» ISP News » 
Sponsored Links

Broadband ISP Virgin Media O2 UK Still Non-committal on IPv6

Monday, Aug 14th, 2023 (7:41 am) - Score 12,552
IP Address Illustration

Seven years have now passed since broadband ISP Virgin Media (VMO2) first informed us that they finally planned to adopt the Internet Protocol v6 (IPv6) addressing standard by mid-2017 (here) and we’re still waiting. Customers of the provider keep asking us for a progress update, but sadly the answer remains unclear.

Internet addresses (IP) help connect you and your devices or software with others around the online world – like an ID number for your connection. Unfortunately, the IPv4 address space could only handle 4 billion addresses and that supply has long since been exhausted (except for the resale of existing address pools), which means that in the future ISPs will need to use the new IPv6 standard in order to continue adding new connections without compromise (e.g. IP address sharing / CGNAT is a compromise that can cause problems).

Sadly, the longer form IPv6 addresses are not directly compatible with older IPv4 addresses and so the two will need to be run side-by-side for many years to come (i.e. until the vast majority of the world has enabled IPv6). The good news is that there are various ways to resolve this challenge, but some providers have been dragging their feet for much longer than others, with one of those being Virgin Media.

Over the years’ we’ve seen the operator test a number of different approaches to deploying IPv6, some of which were better than others (mercifully they seem to have moved on from their Dual Stack Lite phase). But for whatever reason they’ve never actually got around to launching it, despite much of their network and hardware now being in an IPv6 ready state.

So far as we can tell, the engineers seem to want it done, but the bean counters call the shots. More recently, there have also been some indications that the provider might be migrating customers to new DHCP servers, which look setup in a way that could possibly hint toward IPv6 deployment. But for now, the ISP prefers to keep it ambiguous.

A Virgin Media spokesperson said:

“We are continuing with our testing across the network as part of longer-term plans to deploy IPv6 for our broadband customers, and will announce further details in due course.”

In fairness, Virgin Media has a good stock of spare IPv4s left, and on top of that they’re unlikely to face much real consumer pressure on the issue because, to most people, IP addresses are one of those largely invisible aspects of internet magic that you only notice when things go wrong. Some may argue that Virgin’s delay prevents people from connecting to IPv6-only services, but so few of those exists that it’s not (yet) a major issue.

Still, who knows, perhaps they’ll have launched it before we ask again in another two years. Place your bets.

Share with Twitter
Share with Linkedin
Share with Facebook
Share with Reddit
Share with Pinterest
Tags: , ,
By Mark Jackson
Mark is a professional technology writer, IT consultant and computer engineer from Dorset (England), he also founded ISPreview in 1999 and enjoys analysing the latest telecoms and broadband developments. Find me on X (Twitter), Mastodon, Facebook and .
Search ISP News
Search ISP Listings
Search ISP Reviews
35 Responses
  1. Avatar photo Michael Paul says:

    It’s frustrating to see such non-commitment to an inevitable eventuality. Yes we’re still a long way from the retirement of IPv4, but eventually our bag of tricks to extend the life will run out and we’ll need to start NATing IPv6 to IPv4 for legacy services.

    VMO2 aren’t the only ones, it’s silly that we’re seeing alt-nets, especially newer ones, without either day-one IPv6 support, or even in the early years, when they’re clearly not dealing with the same legacy issues you’d expect to hamper incumbents like VMO2. I’m looking at you Jurassic Fibre! Carrier Grade NAT on IPv4 and zero IPv6 support!!!

  2. Avatar photo Andrew G says:

    IPv6 is an archetypal example of how not to do technology change. IETF should have set an end date beyond which v4 addresses would not be used, and worked with registries and government regulators to ensure that was mandated. Instead, IETF too fifteen years or so to ratify v6 as a standard, that was six years ago and adoption’s barely much better.

    Absent being forced, companies will continue to bodge their way forward at the lowest possible cost with CGNAT, and then we stick with what we have – requirements for both v4 and v6 to work, not particular reason for anybody to have to adopt v6, and things continue with the current poor situation.

    On current progress, it seems likely that v4 will be hanging around like a bad smell into the 2050s.

    1. Avatar photo Jon says:

      It’s an absolute mess. I’d prefer IPv4 only if the alternative is CGNAT because I quite like the fact my Virgin IP address hasn’t changed in 4/5+ years. A proper dual stack implementation would be nice though.

    2. Avatar photo Ian says:

      RE Jon: CGNAT is unrelated to IPv6, they only get lumped together because they are both solutions to IPv4 exhaustion, and enabling IPv6 reduces some of the problems of CGNAT.

    3. Avatar photo Andrew says:

      Re: Ian – IPv6 alleviates all the issues of CGNAT to the best of my knowledge. Which doesn’t it alleviate?

    4. Avatar photo Steve says:

      It doesn’t alleviate access to resources that are IPv4 only. Such as twitter/reddit

  3. Avatar photo carlconradw says:

    Every time I read about Virgin Media O2 it suggests a company that’s failing both to innovate and invest. Its mobile network in London is the worst for speed and congestion – try getting a 4G signal around London Victoria. As competitors ramp up their full fibre offerings VMO2 will lose market share hand over fist.

    1. Avatar photo Andrew G says:

      As one of VM’s harsher critics, even I have to say that’s not wholly correct. VMO2 spend almost half a billion quid a year on capital investment, they’ve taken cable from DOCSIS 2 right through to 3.1 and gigabit speeds, they’re working on a network wide XGS-PON conversion that’s only just started but likely to be available to order network wide a good few years before Openreach can get anywhere near the same claim. Whilst O2’s network has looked under-invested for a while, they’re certainly spending money on that too. Judging on localised experience may be true of your experience, but isn’t necessarily indicative of the wider network position – other networks have similar problem locations.

      Despite all that, I’m no longer with VMO2, and that’s down to the grim customer service and unsavoury commercial practices.

  4. Avatar photo kbo19kr@googlemail.com says:

    in a word, apathy

  5. Avatar photo Big Dave says:

    Trouble is the market has become so cut throat there is no room for anything the bean counters see as fripperies. As long as they do the bare minimum to make it work and keep the money coming in thats good enough. When I think back to when I first got adsl in 2002 that was plus £30 a month plus a £15 phone bill on top. Run £45 through the Bank Of England inflation calculator and that equates to £79.00 today. I don’t pay that for BT fibre 900 + digital voice with 700 minute call allowance plus a load of add ons to boot. Mind you I’ve have had IP V6 for the last 7 years. The whole telecom industry has got involved in a race to the bottom. No wonder it’s in trouble

    1. Avatar photo spurple says:

      Pretty sure they have a lot more customers nowadays too, so that kind of negates your argument tht price hasn’t risen in line with inflation. It’s not supposed to. Economies of scale, technology has improved to lower unit costs etc.

    2. Avatar photo Iain says:

      Technology in general is deflationary. Virgin Media don’t pay as much for their Windows XP machines as they used to. Or, at the very least, they get faster, more energy efficient devices, which boost productivity.

      The price they pay for high capacity network switches, and network bandwidth, will similarly have fallen

  6. Avatar photo Karen says:

    What we need are some large websites state they will drop IPv4 by a given date. This is the only way it will become a priority for some ISP’s

    1. Avatar photo Steven says:

      Typical chicken and egg scenario. Big content providers don’t want to drop v4 because ISPs don’t have a large enough v6 user base. ISPs are dragging their feet because v4 “works” just now.

      If Netflix and Amazon announced they’d be switching off v4 at the end of the year… fairly sure the bean counters would be falling over each other to give their v6 deployment the green light.

    2. Avatar photo Berny Stapleton says:

      Steven – AWS have said that they’re going to start charging for IPv4 addresses as they can’t get enough to satisfy demand. That’s the start of the prod from the bean counters…

  7. Avatar photo OverThere says:

    I’m one of those who got shunted to a new DHCP scope last week, still no sign of v6 though and I can’t see it being any time soon, Virgin have no reason to other than a small handful of customers asking for it.

  8. Avatar photo steve welch says:

    I imagine this is a Worldwide tactical issue. I first installed a dual stack ipv4/ipv6 system on a Sun SparcStation in 1999, so it’s not like ipv6 is new. I was then using it in a real-time UNIX development system, and providers such as Andrews & Arnold have had it available for years.

    1. Avatar photo Ivor says:

      BT and Sky have also had it available for years too, and at a far larger scale than A&A.

      No excuse for existing ISPs to not Provide a full Internet Service, and it is even more ludicrous for new ISPs and altnets to not be v6 enabled from day one.

  9. Avatar photo M says:

    It’s like waiting for Three to enable Visual Voicemail. You need it but they don’t think you do.

    1. Avatar photo 4chAnon says:

      No problem, Vodafone Hutchinson 3 UK Ltd will do it (probably)

  10. Avatar photo wireless pacman says:

    I strongly suspect I’ll be pushing up the roses before IPv6 takes off properly.

    I’ve still not see any proper (imho) technical case for it, let alone a business case.

    It was badly “designed” by uber-geeks with no real thought given to normal (non techy) peeps. Said geeks then “decided” IPv4 could cope pretty well for a few decades more so (as mentioned by a previous poster) made zero attempt to kill off IPv4 whilst it was (then) still (relatively) small.

    Normal people like short snappy things that they have a chance of remembering (&/or getting their heads around). Even the non-techies can generally cope with an IPv4 address, but have no concept of an IPv6 one and would probably freak at the sight of one.

    Normal people also don’t tend to like change – especially where there is zero reason (from their perspective) for such change. They are VERY unlikely to be impressed with tech geeks “explaining” that with IPv6 every proton in their house can have its own IP address.

    Oh, and of course there is also (again in my humble opinion) the massive design flaw that IPv6 is not intended to use nat on the home router. At a stroke they removed a very effective (and yes, I know, basic) level of security to stop intrusion – or at least make it more difficult.

    1. Avatar photo XGS says:

      If you don’t mind sharing what’s your technical background and exposure to IPv6?

      It’s difficult to put this without sounding rude but if you aren’t aware of any technical reason to use IPv6 over v4 you might be a little misinformed about v6. If you think it’s just a 2^128 address space with the rest the same you couldn’t be more wrong.

      The security comment invoking NAT is invalid. I’ve heard it used before: it’s simply wrong.

      I’m interested, too, in how often you think non-techies use IP addresses over DNS and why? Only time I can imagine is when being told to change DNS servers.

      Not like random folks going to be configuring their equipment with fixed IP addresses.

      IPv6 has SLAAC, DHCPv6, ND, EUI64 so no need to even number an interface, use that, it’ll take an address from an interface pool, in turn derived from the larger prefix the ISP delegated you, and it’ll advertise itself as the gateway for that interface. v4 can’t do that. Link local for individual nodes to argue amongst themselves, addresses with a v4 address embedded, etc, etc.

      Just scratching the surface. IPv6 + QUIC + TLS 1.3 == HTTP/3 at something like potential. An immediate session restart, if you actually lose your session at all, as there’s no NAT state to get broken

      I was skeptical. Tom, I know you’re reading this, Marek and others converted me. Well, I say ‘convert’ probably I had a few too many Earl Grey IPAs, was a bit suggestable and it made perfect sense. The issue was my ignorance. I was educated, enlightened, read up and said ‘This is so much more powerful, easy to control and index can self repair to a greater extent than v4. Everyone network this’.

      Now addressed in our main product. Dual stack baby and I’m getting my own PI space for my lunatic side venture

      Read up on the capabilites of v6. It’s the only way to go, Wireless.

    2. Avatar photo Ivor says:

      the technical case is that we’ve run out of addresses and we need to expand the address space. Might as well go through that pain once.

      Normal people don’t even know what an IPv4 address is. They just want the internet (often called “wifi”) to work. Good ISPs have seamlessly implemented IPv6 and customers are using it without knowing.

      NAT is not a security feature and should not be relied upon to provide it. NAT is not necessary in IPv6. If you know enough about IPv6, you should also know how to set up sensible firewall rules (or if you don’t, continue using the ISP supplied router which does)

    3. Avatar photo wireless pacman says:


      Am sure the techie bits in your response are valid, but I would question how relevant they are for the (vast) majority of peeps.

      To answer your question, I own/run a small rural isp business. It is all run on v4. Over it’s 20 year existence I’ve only ever once been asked to provide a v6 address (and I did), but they only asked as they wanted to “play around with it” and not for any real business case.

      Totally accept your comment that my knowledge of all things v6 is not what it could be. However,as a (very) small business owner I have to prioritise my time investment based on need.

      My main gripes with v6 remain. It was designed by geeks for geeks and no realistic consideration was given to normal peeps. Then, and im my view this is more important, it was not forced on the industry once it became a ‘real’ standard – which was long before the Internet as we know it took off for real.

      The only reason we ourselves have a v6 block is that RIPE basically insist on it, not because we have any real need for it. I said as much to a RIPE chap I was talking to a few months back. His response (paraphrasing) was pretty much “yeah, get where you’re coming from”. He then told me about a couple of very large Nordic ISPs that gave in to RIPE pressure to invest in and roll out v6. After having done so, they got back to RIPE with (paraphrasing again) “None of our customers seem to give an es aych one tea about it and are more than happy sticking with v4”.

  11. Avatar photo Chris says:

    If NAT was supported in ipv6 as it is in ipv4 then ipv6 would be showing greater adoption, especially in enterprise.

    It’s far better to have a gateway breaking that end to end comms than to just have everything in your lan be directly reachable from the internet.

    There is no need for my dishwasher to be reachable from the internet, there is however some use to me for it to be able to reach the internet.

    IPv6 went too far, all we needed was a bigger address space, not deliberate policies to stop nat.

    Ipv4 had an addresss extension header, the address extension header could have been used to extend ipv4 or as a migration to ipv6 – ipv6 address could have been put in said header, for some reason it was deprecated.

    As for vm not supporting ipv6, what impact is that actually having? Is anyone not able to do something because of lack of ipv6 support?

    1. Avatar photo XGS says:

      Hey Chris,

      ‘If NAT was supported in ipv6 as it is in ipv4 then ipv6 would be showing greater adoption, especially in enterprise.’

      The largest publicly traded company in the world by market cap is dual-stack internally.

      ‘It’s far better to have a gateway breaking that end to end comms than to just have everything in your lan be directly reachable from the internet.’

      The largest user of IPv4 in the world is mostly publicly addressed internally. https://en.wikipedia.org/wiki/List_of_assigned_/8_IPv4_address_blocks

      They appear to have plenty of gateways breaking end to end comms.

      ‘There is no need for my dishwasher to be reachable from the internet, there is however some use to me for it to be able to reach the internet.’

      Agree. Which is why unless there’s been a major configuration error on a router you’re using it won’t be. Stateful firewalling does not require NAT. You seem to be operating on the preconception it does.

      I permit ICMPv6 inbound, it’s part of the standard. Try initiating any other connection it’ll be silently dropped unless there’s an outbound connection matching it for IP protocol and, where used, port.

      ‘IPv6 went too far, all we needed was a bigger address space, not deliberate policies to stop nat.’

      An argument could be made there was too much of a purist demand to be rid of NAT, however…

      ‘Ipv4 had an addresss extension header, the address extension header could have been used to extend ipv4 or as a migration to ipv6 – ipv6 address could have been put in said header, for some reason it was deprecated.’

      IPv6 addressed a number of shortcomings in IPv4.

      ‘As for vm not supporting ipv6, what impact is that actually having? Is anyone not able to do something because of lack of ipv6 support?’

      There are a number of reasons to go v6. BT and Sky didn’t do it for fun or because they were bored.

      Maybe go back nearly a decade for why VM were already mooting it then. https://tv.theiet.org/EmbedPlayer.aspx?videoid=5899

    2. Avatar photo Chris says:


      You’ve written words. It never addressed my points.

      I’ve worked at ibm on the 9/8 net if that’s what your referring to, amongst others.

      Just what are those shortcomings on ipv4 that are addressed in IPv6, the ones people mention are address exhaustion and Nat, what else has IPv6 fixed?

      I don’t see why I should allow unsolicited inbound anything.
      With IPv4 and Nat it’s effectively guaranteed that unsolicited inbound is dropped by default. With IPv6 I have to explicitly drop everything.

      What’s the default IPv6 fw policy in a consumer gateway? I’d assume per IPv6 standard it’d allow any any in unless changed.

      Sky & bt provided IPv6 due to industry and consumer pressure which evaporated once delivered.

      At the end of the day it’s an addressing scheme whose instigators wanted to punish and denigrate those reliant on nat.

    3. Avatar photo XGS says:

      There’s nothing in the IPv6 RFCs / standards requiring ‘allow all’ as a firewall policy any more than there is in IPv4.

      At the end of a firewall policy is an implicit deny. I’m not aware of any equipment that breaches this but if it’s there in something it’s doing it wrong.

      The default on the consumer kit I’ve seen is to behave statefully, not to allow random inbound connections.

      IPv6 fixes some security issues, adds additional functionality and incorporates various things that had to be built on top of IPv4 into the network layer. It’s a superior, if more complex, protocol. The lack of NAT is not a significant security issue any more than it is for the US DoD or other organisations using publicly routable address space internally.

      Your opposition to it seems… vehement. I know of a very few people that are anti-NAT zealots: I’m not one of them.

      Out of interest what’s your solution for somewhere with many, many devices from things like IoT through to laptops and other computers?

      NAT overload only has so many ports available to it.

    4. Avatar photo XGS says:

      I forgot another part, too. Using NAT requires use of ALGs / Application Layer Gateways to allow some things to work.

      My wife and I both have need of IPSEC VPNs for different purposes. I’m fortunate in that I have two Internet connections, but have to route hers out of one and mine out of the other as hers cannot be port translated: it must use source and destinations 4500 for ESP and 500 for ISAKMP. If mine gets there first she just can’t connect.

      IPv6 that goes away. You don’t actually need NAT traversal as there’s no NAT to traverse. Another issue that goes away with v6.

    5. Avatar photo Chris says:

      You mentioned that icmpv6 needs allowing in.

      Some discussion here about it.

      Again you mention things with no detail.
      “ IPv6 fixes some security issues, adds additional functionality and incorporates various things that had to be built on top of IPv4 into the network layer. It’s a superior, if more complex, protocol.”

      What security issues are fixed?
      What additional functionality is added?
      What has been incorporated?
      What is the issue building things on top of ipv4, it’s an attribute that’s arguably made it outlast its design life.
      Why is ipv6 superior?

      Re large organisations that could exhaust nat pools, typically they don’t allow straight access to the internet and everything goes through a series of load balancers and proxies.

      In enterprise that’s how you do it, in a home you’re unlikely to exceed a nat table. Nat allows for 60k+ connections per destination ip & port combo

      It’s a problem solved decades ago..

      Re isakmp vpn’s, if you know the public IP’s at the remote end you can port forward to individual internal hosts at your end.

      It’s unusual in this day to be restricted via a corporate vpn as you described, more likely to see that with site to site vpn’s. Could be your upnp on your gateway interfering.

      Re nat & connection state, nat capable routers made it trivial to then add stateful fire-walling as the cost of the capable hardware was already made & was a little additional coding. This made broadband more secure than dial up for home users without the need for them to be technically minded.

      IPv6 can up the technical burden to users for no good reason.

    6. Avatar photo XGS says:

      You can read the RFC on IPv6, Chris, as well as I can.

      On the port forwarding I don’t have UPNP active and have no IPSEC ALG running. The remote end is dynamic based on DNS round robin for both.

      Given the amount of flows a simple home generates a moderately large office making extensive use of IoT isn’t going to have a good time. It’s actually become a use case for Zscaler, Axis and other Secure Web Gateways.

      Bit pressed for time, please excuse me!

    7. Avatar photo Chris says:

      I don’t think your getting it,

      Regardless of iot or laptops or other things that need to go out to the internet, in an enterprise or corporate we have mechanisms to deal with multiple systems sharing a single or pool of public addresses. It’s not been an issue for decades as mechanisms exist to deal with it.

      Simply put, even in an organisation like a university it’s unlikely to have a need for 60k+ concurrent sessions to a single ip on a single port (like 443 for https), if that’s a likely hood then a load balancer in front of a pool of proxies fixes that, or a proxy with multiple public IP’s etc.

      If your home has more than 69k+ concurrent sessions to the same remote ip & port then you likely need something more than the isp provided gateway.

      It’s not hard to have hundreds of devices on a home lan and that won’t cause an issue with nat, unless your routers nat table can’t cope with the number of sessions.

      Re your vpn issue, it’s unusual as your nat table will block unsolicited in bound comms by virtue of not having an entry for it, upnp sorts that by building an entry to permit bi directional initiated comms. If only the 1st vpn works then that suggests upnp is doing something, unless your both connecting to the same vpn providers same address.

  12. Avatar photo Ad47uk says:

    Plusnet did a IPv6 trial a few years ago, but that is as far as it got as far as I know. I do wonder why ISPs have not gone for Ipv6, maybe it is because there are very few sites that use it, also does it cost extra to run then both side by side? I don’t know if it does or not, I have not really looked into it.
    Zzoomm don’t use IPv6, I thought they would have. I presume all modern routers support IPv6

    Makes no difference to me, zzoomm not supporting IPv6

  13. Avatar photo XGS says:

    This comments section has been quite informative.

    A really big point being how many folks seem to conflate stateful firewalling with NAT.

    NAT relies on connection state, not the other way around. NAT is having a connection state engine and in addition rewriting source/destination IP/port whether using source or destination NAT. The connection state engine marks the connection for NAT or some other action but is the first step in the commonly used masquerade NAT we see on home routers.

    The same connection state engine will drop unsolicited traffic whether NAT is being used or not. No outbound connection no inbound one, in no way dependent on NAT.

  14. Avatar photo Bob says:

    I’ve been using v6 for about 13 years now, on the public internet. It has been a game changer for an IT enthuiast, unlimited public addresses, inter-server backups between family sites has been seamless. I’m amazed that people are still stuck on v4 after a decade or more.

Comments are closed

Cheap BIG ISPs for 100Mbps+
Community Fibre UK ISP Logo
Gift: None
Virgin Media UK ISP Logo
Virgin Media £26.00
Gift: None
Shell Energy UK ISP Logo
Shell Energy £26.99
Gift: None
Plusnet UK ISP Logo
Plusnet £27.99
Gift: None
Zen Internet UK ISP Logo
Zen Internet £28.00 - 35.00
Gift: None
Large Availability | View All
Cheapest ISPs for 100Mbps+
Gigaclear UK ISP Logo
Gigaclear £17.00
Gift: None
YouFibre UK ISP Logo
YouFibre £19.99
Gift: None
Community Fibre UK ISP Logo
Gift: None
BeFibre UK ISP Logo
BeFibre £21.00
Gift: £25 Love2Shop Card
Hey! Broadband UK ISP Logo
Gift: None
Large Availability | View All
The Top 15 Category Tags
  1. FTTP (5514)
  2. BT (3514)
  3. Politics (2536)
  4. Openreach (2297)
  5. Business (2262)
  6. Building Digital UK (2244)
  7. FTTC (2043)
  8. Mobile Broadband (1972)
  9. Statistics (1788)
  10. 4G (1663)
  11. Virgin Media (1619)
  12. Ofcom Regulation (1460)
  13. Fibre Optic (1395)
  14. Wireless Internet (1389)
  15. FTTH (1381)

Helpful ISP Guides and Tips


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