaclear - just compare both companies coverage areas.
The trouble is that it takes time, capital and labour - all in short supply.
Take a look at the situation in Hull where they are going for FTTP - can you imagine if the rest of us had to wait that long on slow ADSL!
Posted by JacktheMac 12 months ago
@chilting; Where are BT doing this ? What’s the product called ? Gigaclear have just proved that while financial backing is needed, it takes VERY little time and labour to install an FTTP network from scratch.
Posted by JacktheMac 12 months ago
@chilting: BT would never, ever even have improved the intermittent sub-2Mbps service our village experienced. They told us a fibre connection was ‘scientifically impossible’ and ‘made no economic sense.’ BT would never have brought FTTP to us - they couldn’t even be bothered to supply FTTC.
You must be logged in to post comments. Click here to login.