Virgin Media scraps wholesale network rival to Openreach Virgin Media has ditched plans to use its network infrastructure to create a UK national fixed line operator to rival BT's Openreach just 18 months after the project was made public. Last year, Virgin Media O2 (VMO2) disclosed plans to form a national fixed network operator (NetCo), backed by its shareholders Liberty Global ( …
That's a shame from a technical viewpoint. It would've been interesting to see how they managed sharing network capacity. Either their wholesale customers would've given up in disgust or else perhaps we'd have seen non-VM CPs offering higher average throughput and lower/more consistent latency than VM on 'their own' network.
It would also have been interesting to see how it affected the overall market.
Oh well.
Presumably it'll come back again. You know the tech and the industry history so I can skip that bit, but it's all very well for Telefonica to mutter about the debt problem, that doesn't make it go away. In ROCE terms, VMO2 are destroying shareholder value. Either the Christmas Fairy comes along and hands over a few billion to reduce the $21 billion of VM debt, or some of the creditors/parent companies take a heavy haircut (unlikely), or VMO2 have to find some growth. The existing business has only seen fixed line customer numbers bounce between 5.7 and 5.8m, they've failed to achieve real terms pricing or margin increases. They've sold down other assets like Cornerstone, the VM brand seems unable to grow (perhaps too many ex-customers like me who will never come back), which really only leaves the hope of wholesale growth to make better use of the network.
Or perhaps the hope is that, just as Telefonica were mugs for merging with the debt laden, ex-growth hulk of VM, they now hope that some greater mug can be found, eg a large US media outfit daft enough to want to own a distribution network.
Possible, but I guess unlikely. Much as I loathe the company, they did invest a lot in upgrading their network, and the backhaul was very, very rarely a problem when I spent time as a superuser in their help forums. Where capacity constraints existed these were 99% of the time on the DOCSIS circuits, usually at the CMTS (the big mega-modem that sat at the head of the essentially analogue cable distribution). I suspect but don't know that VM's network architecture people were very good and really understood their network, and because they were always thinking ahead to the next speed ramp-up to stay ahead of Openreach, they were always mindful of the capacity of each element. Their main problem has always been that DOCSIS had a best-before date some years ago, but VM were putting off the evil day of going FTTP, so they had to stick with the old co-ax technology on the local circuits. I believe they're now part way through changing to FTTP, but whether that's being rolled out to the entire network and existing customers in any credible timescale I don't know - as announced, they were going to put in fibre capability for each local network, but only going to put in FTTP links for new customers, or those upgrading their contract, so existing customers......yeah alright, I'm boring myself, and I'll STFU
AndrueC probably has a much better knowledge of all these matters?
Sadly, my expertise is more around Openreach/DSL. I think VM were always just a bit stingy with the local loop as you say. I think they liked to offer the highest speeds by just upgrading the nodes then hoping that few people would actually sign up to those services. Looked good in advertising but could lead to problems for customers.
It'll be interesting to see how GPON handles contention. It certainly isn't immune as there was at least one complaint I saw on a forum that required engineers to come out and add a CBT. That was blamed on someone editing TV programmes at home which is a very bandwidth hungry process though.
It'll be interesting to see how GPON handles contention. It certainly isn't immune as there was at least one complaint I saw on a forum that required engineers to come out and add a CBT.
GPON (Gigabit Passive Optical Networking) handles contention pretty well, or at least predictably. So it's basically an OLT (Optical Line Terminal) at one end, then ONTs (Optical Network Terminals) at the customer end. Then a 2.4Gbps down / 1.2Gbps up between OLT & ONT with passive splitters muxing that to service up to 128 ONTs per OLT. Then it uses TDM (Time Division Multiplexing) to time-slice the bandwidth between OLT & ONTs.
Which then gets into creative marketing and the 'up to' statements. So each user/ONT gets say, 1Gbps depending on service definition beyond the OLT/ONT, but only for 1/128th of the time. Which can then become an issue for the service layer, eg Ethernet or IP. So a single user shouldn't be able to hog bandwidth at the GPON level because every ONT gets the same timesliced bandwidth. That user might complain they're not getting 'full 1GE service' though but that's marketing for you. Network performance though comes down to how many times you split the bandwidth between ONT & OLT.
It's also why wholesaling gets tricky. Technically, it's doable with VLAN seperation between headend Ethernet boxes and NIDs, but with a network based on GPON, still has the TDM issues. Which might be fine for broadband Internet, but is NBG (No Bleeding Good) for a wholesale customer that wants to get a 'real' Ethernet. But those are harder to find, ie plain'ol fibre or DWDM derived Ethernet as carriers move to using GMPLS and packet based delivery. Then it's studying service descriptions and SLAs to figure out what you're actually getting. But VM Wholesale generally designed their services using seperate infrastructure to retail, which was part of the seperation plan, ie the Wholesale OpCo gets the infrastructure and Retail becomes a customer. So similar to the BT split between Openreach and the BT service divisions.
Which then gets financially challenging, so Wholesale gets a lump of cash for providing Retail's network which then has to cover all the network opex, plus some potentially big capex items. GPON is kind of old tech, and I think some of VM's network was upgraded to NG-PON2. That combines TDM with WDM for 'up to' 10Gbps and the WDM elements use different wavelengths to GPON, so both networks can co-exist using the same fibre. And then next up is HSP, or High Speed PON with 50Gbps and backward compatibility with GPON and NG-PON2.
So scope to upgrade bits of network as needed from GPON->NG-PON2->HSP by filling street cabs & PoPs with new boxes, but that also means backhaul kit needs upgrading from 2.4Gbps->10Gbps->50Gbps. Plus having that core able to deliver plain'ol Ethernet at 1/10/100Gbps for wholesale customers building their own networks.
But such is the life of a network designer. Have a network rolling out while planning for the next upgrade, then FUN! meetings with marketing & finance types to pay for all that when customer perception is to get more for less. It's also amusing that the good'ol days of SDH and TDM, you ended up with 'mux mountains'. SSDD, and PON broadband presents the same challenges.. With some addition quirks, like how to get enough power to a streetcab with a pile of muxes and other tin crammed into it.
Of sure why you got downvotes, that was an interesting read from someone who just works with lan and wifi designs.
It's just politics. I have a skiddie who's automated downvoting any & every post I make, then a couple of admirers who'll downvote just because I dare to express opinions on controversial subjects. You seem to have picked up a downvote probably just for being positive, so have an upvote to balance that!
Interesting read, but incorrect on where VMO2 are and looks to be written by an American as no-one in the UK is using NG-PON2 and we typically call it 50G PON on the right hand side of the pond. (I didn't downvote)
VMO2 have two networks of their own and their join venture Nexfibre (which VMO2 are the only user today).
VMO2's legacy network is DOCSIS 3.0/3.1 HFC network (3.0 where they used RFoG).
Their new network (Project Mustang) and Nexfibre is XGS-PON, apart from Openreach this is what just about everyone is deploying in the UK today.
Openreach wholesale on GPON successfully in the UK, they've got millions of customers on their network all wholesaled, handoff is local to the end-customer.
CityFibre wholesale on a mix of GPON(legacy) and XGS-PON both locally and nationally and were targeting 1million customers this year.
CityFibre also deliver ethernet services over XGS-PON.
Both GPON and XGS-PON are TDM, but the allocation is flexible based on CIR (committed information rate) and EIR (Excess Information rate) and other parameters. The ONTs have a management channel to the OLT (OMCI) that enables the slots to be agreed based on the PON load and the configuration for each service. CityFibre and Openreach sell over 1G on GPON and CityFibre launched a 2.8 and 5.5G service with Sky as the ISP last month.
Interesting read, but incorrect on where VMO2 are and looks to be written by an American as no-one in the UK is using NG-PON2 and we typically call it 50G PON on the right hand side of the pond. (I didn't downvote)
Cheers. I'm mostly British and generally stick to the backhaul/line (Nokia-Infinera) side of networks now. I have been involved in a couple of NG-PON2 rollouts, and yep, those have been on the left side of the pond. I think it also demonstrates one of the challenges of networking & the joy of standards. NG-PON2 doesn't exactly roll off the tongue and now a bit curious what 50G PON is given-
https://www.itu.int/rec/T-REC-G.989-201510-I/en
VMO2 have two networks of their own and their join venture Nexfibre (which VMO2 are the only user today).VMO2's legacy network is DOCSIS 3.0/3.1 HFC network (3.0 where they used RFoG).
I've lost track of where VM's at with all their M&A but did some work with them around some mobile backhaul projects. So at that time, they had 3 business units, wholesale, business and retail, which pretty much meant 3 networks. The DOCSIS bit being retail, and wholesale selling DF and other capacity mostly using Transmode DWDM kit. That worked, but was a bit clunky and burned fibre and space at VM's PoPs. At that point, I know there were conversations about the future and strategy given the way it had already grown via merging cable operators, had spent a lot of money integrating that and needed a lot more investment in network. So then came Liberty, VMO2 and the rest.
I think the biggest challenge was their last-mile access for broadband was majority coax. Going DOCSIS 3 meant fibre to the cab, but still coax on the customer connections. But then market pressure was for FTTC or FTTP, so a lot of expense needed to replace their copper loops with fibre.. Much the same as with BT.
Openreach wholesale on GPON successfully in the UK, they've got millions of customers on their network all wholesaled, handoff is local to the end-customer.
Yep, and those have to be replaced because of the Huawei edict, which gets interesting given there's funding to pay for that replacement, but gives BT an opportunity to upgrade from basic PON to something newer which otherwise would have had to be funded by BT's BUs. So they maybe get a 'free' upgrade to XGS or NG-PON and become more competitive with the alt-nets.. Which would include VMO2 / Nexfibre, and whatever they might want to do with their wholesale busiiness given BT still has the incumbent advantage.
But it's an expensive game to play. I had a look at this lot-
https://the4thutility.co.uk/
who dropped a flyer through my letterbox and-
https://the4thutility.co.uk/news/the-4th-utility-hits-the-ground-running-in-may
With the ramp up of recent investment and increase in demand for Ultra-Fast fibre broadband the 4thUtility has been working hard, very hard. They have successfully completed over 1000 installations in homes across the UK, during April alone!
https://the4thutility.co.uk/news/4th-Utility-secures-game-changing-25m-investment-from-NIBC-Bank
Fast-growing full fibre broadband provider 4th Utility has received a £25m investment to boost its ambitious plans for expansion.
And I have a CityFibre drop outside the house already. But I haven't looked if 4th has been busy digging, which gets expensive here in MK given it's the land of the brick and lifting, trenching and reinstating those gets expensive.. So the acquistion costs for FTTP are high, prices are low and retention can get tricky if I've got a choice of at least 3 FTTP providers already. Not a business I'd really want to be in.
50G PON is part of the Higher Speed PON standards - G.9804.3, key is like GPON and XGS-PON it is single channel whereas NG-PON2 uses multiple wavelengths. Typically single channel means simpler and cheaper kit.
Openreach have long since met their requirements for FTTP and Huawei, that was no more than 35% Huawei in that area. They did this via a lot of build with Nokia & Adtran GPON kit, so no replacement of the existing Huawei needed. They have stopped using them for the ONTs but the OLTs remain. It is rumoured they have been deploying Combo cards so they could enable XGS-PON in the future on the newer kit, but they've just announced two new GPON ONT vendors. I do wonder if these new vendors is because Openreach are now somewhat of an outsider still actively deploying GPON.
There has been talk of XGS-PON and symmetrical products from Openreach, but at the moment these seem to be only government subsidised locations (BDUK) which I suspect is more down to their contract requiring them to provide such a product rather than Openreach wanting to offer it.
4th Utility does have a partnership with CityFibre, I would hope in MK they'd be using that rather than digging again. In MK CF is probably the biggest fibre infrastructure provider.
Nexfibre are rolling out XGS-PON, the exact flavour I'm not 100% on but it's not plain GPON that's for sure....
wonder if openreach bought way too much GPON kit / got it cheap/ wrote the lowest performing spec possible so they could stick their hand out again for a govt handout - "you see minister they say they can do this but I can assure you that their way of achieving it will be unreliable, we at openreach only use the best tried and trusted ways of working and we rolled out gpon as XGS-PON hadn't undergone our in-house very rigorous testing and even we began it,.well our testing is done over a very extended time period to guarantee reliability and we have only just signed off on it but during the testing we didn't want to delay the rollout so we used our then gold standard of GPON which is now obsolete and it will be incredibly expensive to rip out and replace and not commercially viable without public support" - cue another massive govt cheque being written to openreach.......
wonder if openreach bought way too much GPON kit / got it cheap/ wrote the lowest performing spec possible so they could stick their hand out again for a govt handout -
I doubt that, given about from some rural subsidies there isn't much government cash to grab.. In fact for BT, it's been kinda the reverse. So as an example, my FTTP connection uses a Huawei HG8110-H20 ONT/NID with a router hanging off that, and government has told operators to de-Huawei their networks.
Mostly though it's just the planning cycle, which for Openreach can include getting regulatory and other operator approvals to launch a new service. Sometimes that can take a year, depending on regulator or how difficult competitors want to be. Which can sometimes be valid concerns, ie competitors wanting service will also usually want to do some interoperability testing.
But it can be a slow process, especially for incumbents. Market (or marketing) says 'We want FTTP!' and (hopefully) hands over a service description giving desired features & costs. Then if manglement OK's the service, development starst. So then engineers go to market to see who can deliver a few million ONTs, with an ability to hook that in to existing OSS/BSS with as few pain-points as possible. Plus additional features like being able to slap BT/openreach branding on tin, deliver SKUs that include a wall unit to contain the ONT, battery backup, cable management and fixing screws.
Which is pretty much why Huawei got a lot of this business. Other manufacturers often couldn't, or wouldn't offer the volume, flexibility or price carriers need. UK has around 25m homes, China has.. a lot more. Then it's 'just' a case of preparing a national network for deployment, so inspecting thousands of ducts & cable chambers, street cabs, ordering new cabs if needed.. Which can be FUN! because if new street furniture is needed, that can need council planning approval, especially in conservation areas. Which can also mean mandatory notice periods so residents can object to cabs outside their properties. And also arranging power connections to new cabs, that can take months.
Then you might be ready to go to market and discover government does a rug-pull and you need to start vendor selection all over again, rip that kit out, replace it with something different and run two networks while you manage the migration.. Which is probably why my Huawei ONT hasn't been replaced yet, and when it does it'll be a newer version of PON.
But one of the things I learned quickly doing core stuff is the tin & software is often the easy bit, it's the physical delivery & civils that can be the real killer. Even 'simple' things like being able to deliver against an MTTR.. So having enough field engineers to respond to faults, those field engineers having enough training, spares, tools and consumables to do fixes. Then having enought stores to hold spares or installation kits. And sometimes having FUN! conversations supporting ops vs finance when inevitable cost-cutting subjects come up, like why so many field engineers and inventory sitting in vans and stores? Then patiently showing them maps of where their engineers and stores are, where their customers are, the distances.. And where Google can be fun for showing distances and estimated travel times.. Especially when field engineers aren't blue light services & can't speed, even though the customer who's service is down might be, and network failures would likely make the news and result in big fines from the regulators.
So rarely a dull moment, a fun career, but glad I'm mostly out of it now.
Someone used the bandwidth they contracted to VM for and it caused problems for other customers.
Ultimately the VM network was built for tv and I assume performed adequately for that purpose. Broadband was a bolt on for cable and uploads suffered because of how cable systems are designed for sending content to households and not large volumes of data being returned.
I remember they briefly looked at RFoG as an alternative to pon fttp - meaning they could reuse their distribution tech to save costs but I think they’ve decided to go full in now with Xoom so they don’t have to redo connections again in a few years.
I am increasingly of the opinion that the only way to ensure I can get a decent, reliable connection is to buy several technologies and use them together in a redundant manner.
As far as I'm concerned "BT" and anything involved with it is one single point of failure.
I have an expensive Draytek router that can do Ethernet, VDSL2, and 5G directly. That gives me all kinds of options because, clearly, there isn't going to be any regulation in any one area of Internet connectivity for a long time to come (and it's already been too long).
I thought that scrapping traditional voice service would be it, but apparently not. I'm now with an ISP who couldn't give me a voice service (I don't care, never use it) except VoIP and the line only does DSL. But it appears that any kind of competition / regulation just isn't going to happen in this area. So I have 5G and if Starlink wasn't Musk-borne I might have that too (I'm in the perfect area for it with stupendous clear-sky access). I'll just have to wait for Kuiper, I suspect.
But this week there was hope. Gigaclear are running FTTP lines into my road and down my road (but I don't know if they will come as far as my house, even though it's a tiny road). That'll be a good "third option" if I can get FTTP.
Ironically, I now have better connection in the middle of nowhere surrounded by fields, in an AONB and SSO, miles from anything than I did when I lived inside a major town that you will have heard of, inside the M25, a stone's throw from the town centre. If Gigaclear comes in too... that could be a 1Gb line direct to my door. And given the constant power cuts in the area because it's so rural, that might well be the last to fall over in my area.
It's just pathetic how any kind of regulation, market management, etc. that's been posited for decades is just overtaken by actually just waiting for something new to come along entirely. We could have all had FTTP decades ago, but we're apparently too scared to just nationalise OpenReach and make it do its job.
"but we're apparently too scared to just nationalise OpenReach and make it do its job"
Well it wasn't any good at its job when originally under state ownership, I'm not sure why you think it would be any better trying again? I reckon you aren't that different in age to me, and I can remember exactly how crap GPO and state owned BT were. In fact, not that far from where I guess you now reside, I can recall my parents having to wait three months for a phone line to be connected (even with wiring and ducts in place), calls were expensive, and the best they could give us was a "party line" so that we shared a telephone number with the neighbours.
Many of their "engineers" still seem to think it's the 70s where strict lines of demarcation are the order of the day, that these other "providers" are all a shower of Johnny come lately's usurping the natural order of things, where in their view bt should have stood firm and refused to even countenance any internet provision let alone this fibre malarkey.
Local paper had an account of an employment tribunal where an ex openretch engineer shared messages from a manager telling him to splice in a master socket before a fault and then retest the line so they could "prove" nothing was wrong with it as openreach didn't want to spend the money to fix it and wanted to "punish' the customer for being "difficult" and "not taking no for an answer"
One admitted to my face that they would disconnect people's service out of spite / retaliation for complaints etc and leave notes in cabinets and at the exchange so their colleagues would leave it disconnected sometimes for weeks at a time, and that it was common practice at openreach for engineers to do this and managers didn't see any issue with it.
Says it all really - they are the height of arrogant - during a power cut you don't phone the metering company, instead you phone the DNO, in the event of a gas leak you phone Transco/ SGN, phone outage - "phone your provider and they will interact with us, we don't deal with the general public" - absolutely appalling arrogant attitude and something that should NEVER have been permitted by Ofcom or the govt.
Openretch are LONG overdue to be broken up or better yet turned into a non profit but one where former openretch staff are ideally excluded entirely or have to go through a thorough assessment to weed out the many bad apples with their 70s attitudes from their ranks and cut out the managerial deadwood and obstructionists
How exactly is breaking up OpenReach going to help?
As wired telephony dies and everything moves to digital you have as the options:
BT/EE and all the resellers backed by OpenReach
VMO2
A few niche players
5G mobile
Both BT & VMO2 have invested huge amounts of money to upgrade networks or install fibre. With BT remember plain old dial up modems for those of us old enough? Then ADSL, the maximum speed of which is increased until the physics essentially stopped further improvement. Then FTTC as a stopgap to FTTP that as far as I can tell is good enough for most consumers. Whoever provides that has to invest huge amounts of money to get the infrastructure in. Now the basic facts of business mean that VM covered densely populated areas where there was a reasonable chance of demand giving a return. Like it of not BT/OpenReach had the ducts with copper being replaced with fibre to the cabinet and the final bit to the premises is underway.
I don't see how there is space for another player and just saying "unbundle" does not help unless you open ALL ducts up for any company to use, That is existing BT/OpenReach. VM and anything else that is in the ground.
Now he danger with that of course is just like Royal Mail, other companies come in and take all the profitable areas with none of the investment required so you end up with your core USO now losing money because the competition does not have to do any of the hard work.
"I don't see how there is space for another player"
Unfortunately, fixed line telecoms are unusual for UK infrastructure in that government permit and encourage duplication. That's led to the situation of a load of altnets building their own networks, largely overbuilding in areas already served by Openreach and/or VM. In some cases there's multiple altnets, plus VM and OR FTTP. The outcome here is that a good number of the altnets are over-stretched and will need to consolidate or fail, and a much worse national outcome that the altnets have collectively invested about £20bn so far and are on course to spend a further £20bn by 2028, almost entirely to duplicate areas that are already well served, or soon will be. Why's that a problem? Two reasons - first, urban over-build does nothing to help the problem of hard-to-serve properties, rural areas, or urban areas at the back of the Openreach FTTP programme. Second, because a large proportion of the altnet investment is going to be an unproductive investment, and when we're talking tens of billions that is money and resources that could have done quite a few different and more useful things.
Nationally we have already spent almost enough to serve every single property in the land with a modern FTTP connection, but the way it's been done will mean that still half the country is on FTTC or worse, and irrespective of the bungled BDUK programme about 3% of properties are still never going to get modern high speed broadband, whilst some areas have more than four competing physical gigabit networks. That's exactly the sort of thing government is supposed to prevent, and it's failed. Worse, the desperate search for growth means government will continue to encourage this sort of poorly guided investment.
Where did this £20 billion by now and another £20 billion by 2028 figure come from please? The only ones building in any volume are Netomnia, Hyperoptic and to a lesser extent now Nexfibre. Most of CityFibre's build is for public contracts, their private stuff is slowed hugely and nearly everyone else has essentially stopped.
Hard to serve properties aren't the market's problem unless regulation makes them that way.
Hardly anyone is going near them as they've no prospect of ever making their money back on them. Spending £300 per premises for a 20% market share in a competitive area makes more sense than spending £3k per premises, having to charge high prices to pay the interest bill on the debt you built up covering the premises at that price and hoping Openreach don't finish a huge build under budget and decide to put you out of business with the remainder.
"Where did this £20 billion by now and another £20 billion by 2028 figure come from please? "
INCA, the trade association for altnets. You'll need to read multiple of their publications to pull this out because (as far as I can see) they've not published a digest of altnet investment to date. But as scaling examples, in 23/24 altnets invested a total of £7.2bn, and in 24/25 £4.7bn. INCA have previously published statements that total alnet investment by 2030 was forecast at £49bn, and their current forward forecast is £25bn, so I'm being a little more modest than they are, and even then that forward look is likely to shrink significantly.
"Hard to serve properties aren't the market's problem unless regulation makes them that way."
And my point was that regulation needed to address that problem, whether through a market or non-market solution. BDUK remains a ponderously slow and incomplete solution.
"Hardly anyone is going near them as they've no prospect of ever making their money back on them."
We agree the prospects for making their money back are in aggregate very poor on pure commercials, but I could say the same about AI and that has no shortage of investors queuing up to throw their money away. Build has slowed up, but there's still a good number of investors who want to put their money directly into a sector like UK telecoms infrastructure, often in the hope that there will be an M&A frenzy that bids up asset prices, even if we think that unlikely.
"Spending £300 per premises for a 20% market share in a competitive area makes more sense than spending £3k per premises, having to charge high prices to pay the interest bill on the debt you built up covering the premises at that price and hoping Openreach don't finish a huge build under budget and decide to put you out of business with the remainder."
I wasn't suggesting that private investors would do rural and high cost premises of their own accord, I was soapboxing that the policy of infrastructure competition was flawed from the start. But, we are where we are. Now its a case of waiting to see how the altnet sector consolidates, who will take big write downs, and whether there's any networks that nobody will buy (eg because the altent that built them goes bust, but there's already extensive overbuild of their network). And all the while, Openreach are building out FTTP at scale and speed, with much lower price per property, so whether it's next month, or four years away, Openreach will put most altnets out of business, aided by the mass marketing by big supplier brands and one touch switching. I'd also question how many altnets are managing £300 per premises, but that's a bit rhetorical since we doubt the commercials of most of the sector.
Openreach will put most altnets out of business, aided by the mass marketing by big supplier brands and one touch switching. I'd also question how many altnets are managing £300 per premises, but that's a bit rhetorical since we doubt the commercials of most of the sector.
Agreed. I've always held the opinion that privatisation has been a huge mistake, especially in telecomms infrastructure. A collosal amount of money has been essentially wasted with overbuild and utilities are pretty much natural monopolies. Personally I think it would have been smarter to have a national & nationalised InfraCo (ie Openreach) that had the responsibility to support government ambitions of FTTP everywhere and then just wholesale access to ISPs or anyone that wants to run services over that fibre.
So basically make it more like the electricity market where there's competition to bill customers, but no alt-nets are offering me competing electricity cables because that would be collosally expensive and silly. As would alt-nets attempting to compete building new water mains & pipes, or gas. There have been arguments that this would be bad because public sector, but there's no good reason why public sector couldn't hire competent people to manage an InfraCo.
I remember hearing a story about the early days of when BT (now Openreach) was being forced to let rival telcos put their own equipment into BT exchanges - allegedly BT built false walls in this particular London exchange to make the equipment room physically smaller ("oh we'd love to let you put your racks in here but as you can see there's no space available") but somehow word of this leaked out.
("oh we'd love to let you put your racks in here but as you can see there's no space available") but somehow word of this leaked out.
That might have been a bit of Chinese Whispers. BT, being The Phone Company provided services to TPTB. So in the good'ol days when they were just BT Exchanges, staff would be vetted before being allowed to go poke around. Then came LLU and pressure to open up exchange buildings to people who might not be entirely trustworthy (eg ISPs). So then to maintain security, collo areas and sensitive stuff had to be physically seperated. There was a bunch of politics around the seperation of openreach and BT's other business units, especially with finite amounts of space being available. So BT Retail was supposed to compete on equal terms with other LLU operators, but I think that basically ended up being pretty much a first come, first served basis with popular exchanges filling up very fast.
The bastards can't/won't provide me with FTTP even though there's a cabinet and fibre 200m away at the bottom of the road.
They probably would, if you were prepared to pay for maybe 200m of construction at £100/m or more. Also why people are complaining about telephone poles reappearing because it can be cheaper to use those and caternary cables than trying to dig. Also gets to be FUN! sometimes when new-build housing developments didn't bother to contact BT's new-build team to get infrastructure built at the same time as the homes. Especially when you can then run into FUN! with wayleaves because the new-build has private roads.
Sometimes it's luck of the draw which BT salesperson you speak to and whether they can go off-script. Plus how your existing service is delivered, ie copper in ducts, overhead delivery by poles etc and then what's in their cabinets or cable chambers. Something that might help ease the pain is if consumers in a neighborhood could get together and committ to service, if it's FTTP. Challenge for retail broadband is there's not a lot of margin to play with, especially if customers want 1yr deals and churn frequently. But there's also a bunch of alt-nets, like one who dropped a leaflet t'other day. Some outfit I've never heard of before called '4thutility' and reminds me to go give that a poke and see who they are & where they're getting their network from. But a £25 a month introductory offer for 'Ultrafast full fibre'. Color me sceptical, but I don't get out of bed for less than 5Tbps these days.
Blimey! What resolution are you watching your porn in?
Watching? I need it for the upload when partying at the weekend.
But it's one of those things that has constantly amazed me over my career. Like back in the day, backbones were STM-16s and dropping out E1s to today, where consumers want (and sometimes get) 1Gbps with business users 10Gbps.. Then designing and managing core networks to deliver that capacity cost effectively, provision and support it. So then advances in photonics gave WDM, then DWDM increasing both channel count and bandwidth. And now some very funky optics that kinda ignore traditional ITU DWDM wavelength & guard channels to squeeze even more capacity out of fibres.. Which is one of those fascinating 'blink and you miss it' given just how many times a second lasers are being pulsed.
"Like back in the day, backbones were STM-16s and dropping out E1s to today, where consumers want (and sometimes get) 1Gbps with business users 10Gbps."
Back in the day when the Internet started in the UK (early 1990s) the backbones and transits in the UK were 64Kbps (or low multiples of that such as 128Kbps or 256Kbps) until Demon finally managed to afford a T1 straight to the USA via Sprintlink and customer connections were (9600bps & 14.4Kbps) dialup modems and 19.2Kbps & 64Kbps leased lines.
Back in the day when the Internet started in the UK (early 1990s) the backbones and transits in the UK were 64Kbps (or low multiples of that such as 128Kbps or 256Kbps) until Demon finally managed to afford a T1 straight to the USA via Sprintlink and customer connections were (9600bps & 14.4Kbps) dialup modems and 19.2Kbps & 64Kbps leased lines.
I cut my teeth at a certain operator beginning with 'B'. Who's internal network was a lot of 4-wire private circuits at 9K6bps. Then lots of V.24 serial connections into IBM FEPs or DEC terminal servers and curious devices like Ferranti PT7 terminal controllers that used to load off a microcassette (eventually), would stay up all day (if you were lucky) and part of the job was copying microcassettes so there was a stash ready to send out as those tapes wore out. Or there were the IBM3174s driving 3270 coax terminals that use to crash a bit less often and loaded faster cos they loaded off 5 1/2" or 8" floppies. Including 'feature' disks called 'RPQ' which stood for 'Request Price Quote' and cost arm+leg.
But then going from 4-wire to digital, and 64Kbps channels via E1 aggregates mostly X.21 connections. Which meant I could get alarms via that interfaces (sometimes). Unless kit then had to go via X.21-V.24 (ok, V.11) convertors that couldn't, and would often just hang. But that migration also involved getting customer service improvements like a call queue display to show how many calls were waiting! Yey! But at least that let me check the queue status when the phone rang, and if there were 30+ calls waiting, I could just switch to an 'all your operators are busy' message while I went to see what had fallen over. This time.
But that was over about 5yrs from late '80s to early '90s and this newfangled thing called 'Ethernet' was just beginning to appear. Which wasn't the Ethernet of today, so I got to drill baby, drill and practice my vampirism skills. Plus much cursing of possibly the worst connector in telecoms history, the slide-lock, which either wouldn't latch, would vibrate loose or partially latch & make a loose/intermittent connection. HSSI probably came a close 2nd for much the same reasons.
But then I tripped merrily into Internet land and a whole new interesting set of challenges. Luckily Ethernet was mostly RJ45 (ok, RJ11) with probably the 3rd worst connector type. Especially with much increased port densities and tempermental crimps. Plus doing things you shouldn't do, like making Ethernet x-over cables. And not marking them as such. Color coding cables is a GoodThing(tm). On the backbone side, ISPs either opted to go the Cisco route, or hang routers off things like Cascade Frame Relay switches. Which then got interesting given most of that kit came from Sonnet-land, not the Old Countrie's SDH, and they were similar but different. A DS3 is not an E3, and neither is an HSSI interface. But the HSSI interface convertor will happily convince a router interface that it's up, even when the line interface isn't. Ah, such fun!
Then on the consumer side.. 9k6 dialup was still pretty much a luxury and modems were often 1200/75, 2400/2400 and then 9k6 and 14k4. Adoption of which was somewhat hampered by, well, Ascend not realising that UK/Euro POTS isn't the same as US POTS. But had the ability to cram a lot more modems into a few U of Max or TNT dial servers.. At least until those started smoking.
That evolution probably took 5-10yrs. Then the inexorable rise of Ethernet, xDSL and broadband much of core networking converging around that, +/- things like G.MPLS and the ability to just bang out connections at 10/100/1G/10G/100G.. With the occasional curve-ball like 40GE that kept standards bodies busy for years, but I don't think I've ever provisioned. Clients generally prefer to just run Nx10Gbps then jump to 100Gbps because their tin costs end up a lot lower. Which has also been interesting watching the kinda death of the router as a 'core' device, at least for connecting client circuits. Mainly due to cost and interface density, plus most of the time the golden rule is 'switch, don't route' because switching is almost always faster and cheaper than routing.
Which is also back to building a wholesale network, or just supporting retail business derived from that. New entrants have the advantage of not needing to support 'legacy' networking, especially when some of those 'legacy' services might be for safety-critical services like alarm & SCADA networks which incumbents might still need to support.
The houses in our area are in two types; (1) council owned & private owned on estates served by overhead cables and poles. (2) "POSH" estate where to remove unsightliness of poles and cables , the telephone wires were sunk into ground under pavements. Now the (1) houses are upgraded and get good internet and speeds; the (2) area are stuck with what can be transmitted over their very old cables, as no one is prepared to pay for the high cost of digging up roads, pavements, re-cabling and back filling , and making good to surfaces. The era of internet has bypassed the "POSH" houses. Most have gone to satellite services.
Fibre won't come from that cabinet and would take a really long time at regular FTTP prices for you to pay back the cost of build. If you're already paying them for copper might only be another couple of quid a month. Even at the average build cost of £250 a home when they do an entire street and it's easy that's a pretty long payback. If they ignored you because no duct costs are a lot higher.
If you're upset they won't make a beeline to you alone check their excess construction charges and maybe get an FTTP on Demand quote.
And a simple question:
Are there any other providers in the area that will provide FTTP?
I am guessing the answer is "No" for any the infrastructure reasons stated. Just blaming OpenReach because they "will not provide a service" is not the entire story.
To put some context a friend of mine living in a posh village near Stamford asked me about FTTP. They had been notified that is was available. A few phone calls and it is installed, the reason why is because it is poles and overhead cables. Another road in the same village where a friend of his lives cannot get FTTP, only FTTC because it is underground. There you are waiting for the fibres to be put in for everyone. It is simply not viable to put it in piecemeal.
Where I live we had FTTC, nothing else. Virgin had done the main areas but would not cross a railway line and we were too far in the other direction to make it viable. As soon as BT/OpenReach announced there were going to upgrade VM suddenly decided it was now profitable to run over a previously uneconomic distance.
That's true, VM's customer service stinks. But if you're in a VM service area, and there's neither an altnet or OR FTTP then your options are more nuanced. I was an original Telewest customer, and they were pretty good, but service and value turned to shit under NTL and Liberty Global. Even so, the poor service and shite customer experience was acceptable for a fast, stable internet connection when Openreach has nothing good to offer. As soon as OR FTTP came along I took my business to a company that valued it (Aquiss, as it happens), and now there's simply no way I'd give my business to VM, directly or indirectly.
I also signed up to Telewast (Blueyonder) over 2 decades ago when the cable came to my street and initially it was great, far, FAR, faster then BT's ADSL and cheaper as well. However, over the years the service got crappier, prices rose, the cable TV side of things sucked more and more, and last year when they offered me a new "fixed price" deal that included an inflation + 3% increase I told them to bugger off.
CityFibre had put in alt-net infrastructure, maybe not deeply cut/reliable like days of old, but cheaper and far faster on uplink than VM, something that mattered to me for video calls and uploading data to a remote site. Now with a small local ISP (Fibrecast UK, in Dundee) and they are a breath of fresh air, no mid-contract hikes, knowledgeable service if needed, and fair price as well.
Cityfibre were MEANT to do Forfar, but thats never happened and it's not even on their site anymore.
Meanwhile Openretch have been and done most of the town, ditto go fibre, ditto virgin media (though virgin media haven't gone live yet))
My street in Kirrie, go fibre told me they were Def doing my property and id be one of the first to get live service due to my early interest, then they discovered that some of the street is ducted as per the openreach plans but other bits are direct in ground (not in plans) and they then decided NOT to do my 70s street (AFTER painting all the markings etc)
Virgin have stopped 3 doors down from me and noone can tell me why exactly, something about a wayleave due to it crossing council depts outwith roads (meant to have been resolved county wide....about 4 years+ ago). - chased it with them repeatedly, chased the council and I still have no further answers.
openreach have went from.by 2026, to not in plan to build not planned, to by Dec 2026 to "we are building here in the next year"
Starting to think I might have to move home to get better Internet.....
Virgin had me as a captive customer since 2004. Even upto 2024, it was VM on 1000mbps or BT copper line on 3mbps. No contest.
Now FFTH is here, I can get 1.6gbps for less than virgins 1.0gbps.
They said I'm out of contract next month, and the price is going up to £88/month. I started the transfer process, and Virgin contacted me and offered me a "deal" of 'only' £122/month, on a 24month contract.
So the Virgin price went UP via their retentions service. No wonder they're losing customers hand over fist.
The captive audience is leaving for faster AND cheaper services.
Known to be appalling to current customers, you "can't get advertised as that's only for new customers" shtick. Lucky to get any sort of discount year on year for staying with them. When their system works it's great which is most of the time but good value after initial contract is really unlikely.
Yeah, but like all their new customer offers that's a circa 50% discount to the full price, during the 18 month contract there will be one, maybe two inflation busting price hikes, and at the end of the term it's a case of haggle with their retentions team or see the bill doubling. The existing customers all know how to deal with retentions, it's largely a few unlucky ones that the agent won't find a good deal for who see big increases. This is evidenced in VM's investor reports that show that VM have failed to increase Average Revenue Per User since about 2016, so even with their annual hikes ARPU is still stuck at £48.50 a month. Factor in inflation, and the real terms value of ARPU has fallen by a third over that time.
"said to be hindering the company's ambitions for growth" Stop, why do all these companies that are already massive want "growth". You could grow but spending money and time on good customer service then more people would stay and more people would sign up. But I know they'll never do that.
"You could grow but spending money and time on good customer service then more people would stay and more people would sign up. "
So we'd like to believe, but unfortunately there's only a small minority of customers will pay more for what they believe is better service. Ask people in a survey and between 30-60% will say that they will pay more for better service - plenty of examples published online. But when it comes to actually parting with money those numbers fall by 90%. So unless companies are happy with a circa 3-5% market share like Waitrose, then it makes zero sense to spend extra to improve customer service.
Even where customer service is crap, that only matters if customers have cause to use it. Talktalk are a byword for poor customer service, yet Ofcom data shows 77% of their customers are satisfied. If almost 8 out of ten of your punters are saying they're satisfied, is it worth spending quite a bit more money to push to just the other side of 80%?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
