TalkTalk and Virgin Media Moot UK Broadband Infrastructure Sharing
The rumour mill has burst into full swing this morning with reports that UK ISPs Virgin Media and TalkTalk are tentatively exploring the possibility of passive infrastructure sharing, which could enable the budget provider to lay their fibre optic lines in Virgin’s cable ducts and vice versa.
At present most of TalkTalk’s broadband ISP services piggyback off Openreach’s (BT) national copper and fibre optic network, although they have invested in their own 1Gbps capable Fibre-to-the-Home (FTTH/P) network to cover the city of York (here). On top of that they recently announced a joint £1.5bn deal with InfraCapital (M&G Prudential) to provide ultrafast broadband via FTTP to more than 3 million premises in mid-sized UK towns and cities (here).
Meanwhile Virgin Media’s Project Lighting has been busy investing up to £3bn on a national expansion of their own 350Mbps capable ultrafast broadband network, which aims to reach an additional 4 million premises by the end of 2019 or 2020 using a mix of hybrid fibre coax (HFC) and pure fibre optic (FTTP) lines via DOCSIS / RFOG technology (around 60% UK coverage).
Suffice to say that any sharing of passive infrastructure would deliver a huge benefit to TalkTalk’s plans by cutting their civil engineering costs (as well as speeding up their timescale) and that seems to be precisely what is being reported in the Telegraph. The talks are currently said to be at an “early stage” and neither side is willing to comment.
Virgin Media and parent company Liberty Global have generally always preferred to retain a closed network (except for corporate connectivity solutions) and thus working with a rival ISP like TalkTalk in such a way would represent a major shift. The newspaper report also suggests that Virgin could benefit by following TalkTalk’s build into new territory, but we’re sceptical.
The cable operator already has good coverage in urban areas and TalkTalk’s rollout will no doubt target some of those same locations. Certainly by working together and perhaps sharing rollout plans then a good degree of overbuild could be avoided, although on the surface this would still appear to benefit TalkTalk more than Virgin. At least in the short term, until Virgin goes DOCSIS 3.1 and we see 1Gbps packages that can compete.
On the other hand Virgin Media do appear to still be lagging behind with their Project Lightning deployment and anything that can help them achieve their rollout plan on-time, as well as within budget, might well be something that they’d consider exploring. In the meantime we’d take this all with a big pinch of salt.
Lest we also forget that Virgin Media and TalkTalk once proposed a joint bid for funding from the government’s Broadband Delivery UK scheme in 2011, which might have resulted in the construction of a rural open access FTTH broadband network for 5 million homes (here). The proposal never amounted to anything but it shows that the two have form in working together.
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.
« Over 50 Cross Party UK MPs Call for Better Rural 4G Mobile Coverage
Latest UK ISP News
- FTTP (5531)
- BT (3518)
- Politics (2542)
- Openreach (2298)
- Business (2266)
- Building Digital UK (2247)
- FTTC (2045)
- Mobile Broadband (1977)
- Statistics (1790)
- 4G (1668)
- Virgin Media (1621)
- Ofcom Regulation (1465)
- Fibre Optic (1396)
- Wireless Internet (1391)
- FTTH (1382)
Also worth bearing in mind that the TalkTalk LLU backhaul network uses leased fibre from Virgin Media in places, eg North Scotland despite VM not having a retail (cable) presence north of the central belt. I believe Sky also have a similar arrangement with VM in places.
VM don’t have any core fibre in Northern Scotland. Indeed they lease some fibre to backhaul their cable network.
https://www.slideshare.net/VirginMediaBusiness/optical-networking-trends-the-virgin-media-business-view – slide 2.
@CarlT
I wasn’t saying there was a significant project to replace coax underway at present. I’m perfectly sure they will stagger on with their horribly unreliable network until absolutely forced to upgrade/fix it.
If the coax was mint then I totally agree that I would have the reliability/utility for decades to come.
And I also agree that the main issue with VM capacity is from the nodes back to where it meets the real fibre.
I’m afraid in some areas it was direct buried: I can personally remember it being installed in a few places that way. It all depended on which of the very many individual cable-co’s carried out the work and the level of financial pressure they were under when the work was done. In most case every expense was spared.
The nodes *are* where it meets the real fibre. Nodes short for optical nodes.
There might be some directly buried stuff but it’s such a minor consideration it’s not even on the radar right now, not least because there’s so little utility in direct burying – those same cable runs are usually wanted for drops, so swept tees and ducts are a better option. I’m not aware of anywhere where cables were direct buried to homes, always ducts. If there’s ducting outside homes why put the coaxial trunks directly in the ground?
There are some areas where the builders ran out of cash where there are empty ducts and cabinets.
Any areas where every home was served in a MATV network, the most likely candidates for direct burying, have been overbuilt or retired now.
Regarding running power to cabinets – it’s not an issue. It’s already there and has been since the beginning.
“Lest we also forget that Virgin Media and TalkTalk once proposed a joint bid for funding from the government’s Broadband Delivery UK scheme in 2011, which might have resulted in the construction of a rural open access FTTH broadband network for 5 million homes (here). The proposal never amounted to anything but it shows that the two have form in working together.”
—— or not
Lots of talk, but no action.
If they do do this it could mark the start of some proper competition for OR, and would be welcome certainly by me 🙂
I do have this feeling though that it may be a ploy to get VF to think a bit more seriously about a purchase of VM, while talking up talk talk too.
I had the same feeling 🙂 .
It might also be another indication just how cash-strapped TalkTalk is, following on from its recent FTTP deal. Anything to minimise CAPEX since it doesn’t have the funds to compete with the big players.
Possibly
But also VM are at some point going to have to do something about all the direct buried coax some of which must be getting close to EOL. There is a distinct choice between making all VM fibre to the cab (and getting rid of all of the splitters on splitters on splitters) and keeping the last 100m of coax or changing over to full fibre. And that is a toss up between running power to a lot of cabs to power the active fibre to coax or keeping it fully passive with fibre.
Wasn’t aware there was a mass of direct buried coax in the VM networks. In fact haven’t known any that are direct buried, all ducted. Might be some outliers but most definitely not the majority or even a significant minority.
There’s power to virtually all cabinets already.
In fact a very short bit of research indicates there is no significant project to replace the coax in the HFC networks in any significant quantity nor is it required. There is no plan beyond business as usual splitting of nodes to overbuild coaxial with fibre.
The only real network upgrade work is upgrading the equipment that’s in cabinets already, alongside some in-home installation replacements.
The coaxial plant isn’t having 1/3rd of its capacity used. The restrictions are in the active network.
Another reason for them/BT not to do FTTP is the lit fiber tax.
Neither of those two pay that tax in the same way others do. VM pay a tax per premises passed whether coax or fibre. BT’s taxation arrangements don’t change whether it’s fibre going to properties or copper.
The business rates were there before fibre was really a thing.
Nice to read someone with real knowledge correcting all the regular dud information. Nice work CarlT
@Alan
I’m not at all sure that it is real information.
Following @CarlT’s responses I was quite concerned that I had posted dud information, I don’t like doing that, and so I have spent quite a bit of time doing some proper research.
The slide that @CarlT seems to be (mis) quoting at
https://www.slideshare.net/VirginMediaBusiness/optical-networking-trends-the-virgin-media-business-view
States that
“Over 37,000 powered street cabinets all with fibre attached now”
It does NOT say that all cabinets are powered and all cabinets are fibre attached as my quite thorough research has shown that would not be a true statement.
37k cabinets sounds like a lot when compared to BT DSLAM’s but some of the cabs only serve groups of around 24 houses.
As I understand it, and it is nearly 20 years since I was designing HF filters and splitter setups as part of my day job, the VM network is structured:-
a) big VM cabs are powered where the fibre is converted into the big fat double screened coax; and
b) smaller local cabs are purely passive as all they have in them is large multi notch splitter/filter and a big pile of small passive splitters
My understanding of the VM capacity issue is that it lies between the big active cabs and the small passive cabs down to the fat double screened coax and the way that it is used. This is particularly an issue where the fat coax is daisy chained to multiple small passive cabs.
And this depends on how the original network was architected way back when.
Granted some of the newer installs seemingly have the fibre going straight to an active cab with 3 (or more) big splitter blocks on the backplane and this goes directly to homes. In this setup my comments are not relevant and others are correct that the only limitation is the fibre backhaul.
So given the intensity of discussion, and debate over fact, that kicked off here I did wonder if I had got it totally wrong and so I called someone I used to work with who did HF/RF design work for VM and he confirmed that was the way the network used to be set up a few years back when he was involved. He was quite clear that the small satellite cabinets were not powered.
My comments on EoL related to how much bandwidth you really can stuff down even a very good quality fat well screened coax. In some cases hundreds of homes are connected to the one piece of fat coax, via the daisy chain arrangement of the cabs on the fat coax. There is a physical/practical limit and it is utter fantasy to think that this sort of architecture can cope with hundreds of homes at 1G/1G. Further at some point you have to accept that the RF properties of the coax are degraded/degrading. Coax depends on the air gap being dry gas and is therefore sensitive to degradation.
There is no practical problem with getting 1G/1G down a single decent coax that runs from the cab to the property that after all is the point of DOCIS3.1 etc. And this is very well established and accepted tech. So provided the coax is in good shape all is well.
Coax isn’t fibre you cannot just change a couple of media modules at either end and get another order of magnitude down the coax. How much you can get down the fat coax is also length dependant and the resistance, capacitance, inductance and general dielectric properties all play a cumulative part. And that changes with ageing.
Stir into that, high frequency equipment does behave in strange non linear ways as it ages or contact corrosion provide slight reflectance of signal. Non linear behaviour is exactly what you do not want in a data system.
Hence my perhaps not very well expressed comment that at some point VM have to bite the bullet and power the satellite cabs and take fibre to them. The balance is wether it is cheaper to just pull fibre through, if indeed ducted I would be interested to see CarlT’s source for the % of the VM network that is/is not ducted, and keep the whole thing passive or to spend money producing powered cabs closer to the end point user.
And still wanting to be 100% sure that I had got this right I asked one of the VM engineers working close to one of my sites last week, wether I was right in stating that some of the cabinets were still not powered. He was absolutely clear that a lot of the “end of chain cabs” were not powered. He cheerfully showed me the inside of the cab he was working on: no sign of any power. Totally passive.
I’m happy to be proved wrong, I don’t believe that I am now but was worried that I was, but that does mean citing primary sources.
Shared infrastructure is the way forward for fully utilised investment however what is inferred here is that TT will fibre the same areas as VM are present which means they will become direct competitors in the same locations along side OR offerings. Surely they should avoid each other and have a shared network arrangement.
That is too much like common sense to be a real plan.
But I don’t think Virgin have the wherewithal to allow TalkTalk to use the existing cable system as a competitor alongside themselves. SO if TT want to sell into Virgin cable areas then they need to build their own infrastructure.
With a new FTTP system, then there are no issues with sharing common infrastructure. You just create a new Openreach type company who do all the common work. Then VM and TT can sell their own packages on top.
Surely we want more ultrafast coverage not multiple providers in the same streets.
What you want and what’s profitable don’t necessarily correspond. Operators aren’t going to go out of their way to avoid one another if they feel there’s a business case to overbuilding another network and many rural areas are likely to be waiting a while or require taxpayer subsidy, for a change.
Multiple providers and deployments means more choice for the consumer, why any normal customer would have an issue with that i do not know.
TalkTalk and Virgin Media? What could possibly go wrong?
Worth bearing in mind that talktalk (and all other service providers) have been able to do this with bt ducts for over 10 years and haven’t done.
Openreach’s original PIA solution had a fair few problems and some key limitations, which is why Ofcom’s 2016 Digital Review pushed for greater change. A few smaller ISPs did use the original PIA but it seems to be becoming more popular since the enhancements.
https://www.ispreview.co.uk/index.php/2018/02/ofcom-changes-boost-uk-full-fibre-broadband-cut-fttc-prices.html
Virgin already use Talk Talk and Vodafone tails for off-net connectivity and have done for sometime.
I have heard thought that Talk Talk may start using Virgin tails also although I’d prefer them to stick with BT as Virgin are a nightmare to work with from a civils point of view
It’s not very obvious to me it would make much economic sense for either side of that conversation to put in a parallel “last mile” network as, even using existing ducting, it is not going to be cheap and could pirate VM revenue (albeit, no doubt, that would be partially recovered through the wholesale charges).
What would make more sense is to uplift the local VM network, but rather than the expensive and laborious exercise to replace the coax, that could probably be done more cost effectively by upgrading nodes and/or reconfiguring the loops where feasible. It doesn’t meet the “it must be 100% fibre” folk’s demands, but a great deal more can be achieved.
Of course, if the existing co-ax is nearing EoL, that might be an issue, but is there any sign of that?
One downside of this is that DOCSIS doesn’t lend itself to unbundling. It would be very much a virtual service.
nb. in the area where I used to live, it wasn’t so much the below-ground assets that were showing their age, but the above-ground VM equipment. Lots of rusty boxes, often half fixed to the plinths and sometimes with holes and doors hanging off.
Oh it wouldn’t be unbundled, it’d be bitstream if they were to use an active option.
Unbundling it without causing VM grief would be a PITA involving building an overlay network. May as well simply use the passive infrastructure.
Which is why I called it a virtual service. Of course, that’s also true of the OR GEA services.
In any event, it’s unclear that there’s a business case for a co-ax replacement or a fibre overlay on existing VM networks.