back to article OK, this time it's for real: The last available IPv4 address block has gone

You may have heard this one before, but we have now really run out of public IPv4 address blocks. The Internet Assigned Numbers Authority – the global overseers of network addresses – said it had run out of new addresses to dish out to regional internet registries (RIRs) in 2011. One of those RIRs, the Asia-Pacific Network …


      1. Long John Brass Silver badge

        Re: IPv6 in the DMZ

        those private blocks are already to small for them = 16,777,214 hosts! I mean yeah sure you generally carve it up into sub-nets but still!

      2. Nutria

        Re: IPv6 in the DMZ

        The 10/8 block has 16 million addresses. Sure, there are companies which need more... but not many.

        1. simpfeld

          Re: IPv6 in the DMZ

          One of the big drivers has been US mobile companies. They run out of private 10.x space quickly, as 19 odd million isn't large in mobile subscriber numbers.

  1. Dodgy Geezer Silver badge

    Who got the last one?

  2. Joe Montana

    Incentives / Demand

    Currently there is very little reason for users to demand ipv6, about the only vendor doing anything positive is microsoft who publish documentation for the xbox one which encourages you to use ipv6 for a better experience. Users are not asking for ipv6, so providers don't bother offering it.

    If users were demanding ipv6, isps would start providing it or lose customers, and sites would start offering dual stack at least.

    A lot of US government sites are available over ipv6, because the government demanded it... In the UK, there are no government sites available over ipv6 that i'm aware of, even the relatively new site is ipv4-only.

    Even when everything supports ipv6, many people will not bother to configure it or even explicitly disable it.

    One approach would be for the likes of google and facebook (who both already fully support ipv6) to start offering new (ie beta) features over ipv6 first, and displaying warnings to users accessing services without ipv6. Having beta services available over ipv6 would result in better beta testers in the short term (people with ipv6 now are more likely to be tech savvy), and result in more users demanding ipv6 from their isp.

    1. Anonymous Coward
      Anonymous Coward

      Re: Incentives / Demand

      We offer v6. We get 10 million hits/day. Of those about 100 come from v6 addresses. Of those 100 all but 3 come from Googlebot.

      So we support v6 for the benefit of 0.00003% of our customers. You can see why management wouldn't want to invest.

    2. Anonymous Coward
      Anonymous Coward

      Re: Incentives / Demand

      > "google and facebook to start offering new features over ipv6 first."

      Trouble with that is that most new 'features' from either of these two are simply better ways to monetise their userbase. Not really an incentive now is it?

  3. Lee D Silver badge

    I'm implementing my rule again, Reg.

    When YOU BOTHER to put an IPv6 address on your website, as already supported by your browsers, DNS host, webserver, content delivery network, and everything in between... THEN you can be sarcastic about a poor IPv6 deployment statistic.

    It's companies like you that are precisely the problem. "We've got our IPv4, and it would 'take effort' to make everything work for IPv6, so why bother?" is the attitude you've given for... what... 8 years? Maybe more. I'll check my comment history where I have about half-a-dozen annual "Yeah, we're going to look into that next year" things.

    I mean, at least you did eventually get around to SSL. But, honestly, you should restrain yourself from sarcastic IPv6 comments until you at least have an AAAA record on a beta-domain:

    Your DNS hosting provider is "Cloudflare"

    For anyone else, it would just be annoying but for a TECH SITE it's downright rude. It's like writing articles dissing Windows 10 for not keeping up to date while running XP in all your offices.

    1. I ain't Spartacus Gold badge

      No. Because the article is taking the piss out of IPv6, and its poor adoption, because almost nobody wants it. Including El Reg - who don't bother with it.

      So their article is fine.

      IPv6 has failed. There's been pressure to fully adopt it for years - even decades now. But people just keep on with IPv4, because kludging that a bit more is easier than switching over.

      What this should tell everybody is that we should abandon IPv6 and start from scratch with something that people will actually be willing to use. Otherwise I foresee decades more of the current mess.

      So let's say IPv6 is Vista. We need an IPv7. Which almost everyone will like The trick then is to beat to death with sticks anyone who suggests IPv8...

      1. CrazyOldCatMan Silver badge

        The trick then is to beat to death with sticks anyone who suggests IPv8...

        IP Mountain Lion? IP High Sierra? Will those do?


      2. Lee D Silver badge

        IPv6 is present in all modern smartphones - it's a requirement of the protocols involved.

        IPv6 is present in all modern communication protocols - including DOCSIS.

        IPv6 is present in all modern operating systems. It took decades to get it in there.

        IPv6 is present in all modern switching/routing hardware. It took decades to get it in there.

        Nobody is going to supplant IPv6.

        You know what hindered it? That NONSENSE about it meaning that every device had to have a globally addressable address. That was the problem. Nobody wants their local devices to have an address like that. NAT is perfectly fine. And converting a NAT network to IPv6 consists of this... add IPv6 to the gateway device. Done. Everything else can be done at leisure, or stay IPv4 into perpetuity - nobody would ever care.

        That nonsense literally held back adoption, because who the hell wants to go through every switch, router, server, client, phone, printer, etc. and give them all IPv6 addresses and then address them only by that? Nobody. Internal networks, it does not matter how they operate. That's why they're internal.

        But the anti-NAT brigade set us back 10 years on IPv6 because of that.

        You are not going to get anything but IPv6 for the next 20 years. Deal with it. Activating it, using it testing it, and understanding it takes about an hour tops for any IT professional, with a deployment plan then going into normal change management.

        Sorry, but you can make all the excuses you like, like The Reg does. All my servers, domains, etc. are IPv6 capable and have been for years. It really doesn't take much and things like log-file analysers and custom-made sticking-plaster scripts are the things that need time to be converted. The protocol support? It's just there. In your device, in your OS, in all the things you use that OS on.

        And deploying it affects nothing IPv4-wise, so there's no reason not to. Do it using and say it's a test. Google report that something approaching 10% of their traffic is IPv6 now. It's not going anywhere.

        1. Lee D Silver badge

          I lied:


        2. Orv Silver badge

          That nonsense literally held back adoption, because who the hell wants to go through every switch, router, server, client, phone, printer, etc. and give them all IPv6 addresses and then address them only by that?

          Eh? I don't understand what the complaint is here.

          Yeah, you plug in the printer and your router will automatically give it a global IPv6 address. Your router will also have a firewall in it, probably on by default, so there's no real security concern here.

          If you don't like that global address, that's fine, because IPv6 allows multiple addresses per interface, and the fe80:: block is set aside for local use; you can set up all the private, non-routable IPv6 addresses you want there, and in the fd00:: block. Actually, given modern discovery protocols like Bonjour, your printer will probably do this automatically and broadcast its existence to your computer, so you never actually have to enter the address manually.

          The only real difference here is you don't have to go through a broken NAT layer that has to keep track of every single connection, and guess when they're idle in order to clear out that memory table. The limitations of this become readily apparent when you launch BitTorrent on a NAT'd machine and everyone else's SSH connections drop because the NAT table filed up.

          1. Lee D Silver badge


            You've gone from "Just add an IPv6 address to the device already running NAT on the front-end of your Internet connection" which is centralised, easy to diagnose and easy to revert to "set up IPv6 local DHCP which could interfere with local services if they aren't already set up for IPv6, while making sure that all your internal access lists, subnets, etc. are also configured for IPv6, etc. etc. etc." not to mention "now you have to consider that every machine has a globally routable IP", so your firewall config just expanded from securing ONE IP to an entire subnet on a protocol you aren't familiar with.

            Worrying about NAT literally held everyone back. NAT isn't broken. It works for the vast majority of the world. You know how we know? Because the vast majority of the world has a NAT router on their DSL connection. And the solution to "poor" IPv6 deployment is now likely to be carrier-grade NAT on IPv4. Ironically, the "problem" cited by everyone like yourself - spewing NAT-hate - actually CAUSES PEOPLE to stay on IPv4, which means ISPs are forced to NAT them as they can't get any more public routable IPv4 addresses.

            Nobody is saying "stay like that forever", but the initial transition is literally an hour of work, for a site with an unlimited number of existing machines, with no changes to internal services whatsoever. But NAT-fear stopped people doing that, because "with IPv6 you should ditch NAT too", etc. etc. Which turns it into a 6-12 month project of testing and reconfiguration.

            Your post is the epitome of demonstrating my explanation. NAT or not-NAT has nothing to do with security either. I'm not even claiming that. NAT is a "sensible default" applied to the technology that happens to translate to a "block all incoming" as the final rule by the way it works, and that should be your default rule anyway.

            What you did was tell people: You're an idiot to use NAT, turn it off. When everyone is using NAT and there are no inherent problems with a proven technology that serves a practical purpose. And because you conflated that with "here, have a bunch of new-style IP addresses", nobody moved to new-style IP addresses because they were afraid they'd also have to change EVERYTHING about a technology they've been using successfully for decades.

            P.S. Your IPv6 router/firewall, no matter how basic it is, still has to keep track of connections. Stateful firewall is the norm. If it's not, you should worry. And though connection tracking on IPv6 does technically take up slightly more memory... there's no way you should be hitting limits on any router advertising itself as IPv6-capable.

            P.P.S. I've run Bittorrent on NAT'd connections, like I imagine the majority of the world has. It's never dropped unrelated connections. That's a factor of "crappy router" not NAT. I've literally never witnessed the symptoms you describe (but sheer bandwidth can fill up your outgoing line, which knocks your users for six if you have asymmetric connections and they can't get TCP request and acknowledgements etc. back out. Solution: QoS, not removing NAT.)

            1. Nanashi

              You kinda are an idiot to use NAT when it's not necessary. If you use it when you don't need it, the only thing it does for you is make your network harder to admin, and make your security harder to reason about. It makes sense if you're a masochist, I guess.

              (By NAT I specifically mean iptables' "-j MASQUERADE" mode; the one that you apply to outbound connections only. There are various other targeted cases of address translation that can be handy, like NAT64/NAT46 or load balancers, but we're talking about the type of NAT that people use on their home connections here, right?)

              Of course it is often necessary -- you need it if you aren't receiving enough IP addresses for your network from your upstream ISP. That is why you see it used everywhere for v4. It's because we're so short on IP addresses that you're lucky if you can even get one single v4 IP for your entire network.

              As a side note, you're going to need to deploy v6 on your local network and not just on the WAN side of your router, because there's no way to fit v6 addresses into v4 packet headers. Your LAN machines will have no way of reaching v6 hosts without v6 on the LAN. This is just an unavoidable consequence of the way v4 works, and the only way to fix it is to deploy a new protocol. (Or you could use a proxy, but nobody wants to use proxies.)

              > NAT is a "sensible default" applied to the technology that happens to translate to a "block all incoming" as the final rule by the way it works

              Woah, woah, woah... where did you get this idea from? NAT doesn't block any connections. Literally the only thing this type of NAT does is change the apparent source address of outgoing connections. It doesn't do anything to inbound connections.

              Okay, I know the answer to this one: it's because you normally use NAT together with RFC1918, and using RFC1918 does make it difficult for most, if not all, of the internet to connect to you. But the NAT part of that does nothing to inbound connections. This is the "makes your security harder to reason about" that I mentioned above: it's causing a misunderstanding here that could potentially be dangerous if you try to rely on it.

            2. Orv Silver badge

              ...your firewall config just expanded from securing ONE IP to an entire subnet on a protocol you aren't familiar with.

              That doesn't actually make your firewall config more complex if the default is "block all incoming," which is what you're arguing we should use NAT to do anyway. (This is assuming a bridge-style "transparent" firewall, but those are common even on IPv4 networks at this point.)

              ...everyone is using NAT and there are no inherent problems with a proven technology that serves a practical purpose.

              If you think NAT isn't broken, it's because you're used to the brokenness. I started using it when it was called "IP masquerading" and was an experimental Linux kernel module. It's always been hacky and buggy, people now just think the breakage is normal.

              Besides the problem I noted earlier, there are others:

              - Double NAT. Right now most traffic only has to traverse one layer of NAT, at the home router. It does usually work OK if you only go through one layer of it. Try to go through two -- say, you're using a mobile hotspot (NAT'd on the phone) on a mobile carrier that's also using NAT -- and things start to break. FTP simply stops working, for example, even in passive mode. As time goes on we'll be seeing more and more levels of NAT applied and more and more protocols will fall apart.

              - Peer-to-peer protocols that work across the Internet but fail if you try to use them with someone on your LAN, because the IP addresses don't match up. I've played some online games that I could play with literally everyone in the world except my own family.

              - Idle TCP/IP connection timeouts. After 5-10 minutes of silence, a lot of home routers will decide a TCP/IP connection is no longer needed and drop it to free up space in the NAT table. This is why SSH sessions over home routers tend to drop if you step away for a few minutes. This has resulted in a lot of hacky keepalive systems that send useless data every minute or so just to keep the channel open.

              - The security disaster that is UPnP, which exists mostly to give NAT'd devices an automated way to request port forwarding.

              Anyway, the fact is that IPv6 *does* support a form of NAT. It's called Network Prefix Translation. But it's not commonly implemented because it doesn't actually serve a useful purpose.

          2. P. Lee Silver badge

            Mostly wherever you do nat, you will still have a session table in IPv6 because it's a firewall.

            On home lans, the issue is rubbishy kit breaking connectivity. On enterprise lans, bonjour is not a good thing, and troubleshooting without dns is certainly a thing. That makes things harder with IPv6.

            We probably need routing protocols to include not just addresses, but names, which provides local resolution with the same trust level as numerical addresses. That's not dns, that's just route mnemonics.

            Then IPv6 would be a lot less scary.

            We could also start demanding voip over IPv6 only, so we can blacklist those scammers.

            1. Roland6 Silver badge

              >On enterprise lans, bonjour is not a good thing

              Like to know more.

              Certainly, single LAN segment services like Bonjour are a pain in the arse in (larger) enterprise office deployments, but in the main that is down to a disconnect between logical and physical network structure and people's perceptions: why can I not see the office printer that is just across the corridor from me?

              Hence to me services such as Bonjour are indicative of omissions in our network protocol suites that really need to be addressed.

  4. TrumpSlurp the Troll Silver badge

    Various inane ramblings.

    Back in the day a certain large Telco had several Class A allocations because at the time few companies wanted one or even knew what one was. Have they been returned, or are they an appreciating asset?

    Users demand V6......come on, 95% of users at least have no idea what IP stands for, what DNS is, MAC address, anything that underpins the easy names used to identify and connect to network resources. If all new installs from ISPs were IPv6 only, the end user wouldn't even notice as long as all the devices in the home were V6 capable. Even then the ISP supplied router could act as a gateway for legacy kit.

    I suspect that one of the main anchors is ISPs continuing to roll out obsolete kit {cough}SH3{cough} and not wanting to replace the enormous real estate which is end user devices. The current service is "good enough". In the States, allegedly, there is a vast estate of cable routers which are old, slow and tired {know how they feel} which are also a massive cash cow. Why kill the cow that gives the golden milk?

    Until major services start to fail in an obvious manner and there just aren't any more bodges (NAT, carrier grade NAT) to keep the current edifice stumbling along there is no incentive to go V6.

    In fact, didn't NAT completely take the wind out of IPv6's sails by making it an interesting but non-urgent idea?

    1. I ain't Spartacus Gold badge

      Re: Various inane ramblings.

      Users demand V6......come on, 95% of users at least have no idea what IP stands for, what DNS is, MAC address, anything

      Rubbish! V6 is a fruit juice. IP is what happens after I've drunk too much of it. And a MAC address is where the Deliveroo drive brings my burger to.

  5. Anonymous Coward
    Anonymous Coward

    It's bad to say....

    ...but I still haven't got my head around IPv4 yet. With subnet masks etc. I know I should know it by now off the top of my head but I don't. I don't use them enough so always look them up.

    But I like going round the house and knowing router is Main PC is 2nd PC is and so on. Can't very well do that with IPv6 can I.

    AC due to being embarrassed.

    1. A Non e-mouse Silver badge

      Re: It's bad to say....

      I don't use them enough so always look them up.

      Don't be embarrassed. You have to be regularly dealing in subnet masks to easily switch between /28 and I've seen many network engineers with little crib sheets on the side of their monitors to cross reference between the two formats.

      1. Alistair Silver badge

        Re: It's bad to say....


        or ... linux -> ipcalc

    2. Lee D Silver badge

      Re: It's bad to say....

      I always use a subnet calculator ( Anything else is just going to introduce errors, because a lot of things WILL still work with an incorrect subnet.

      For instance, the range I inherited was 48.0/22 ( - that's a really odd range.

      They were using the 48's for client DHCP initially (again - NO IDEA WHY, it's within a local range!). Then they needed more addresses, so some fool decided to do the above (which gives you the 48.49.50 and 51's). But they didn't update the subnet everywhere. So what you get are a lot of computers that can get an IP, log on, talk to the gateway, connect to the Internet, etc.

      But when you try to talk to printers or, say, anything broadcast - DLNA, Chromecasts, Airplay, etc. then it doesn't work properly.

      And you get things like... the 48's are filtered for web, but the rest aren't. All kinds of issues. And I guarantee you that the CCTV, access control, etc. guys will just read it as their bog-standard, "we-don't-know-why-just-type-it" no matter how much you highlight the fact because they don't understand what a subnet is (or a VLAN or VPN or STP or anything, for that matter).

      The solution, of course, is to stop faffing about and use well-known subnets. Very few places have IT big enough to worry about broadcast floods, etc. and hence want to limit their subnets down to the bare minimum necessary, but no IT department that understands the issue... just use the whole damn range and a bog-standard subnet and be done with it.

      Then you have the numbering issues? Then don't. Nobody needs to care about IP addresses any more. I wouldn't be able to tell you the IP address of any of the 1000+ devices on my networks except for a) the gateway, b) the primary and secondary DNS, c) the main DC (which is actually the primary DNS anyway, but I don't actually NEED to know that, I could just use it's name!).

      At home is the same. My router gives everything a name. Sure, at one point you have the IP there but it's DHCP and then you "reserve" the lease and it's permanently on that address but... more importantly... you then just give it a name. Anything that doesn't have a name will autodiscover, I assure you (e.g. Chromecasts by using the broadcast address).

      And it's a damn sight easier for grandpa to remember to type in "backup" into his browser than "" for his backup NAS, or cctv, or printer whatever else.

      As far as I'm concerned, if I don't need to know anything more than gateway and DNS (the two things you really CAN'T refer to by a DNS name), then nobody else does either. I've memorised my VLANs and subnets on each VLAN, though. That matters. But the IP of individual machines? Nope.

      And, to be honest, it REALLY shouldn't matter. Anything that needs to talk to a server should be using the name. Because then transition and retirement is much easier because you just change what the name resolves to without having to have two machines with the same IP trying to failover to each other etc. as you make the switch. Anything else should be picking up a random from DHCP, or literally a "fill-in-the-gap" on your static lists as necessary.

      Too many simple problems are caused by referring to machines DIRECTLY by IP or MAC. Whereas we solved that problem for the Internet by making them all invisible behind a chosen nomenclature.

      Do you know, I don't even know my outside static IP. Because it literally doesn't matter as NOWHERE is it referred to, except the DNS records of my domain. And yet I have a dozen or more outside services for hundreds of users.

      Make your life simple. Choose simple, well-known subnets (the entire 10.0 range is perfectly fine for a local network, nobody will ever have that many devices that it will matter, without having a switch capable of handling such things). Name everything. Use the .1 and .2 as gateway, DNS, etc. done.

    3. James 29

      Re: It's bad to say....

      Thats why we have name resolution. I never bother remembering IP addresses when I can just use a name

    4. Orv Silver badge

      Re: It's bad to say....

      You can always give them arbitrary local IPv6 addresses.

      fd00::1 is the router, fd00::2 is the printer, fd00::cafe is the coffee machine, etc. ;)

      These aren't routable but there's no assumption that a device will only have one address, in IPv6.

    5. Ken Hagan Gold badge

      Re: It's bad to say....

      "But I like going round the house and knowing router is Main PC is 2nd PC is and so on. Can't very well do that with IPv6 can I."

      You can:

      As a bonus, fe80::1 is only 7 characters.

  6. Anonymous Coward
    Anonymous Coward

    In windows this is easy - just need to tick IP6 in the properties.

    So why they don't switch it on everywhere? Must be the Linux that can't do the large numbers?

    1. Roland6 Silver badge

      >In windows this is easy - just need to tick IP6 in the properties.

      You only need to do this if you had previously unticked IPv6 in the properties, Windows has for many years come with v6 enabled by default.

  7. Andre 3

    Stop doing IP addresses and use DNS

    IPv6 will only start taking off when people (thats *you and me*) stop working with dumb IP addresses and start using DNS correctly. The amound of times I have to shout at people who hard-code IP's into apps, files etc gives me an ulcer. THe *point* of DNS is that we dont need to remember stupidly long IP's.

    And don't get me started on the abombination that is CGNAT. WTF - its a too-small plaster for a festering boil, and I hope it breaks and takes all its apps with it.

  8. mark l 2 Silver badge

    My VPS comes with 1 IPv4 address and 30 IPv6 addresses, but yet my ISP doesn't offer IPV6 on its network even though the router does support it. I tried using a IPv6 tunnel but gave up after a couple of hours of unsuccessful messing about so currenly not using any of the 30 IPv6 addresses on my VPS.

    I am sure that my ISP must be running low of IPv4 addresses since their leases are so short that if i switch off the router for 30 seconds and then back on the IP address changes.

  9. Anonymous Coward
    Anonymous Coward

    A few years back, I was working at that company which had a whole /8...

    ... and a few /24 on the side (when they got them, they were actually called Class A and C).

    And they were using a good dozen IP addresses out of them! Okay, maybe a couple of dozens, let's be generous.

    And it got funnier: internally, they were using a different range of public addresses, not allocated to them but to another continent. Their reason: they starting using that range before RFC1918 was published, and the effort needed to change was too big.

    IPv6? They'd heard of it, but rather than not planning for it, they were more like actively resisting the newfangled protocol.

    Aerospace company. Not a very well known name, anyhow, let's leave it at that.

  10. John Sanders

    The lack of IPv4....

    No one is going to move to IPv6 because of this, ISPs have plenty of IPv4 addresses, and not just that, they have the technical expertise to free many more in case they need them from their old networks.

    All it does is increase the price of the commodity which has been made scarce, and prevents competition as the primary resource to start an ISP is IPv4 addresses.

    Enjoy your IPv6 designed at a time where people couldn't possibly anticipate what its adoption would entail 20 years later.

  11. BinkyTheMagicPaperclip Silver badge

    Carrot will always work over stick

    As one of the other commentators say, there has to be an advantage for using IPV6. Either companies have to be generous and offer freebies to IPV6 connections (i.e. Valve offering extra free game weekends over IPV6), or perhaps funding companies to disable adverts on IPV6 for now.

    Go on, reg, if you actually care. Enable IPV6 on thereg, and disable adverts over it for a few weeks, or say ten minutes of no ads for IPV6 users a day. Put your money where your mouth is.

    Consumers do not care about IPV6, but they won't change because their ISP doesn't support it, and their router may not either. If the ISP fixed those two problems, their operating system would very probably 'just work' (all currently supported operating systems support IPV6, and it's available to people who are still using XP..)

    1. DropBear Silver badge

      Re: Carrot will always work over stick

      What you are talking about is the equivalent of bouncing a basketball off a moored oil tanker and expecting it to float away. Not even close.

  12. Paul Johnson 1

    Its the business case, stupid.

    Look at it from an ISPs point of view. Most of their customers have never heard of IPv6, and will require a lot of support and hand-holding as they learn the new system. All sorts of old and obsolete bits of kit will break. Some software won't work. Imagine explaining to Joe Homeworker that they need to tell their corporate IT to upgrade to IPv6 before they can start work again. There will have to be an IPv4 to IPv6 gateway for the foreseeable future, and that is going to be a big headache for all sorts of applications.

    And from an ISPs point of view the current situation is actually very good for business. If you hold a block of IPv4 addresses then you own a valuable commodity. If nobody else can get them then this is a barrier to entry for competitors, which is something that all the business strategy books say is a very important thing to have. So the ISPs are highly disincentivized to migrate to IPv6.

    1. Gerhard Mack

      Re: Its the business case, stupid.

      It is less work than going CG-NAT. For customers, it makes no difference. That's the beauty of dual stack. I have a techie friend whose ISP switched to IPv6 and he didn't notice until I pointed it out. ISP updated the router firmware, assigned a IPv6 IPs, and his computers picked them up and started dual stacking.

  13. johnnyblaze

    We currently have a /21 public range, and we're probably using less than 5% of the address. I occasionally get companies ring asking if we have any address space for sale. I guess those calls could get more commonplace now!

  14. Anonymous Coward
    Anonymous Coward

    I tried it once and didn't like it.

  15. AndrueC Silver badge

    My ISP (IDNet) has been offering it dual stack for several years now.

    I used to be with Plusnet and it seems they still don't offer it even though they rebuilt their network a year or so ago. In fact the only thing IPv6 related that happened during that rebuild is that their IPv6 test servers were turned off so those lucky few on the IPv6 trial lost it.

    Meanwhile their parent company has at least been offering it for a year or so.

  16. Panicnow

    Just Make IP4 addresses 8bytes

    IPv6 is a fail. It was compromised by special interests at standards time.

    Adding a class of IPv4 with 8 bytes (IPv4+4) with being old IPv4 addresses.

    Practically no programming changes needed. and we KNOW IPv4 works and scales!

    OK let IPv6 live next to IPv4+4 and see who wins

    Are the router manufacturers, MS, Apple and Google brave enough? Surely its in their interests to have a working scaling Internet. Otherwise it open up the opportunity to have a "private" proprietary Internet... Oh that's their plan!...

    (I was the founder of First Commercial Internet company in Europe)

    1. Christian Berger Silver badge

      That's called 6to4 and already works nicely

      It's for people who only have an IPv6 connection to connect to IPv4 hosts. AFAIK it uses some sort of NAT mechanism for this. It cannot be done directly as the legacy host would only get the truncated address and therefore couldn't reply.

      Your suggestion would essentially be the same as IPv6, but with much shorter addresses. You'd still have all disadvantages of the switch, but without any of the advantages.

    2. Anonymous Coward
      Anonymous Coward

      Re: Just Make IP4 addresses 8bytes

      "(I was the founder of First Commercial Internet company in Europe)"

      Perhaps you'd like to give your real name then. Because you seem to demonstrate a total lack of understanding of why v6 was designed in the way that it is, and the problem you'd create by introducing your proposal.

      (which is to say that you'd still have to update every bit of network infrastructure anyway, as has already been largely done for v6 - but you'd get few of the benefits of v6)

      The IPv6 debate is settled. It's happening whether the luddites like it or not.

  17. Joe Harrison Silver badge

    Happy New Year!

    Welcome to 2050 - doesn't time fly! Seems like only yesterday we were driving around in manually-controlled cars and buying things with pounds instead of cryptocredits. That was before Korean War II of course.

    Anyway to business - IPV4 addresses are really really REALLY about to run out. What to do about IP V6?

  18. Anonymous Coward
    Anonymous Coward

    Money to be made from a feudal internet

    If all the clients have to connect through CGNATs this enables what the proles do to be more rigorously controlled by everyone from governments which don't like encryption to cloud service providers which will be in control of person to person exchanges and all the crap IOT stuff they flog which captures our every thought and movement. Allow ordinary users to run programs acting as servers on their own premises and all hell breaks loose with this centralised money making control model. Many if not most of the powerful players here clearly have little interest in us upgrading to IPV6.

  19. Alistair Silver badge

    Teksavvy, ontario, DSL

    Dual stack. Have been since I subscribed. I actually can iPV6 VPN to work from here. Although I did immediately get a call from my buddy in the security team. They didn't like that because... they had no throttling on the IPV6 connections into the VPN.

    As for 6 to 4 or 4 to 6. There are a couple of different ways this can be handled, including one particularly interesting suggestion -> to allocate an IPV6 block to be the gateway to IPV4, and to have any device that had IPV6 only on one side and both stacks on the other side effectively become a NAT. DNS lookup on IPV6 only device bolts in the predicate IPV6 block on any ipv4 only return. Going 4 to 6 is somewhat harder, but there are implementations that work, although in the cases I've seen they are effectively reverse proxies.

  20. Pink Duck

    To the max!

    Nothing like a graph that goes to from 0 to 11 :)

  21. J J Carter Silver badge

    I have a cunning plan

    Pass a law that pr0nz sites can only be IPv6, that will force a mass migration as peeps can't do without their interweb o' filth

  22. J J Carter Silver badge

    In the real world

    Our BOFHs can't even get IPv4 subnets to link reliably between on-prem and for home workers on VPN. The chances of them understanding IPv6 is nill.

  23. glnz

    Just moved to Verizon FIOS in NYC, which does NOT support IPv6.

    Let me repeat - the entire US Verizon FIOS network does NOT support IPv6.


POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019