back to article IPv6: It's only NAT-ural that network nerds are dragging their feet...

It has been twenty or so years since the publication of the first drafts of the IPv6 standard, in response to the growing realisation that the IPv4 address space would sooner or later be entirely filled. Fast-forward to today, and amazingly the world is clinging stubbornly to IPv4, with the rate of adoption of IPv6 possibly …

  1. Lee D Silver badge

    *cough*

  2. Degenerate Scumbag

    If past ipv6 articles are anything to go by...

    Cue under skilled network engineers commenting on alleged serious design flaws with ipv6 to cover for their own inadequacy.

    1. Jellied Eel Silver badge

      Re: If past ipv6 articles are anything to go by...

      Ok, so here's one...

      It also simplifies the packet header structure, moving rarely used fields in the IPv4 header into optional header extensions, enabling packet forwarding in routers to become more efficient.

      Or less efficient. v4 has a nice, fixed packet structure. So a 32bit address field. v6 uses 128bits, so 4x larger. Which means 4x more memory needed in forwarding tables on routers. And 4x longer to read an address. Or double that because a packet has a source and a destination. And options in IPv4 are in fixed places, which makes locating and acting on them more efficient given you have a fixed buffer size for a v4 header.

      That also translates into bandwidth efficiency. There's an awful lot of small packets flying around the Internet, so 64 or 128bytes. If v6 is using 32bytes for just addressing then the ratio of header : data gets a lot worse, and those small packets become larger, and link utilisation increases due to header bloat.

      Then there's the issue of checking devices support v6. Which might be nice, although users (especially businesses) may not want every device on their network directly addressable from the outside. Especially if they're IoT things staff have brought in and connected when they shouldn't have. Firewalls are better at dealing with v6 now, but it's still a security concern.

      Then there's native multicasting. That's always been an issue because it pushes delivery handling from the content supplier to the deliverer. So it may save the streaming service money on bandwidth, but pushes stream replication onto ISP's edge routers. And oddly enough, content providers don't want to pay for that... which is why mcasting never really took off.

      1. Paul Johnson 1

        Re: If past ipv6 articles are anything to go by...

        Actually IPv6 will reduce router memory load. The IPv4 space is heavily fragmented. IPv6 will be much less fragmented, so routing tables will have much fewer entries.

        The issue with bandwidth efficiency is not significant. While it is true that there are some tiny packets flying around (ssh sessions spring to mind, with each keypress generating two packets), they don't make a big dent in the overall bandwidth because once the bandwidth goes up the data can be chunked, and the high bandwidth applications like video streaming already use big packets anyway.

        1. Jellied Eel Silver badge

          Re: If past ipv6 articles are anything to go by...

          Actually IPv6 will reduce router memory load. The IPv4 space is heavily fragmented. IPv6 will be much less fragmented, so routing tables will have much fewer entries.

          That's a theory. It's really only an issue for routers that need a full table, but you can track growth here:-

          http://bgp.potaroo.net/v6/v6rpt.html

          v6 has some of the same challenges as v4, ie support for PI addresses, but like for like, a v6 route is longer simply due to header length. And during transition, routers that need full tables may need to hold 2 tables (or VRFs) to support dual stacking. If v6 ends up with similar fragmentation levels as v4, then that also means RIB space, and also FIB. The RIB bloat is less challenging given that changes with routing updates (ie slow) but TCAM memory for the FIB is expensive.. Especially if you need a large FIB on an edge/peering/transit router running Nx100Gbps links. It's also more computationally expensive given you may need to read deeper to get a prefix match than you do for v4. Then there's fixed header length vs v6's 'option' fields.

          (I also think v6 missed a trick by keeping v4's source-destination in the header rather than flipping that to desitination-source. After all, if you're routing/forwarding traffic, you're more interested in the where it's going than where it cam from.)

          The issue with bandwidth efficiency is not significant. While it is true that there are some tiny packets flying around (ssh sessions spring to mind, with each keypress generating two packets),

          Hello, cloud! That's down to looking at link utilisation and packet length distribution. But if a v4 link is currently congested, it'll be more congested if it's converted to v6 simply due to the increased header size.

      2. SImon Hobson Bronze badge

        Re: If past ipv6 articles are anything to go by...

        Or less efficient. v4 has a nice, fixed packet structure. So a 32bit address field. v6 uses 128bits, so 4x larger.

        Nice try there - cite one feature to support an erroneous statement about another.

        The thing is, there are efficiencies in IPv6 by fixing the header format as it does. For example, the hop counter is included in the CRC in IPv4 meaning that it needs to be recalculated on every hop - while in IPv6 it's excluded for efficiency. The extension headers shouldn't need to be checked for routing unless you are doing something quite esoteric and probably won't be present in the majority of traffic anyway - so they are unlikely to have any impact on routing tables.

        And considering that I was using 32 bit addresses back in the days of dial up modems, adding an extra 192 bits to a packet isn't a major issue for most* users. I don't recall too many complaints that the 64 bits of source-address per packet was a problem back then, any more than the 256 bits in IPv6 is today (for most* users).

        * Yes, there will always be some site somewhere stuck with a 1200/75 dial up modem**. But they are unlikely to be connecting directly to the internet and there are various proxy techniques that would allow them to carry on as they are.

        ** OK, suitably antique spec chosen for dramatic effect - but you know exactly what I mean.

        And memory usage is not an issue either. Memory capacities have increase many orders of magnitude, and quite frankly, it's hard to find small memory chips these days that were "cutting edge" only a few years ago.

        And accessing 128 bit addresses need not take longer either - as well as memory getting bigger, we've had a many-fold increase in memory widths. So a 32 bit address meant 4 accesses on an 8 bit system, 2 on a 32 bit system, and you had to be up to a 32 bit system before it because a single access. 64 bit systems are quite common these days, so still only 2 accesses. And don't forget that a lot of what you pay for in "proper" router gear (vs doing it in software) is custom hardware to do the routing table lookup and packet forwarding. So yes, new hardware required, but that's going to have hardware to handle 128 bits address in the same time (well faster now with newer silicon) as older hardware handled 32 bit addresses.

        1. Jellied Eel Silver badge

          Re: If past ipv6 articles are anything to go by...

          The thing is, there are efficiencies in IPv6 by fixing the header format as it does. For example, the hop counter is included in the CRC in IPv4 meaning that it needs to be recalculated on every hop - while in IPv6 it's excluded for efficiency.

          That's only efficient if you assume reliable links. If they're not, then you're reliant on the endpoints doing CRC checks, and potential delays in requests for retransmission. And your lest reliable links are likely to be the most expensive and prone to congestion.. Especially if they're dropping frames at layer-2. Then of course there are potential issues with MTU matching and IPv6's rules on packet fragmentation..

          And memory usage is not an issue either. Memory capacities have increase many orders of magnitude, and quite frankly, it's hard to find small memory chips these days that were "cutting edge" only a few years ago.

          I think you're missing the point. Regular RAM is too slow, especially on BFRs with multiple high capacity links. So FIBs use content addressable memory, or emulations in ASICs that are a LOT more expensive than commodity RAM. Any option fields also become an issue there because the last thing you want to do is have the router process switch packets. If you assume networks are simply routing/forwarding in a dumb way based on source/destination, then you may find it's difficult to implement any fancy SDN features or services.

          And I'd also suggest running Wireshark on your own connection(s) and looking at packet size distribution. Where the payload is only a few bytes, then the increased header size is obviously less efficient.

    2. bjorntoulouse

      Re: If past ipv6 articles are anything to go by...

      Agreed - NAT is embedded in the IPV4 space - the concept of not having to care about address depletion is alien

    3. bjorntoulouse

      Re: If past ipv6 articles are anything to go by...

      One other thing - I always found NAT easier than tunneling - I mean where's mplsv6?

  3. Detective Emil

    Masterly inaction

    $ host -t AAAA www.theregister.co.uk

    www.theregister.co.uk has no AAAA record

    [Was that what you meant, Lee D?]

  4. Pascal Monett Silver badge
    FAIL

    "the world is clinging stubbornly to IPv4"

    If it works, don't fix it. There may well be no new IPv4 addresses, but as I've said before, those we have still work. And NAT is a nifty feature and not only for businesses.

    "as IPv6 takes hold more websites and online services will begin to only be available via IPv6"

    On God's Green Earth, why ? Is there a sudden drop in IQ where IPv6 is concerned ? Have the academics forgotten that businesses exist to make money, not to ensure IPv6 adoption ?

    When a business feels it is pressured enough to have an IPv6 website, that business will ensure that it can still get money from the IPv4 holdouts.

    The day a business will no longer have an IPv4 website is the day funding for IPv4 will become economically unfeasible. That has nothing to do with the day an IPv6 website is accessible.

    1. Chronos

      Re: "the world is clinging stubbornly to IPv4"

      Have the academics forgotten that businesses exist to make money, not to ensure IPv6 adoption ?

      Have the businesses forgotten that the Internet was created by academics to share knowledge and their bottom line is as much a concern as the excretory end of genus rattus?

      1. Will Godfrey Silver badge
        Meh

        Re: "the world is clinging stubbornly to IPv4"

        @ Chronos

        There are countless ideas from academics that were then found to be generally useful.

        Most such academics seem quite happy to receive funding to continue to develop their ideas.

        Currently (rightly, or wrongly) IPv6 is not yet seen to be particularly useful, and is virtually unheard of outside the narrow field of IT.

        1. Len
          Headmaster

          Re: "the world is clinging stubbornly to IPv4"

          To be honest, IPv4 is just as unheard of outside of the narrow field of IT and that is how it should be. Why would the average internet user need to know what TCP/IP is?

          1. vtcodger Silver badge

            Re: "the world is clinging stubbornly to IPv4"

            The world is clinging stubbornly to IPv4. And IPv6 zealots are clinging stubbornly to the notion that everyone MUST switch to IPv6 despite the fact that doing so costs money and provides little or no benefit to the end user. What's wrong with a world where home users and small businesses use IPv4 and their ISP bridges their traffic onto IPv6? Note that most home/small business users are not only uninterested in switching to IPv6, they are quite incapable of setting up IPv6 gear or a dual IP stack.

            Further, unless/until the IoT mess is straightened out, many, probably most, IPv4 users are best off sticking with IPv4 which makes world access to badly designed, poorly secured, digital enabled junk difficult or impossible..I don't know about you folks, but I don't want my scale, toothbrush or printers talking to bored teenagers in Bratislava.

            1. Nanashi

              Re: "the world is clinging stubbornly to IPv4"

              > What's wrong with a world where home users and small businesses use IPv4 and their ISP bridges their traffic onto IPv6?

              The primary issue with this is that it's not possible. I mean, you can use a proxy, but approximately nobody wants to use proxies (as evidenced by the fact that people go straight for NAT, despite its problems).

              (There is 464XLAT or 4in6 to transport v4 traffic over v6, but you still need to deploy v6 to your network or else your hosts will only be able to connect to v4 servers. v4 only has 32 bits available for the dest address, after all.)

            2. defiler

              Re: "the world is clinging stubbornly to IPv4"

              The world is clinging stubbornly to IPv4. And IPv6 zealots are clinging stubbornly to the notion that everyone MUST switch to IPv6 despite the fact that doing so costs money and provides little or no benefit to the end user.

              And then there are those of us looking at IPv6 as (as per the article) an inevitability. Surely it makes sense to have the deployment underway by now at ISPs, and leave the consumer endpoints until later (when they ask, or when the equipment fails, or when the ISP decides that IPv6 is prevalent enough to justify a blanket deployment).

              As has been pointed out, the core routing equipment has been dual-stack for years. The devices at the ISPs are probably all ready to roll, just needing an assignment and a shitload of configuration. That configuration has had years to happen. I would have expected by now that I could call my ISP, say "IPv6, please" and get at least a /64 out of them. On my router. But no - they just don't have it.

              I'm not suggesting that everyone needs to be funneled into IPv6 right now, but that it should at least be an option.

              Of course, if we were to punt all the porn off IPv4 tomorrow, we'd have universal IPv6 by Monday morning...

      2. pɹɐʍoɔ snoɯʎuouɐ
        Trollface

        Re: "the world is clinging stubbornly to IPv4"

        "Internet was created by academics to share knowledge"

        Ahhhh.... I remember those days well...

        you would search for something via webcrawler and actually get a page of relivant results where as now you google and the first page or two are sites trying to sell you a product related to what it is you want to learn about....

        1. Chronos
          Thumb Up

          Re: "the world is clinging stubbornly to IPv4"

          you would search for something via webcrawler and actually get a page of relivant results where as now you google and the first page or two are sites trying to sell you a product related to what it is you want to learn about....

          Spot on. If it was so bloody good, ad men, I wouldn't be looking for firmware to fix its broken-arsedness.

          The microsecond someone comes up with a web search that filters out the SEO chancers, coupon cretins and comparison sites, I'm so there. Remember, if they have to advertise it, you probably don't need it.

          1. JohnFen

            Re: "the world is clinging stubbornly to IPv4"

            "The microsecond someone comes up with a web search that filters out the SEO chancers, coupon cretins and comparison sites, I'm so there. "

            I am so with you on this. Further, I'd be thrilled with a search engine that would optionally omit all commercial sites. As often as not, they just clutter up my search results.

        2. Doctor Syntax Silver badge

          Re: "the world is clinging stubbornly to IPv4"

          "related to what it is you want to learn about."

          And that's only if you're lucky, otherwise they're totally unrelated.

      3. PacketPusher
        Unhappy

        Re: "the world is clinging stubbornly to IPv4"

        Have the businesses forgotten that the Internet was created by academics to share knowledge and their

        bottom line is as much a concern as the excretory end of genus rattus?

        That is correct, and they can use IPv6 to share all of the knowledge that they want, but business is about making a profit and until there is no profit in IPv4, they will continue to support it.

      4. gc23

        Re: "the world is clinging stubbornly to IPv4"

        There is an excellent book called "The Dream Machine: J.C.R. Licklider and the Revolution That Made Computing Personal" that talks about how the Internet came about and how it was very much a military, not academic, project for a long time. Academics weren't really involved with the Internet (nee ARPANET) until much later outside of the Department of Defense scope.

    2. jarfil

      Re: "the world is clinging stubbornly to IPv4"

      I think right now the only online services unavailable over IPv4 are P2P nodes of people on IPv4+IPv6 who didn't care to open ports on their routers. We might actually see P2P services driving IPv6 adoption!

      Other than that I've only seen some servers on just IPv6, but only internal ones, there's always some IPv4 gateway in front of them for the public.

      1. Yes Me Silver badge
        Happy

        Re: "the world is clinging stubbornly to IPv4"

        Anybody running a web site who wants visitors will continue to run IPv4 indefinitely. Obviously. But what is happening now (in some countries but not others) is increasing numbers (I mean millions) of subscribers, especially cell phones, whose primary connection is IPv6, with IPv4 being a second-class service. As major sites react to this (as Google and Facebook did long ago), the major sites will support IPv6 natively as well as IPv4. Obviously, because they want to provide first class service to everybody.

        That being so, it doesn't really matter when the elusive tipping point arrives. Sites will add IPv6 support one at a time. Subscribers will move to IPv6 without knowing it. One day you'll look around and it'll be all over.

    3. Version 1.0 Silver badge

      Re: "the world is clinging stubbornly to IPv4"

      When you look at the world today, you have to ask yourself if adding more address space is actually a good idea ... who really benefits? Facebook I guess, and GCHQ, NSA, the great China Firewall etc.

      I don't really want to be a stick in the mud but the fact is IPv4 is working fine, as it was designed to be - many of those "problems" that IPv6 are supposed to "fix" are really advantages these days. I like the fact that NAT means I can kill a lot of external access at the firewall.

      1. AMBxx Silver badge

        Have the businesses forgotten that the Internet was created by academics

        Don't let the US military hear you talk like that. ARPANET.

        1. Joe Gurman

          Re: Have the businesses forgotten that the Internet was created by academics

          ARPA (now DARPA) doesn't do research. They pay other people to do it. Such as academics and researchers at private industry labs.

      2. Anonymous Coward
        Anonymous Coward

        Re: "the world is clinging stubbornly to IPv4"

        Psst. There is this dark and evil technology out there that allows you to use a NAT router on IPv6. It's got a funny name that starts with an L and it's not unix.

        1. Jamie Jones Silver badge
          Trollface

          Re: "the world is clinging stubbornly to IPv4"

          Psst. There is this dark and evil technology out there that allows you to use a NAT router on IPv6. It's got a funny name that starts with an L and it's not unix.

          Naaa, FreeBSD begins with an "F", and pfsense begins with a "p" !

        2. Yes Me Silver badge

          Re: "the world is clinging stubbornly to IPv4"

          The difference is that for IPv6, things work just fine without NAT. Your firewall works just the same in both cases.

      3. Anonymous Coward
        Anonymous Coward

        Re: "the world is clinging stubbornly to IPv4"

        "I like the fact that NAT means I can kill a lot of external access at the firewall."

        If your preferred firewall doesn't come out of the box with a default 'deny any any' inbound from the 'untrusted' interface for packets that aren't in response to an outgoing connection, then it's rubbish and belongs in the trash anyways.

    4. Anonymous Coward
      Anonymous Coward

      Re: "the world is clinging stubbornly to IPv4"

      IPv6 is necessary for IoT, and IoT isn't just about 'smart' kettles etc.

      For example, yesterday I read the RAIB report on the derailment at Loch Eilt on 22nd Jan 2018, which was due to a landslide. The landslide occurred at some time after the last train of the 21st had run and wasn't detected until the first train of the 22nd ran in to it*

      Whilst systems are available to detect landslides in remote locations they could be described as being fairly primitive, usually relying upon the deflection or separation of a length of one or more wires, strung alongside the tracks, by falling rocks or movement of the underlying ground; they work ok but are a bit Heath-Robinsonish and, being primarily mechanical, are rather fragile (as a safety system) and need regular maintenance.

      An IoT solution, using simple strain gauges and movement sensors, would be both cheaper, provide more information i.e. the exact location and scale of the slide, be more reliable and require less maintenance. But you needs lots and lots of them - IPv6 is ideal for this sort of stuff.

      * The location at which the landslide occurred wasn't considered to be at high risk from such events so the expense of installing one of the current types of warning system there couldn't reasonably be justified; an IoT solution should be a lot less expensive though, through lower initial & maintenance costs.

      1. I ain't Spartacus Gold badge

        Re: "the world is clinging stubbornly to IPv4"

        Exposing rail safety infrastructure to the open internet, rather than having it on an internal network? Ugh! Please no! What when the webcam hackers start taking over your safety systems instead?

        1. Len

          Re: "the world is clinging stubbornly to IPv4"

          What makes you think he is suggesting to put it on the open network? Many IPv6 implementations are in closed networks. Facebook famously switched to using IPv6 only networks internally because they ran out of RFC 1918 address space! Their solution was to just drop IPv4 altogether and only have a bunch of edge servers speak dual-stack.

          1. I ain't Spartacus Gold badge

            Re: "the world is clinging stubbornly to IPv4"

            Well if it's an internal network they can use IPv4 or v6. Firstly nobody else need care, and secondly they're never going to have over 4 billion nodes on their internal network so it makes no bloody difference anyway.

            1. Alan Brown Silver badge

              Re: "the world is clinging stubbornly to IPv4"

              "Well if it's an internal network they can use IPv4 or v6. "

              Perhaps, but the moment they step outside the private IP ranges you run into problems of collisions with resolution of them with external IPv4 ranges.

              I ran into that a number of times 20 years ago when connecting outfits who'd pulled numbers out of their asses when setting up IP networks on the basis of "we'll never connect to the Internet, so any range will do" (128/8 was common). The usual method was double NATting, and you'd eventually get them phoning up panicstricken that they'd been hacked by UC Berkeley due to their internal ranges overlapping with that outfit's public ones.

              So, 4 billion internal IPv4 addresses? Not a good idea.

          2. JohnFen

            Re: "the world is clinging stubbornly to IPv4"

            "What makes you think he is suggesting to put it on the open network?"

            Because putting this on a private network means that the argument is irrelevant to IPv6 migration on the internet.

          3. Anonymous Coward
            Anonymous Coward

            Re: "the world is clinging stubbornly to IPv4"

            > Their solution was to just drop IPv4 altogether and only have a bunch of edge servers speak dual-stack.

            The large finance company I work for is going down the same road. We had some struggles during the last corporate merger due to significant overlap in RFC 1918 network ranges. We also have to use source and destination masquerading on our private tunnels running to third parties due to similar range overlap.

            Our long term goal is to have our perimeter firewalls and load-balancers running dual-stack so they can perform NAT-4to6, while having everything else behind them running native IPv6.

            Besides eliminating future potential merger issues, the company is motivated to complete this migration so that they can sell their sizable holding of IPv4 addresses while they still have significant value.

        2. Chronos

          Re: "the world is clinging stubbornly to IPv4"

          But you don't have to. As has been said time and again, you can NAT IPv6. Consumer kit should NAT IPv6. Just don't take the choice away from those of us who want to run services by sticking us all on CG-NAT v4.

        3. gc23

          Re: "the world is clinging stubbornly to IPv4"

          I guess you've never heard of VPN then, right? There's a lot of critical infrastructure that uses the public Internet through VPN.

        4. Yes Me Silver badge
          Headmaster

          Re: "the world is clinging stubbornly to IPv4"

          Just because a sensor has an IPv6 address doesn't mean it's exposed to the open Internet. Just because a baby alarm with an IPv4 address is behind a NAT doesn't protect it from a malicious user. Security is a completely separate thing from addresses, and NATs are not security devices.

      2. Roland6 Silver badge

        Re: "the world is clinging stubbornly to IPv4"

        >IPv6 is necessary for IoT...

        So what does that have to do with the deployment of IPv6 to corporate desktops?

        Which I think gets us back to the entire premise of the article, a total misunderstanding of the situation.

        The major network equipment manufacturers have for years been shipping kit that supports both IPv4 and IPv6, Microsoft since Win7 [aside might have been Vista but...] have shipped Windows with an IPv6 stack that works out of the box. Apple with iOS 4.1 (2010) and OS X (since 2002) included IPv6 support.

        So in principle, there should be little reason why your typical home (a couple of MS/Apple systems, a printer, a games console, a couple of phones and an ISP supplied and configured router) shouldn't be using IPv6 for many functions - with the users being largely unaware.

        However, I suggest the following should be giving concern:

        1) The vast majority of ISP supplied routers are IPv4 only.

        2) Many high street printers are still IPv4 only

        3) Only with the release of Android 4.0 Ice Cream was support for IPv6 included. However, from various forums it seems that even today not all devices support it and even if they do it may only be on the mobile data connection and not the WiFi connection.

        4) Most residential ISPs only deliver an IPv4 service.

        5) ? What isn't clear is what is being used natively on 3/4G networks. I would assume from the lateness of IPv6 support in mobile devices that 3&4G are natively IPv4 but stand to be corrected.

        Address these and the amount of traffic using IPv6 will significantly increase, without joe public even realising it...

        With a large user/consumer-base using IPv6 it becomes easier to get people to make their websites accessible on both IPv4 and IPv6 - it should only be a configuration tickbox.

        As for corporate networks, well back in the 80's and early 90's they were running all sorts of stuff: SNA, DECnet IV, OSLan, TCP/IP, XNS, Novell etc. yet within a very short space of time the vast majority of networks switched to IPv4, in part because the vendors (including IBM) stop developing their proprietary networking solution and migrated their product to IPv4... So I suggest once you are running an IPv6 service externally and probably also within your cloud-based datacentre, switching off IPv4 becomes a no brainer cost saving.

        1. gc23

          Re: "the world is clinging stubbornly to IPv4"

          > The vast majority of ISP supplied routers are IPv4 only.

          > Most residential ISPs only deliver an IPv4 service.

          I think you hit the nail on the head with that. I'd say that it's also a supply and demand issue: if people started demanding v6, the ISPs would start supplying it. Outside of professional IT, however, no one cares what happens behind the scenes if their Internet connection works as they feel it should.

        2. vtcodger Silver badge

          Re: "the world is clinging stubbornly to IPv4"

          "Microsoft since Win7 [aside might have been Vista but...] have shipped Windows with an IPv6 stack that works out of the box."

          Right. And you are aware that Microsoft had substantial difficulty switching to IPv6 internally not all that long ago? e.g. https://www.theregister.co.uk/2017/01/19/windows_10_bug_undercuts_ipv6_rollout/

        3. casperghst42

          Re: "the world is clinging stubbornly to IPv4"

          Any ISP who has IPv4 only Routers has not upgraded their hardware since before 2012 ... the question is if they have current code on them, and if they have the latest IPv6 stack. But any serious IPS can support IPv6 on the router side, and any homeuser should have Routers which support IPv6.

          1. eldakka

            Re: "the world is clinging stubbornly to IPv4"

            Any ISP who has IPv4 only Routers has not upgraded their hardware since before 2012 ...

            Many ISPs use 'own branded' routers, like many large telco's use 'own branded' mobile phones. They are big enough to have the actual hardware manufacturer remove their branding from the device, both hardware and firmware, rebadge it with the ISPs branding, and provide custom built firmware to the ISPs specifications.

            I recently got an 'own branded' router/modem from my ISP, and there is no firmware available for it from any other source than my ISP. Even tho I do know the manufacturer of the device, they only build devices for 3rd parties. So I can't download a generic firmware from the manufacturer directly, as they provide custom builds to the ISPs that only enable the specific features that the ISPs want in the devices. The device is branded with all the ISPs brands, the firmware is all ISP-branded with no references anywhere on the device or in the firmware as to who the actual manufacturer is.

            This is pretty typical for the cheapest (i.e. most popular) devices available from an ISP. Therefore there is much new hardware out there that only provides very limited functionality, because the ISP doesn't want the end-users to have the ability to make any of their own choices in the configuration.

            Luckily the router does have a bridge mode option, so I can use it just as a modem and use a real router behind it. However, most home users won't want to spend the money, let alone have the knowledge to, configure a separate routing device from the modem.

        4. Yes Me Silver badge

          Re: "the world is clinging stubbornly to IPv4"

          When it changes is when your home network includes internal routers and many wired and wireless segments that are not bridged together. That's coming real soon now and only with IPv6 and HNCP (Home Networking Control Protocol).

          1. Roland6 Silver badge

            Re: "the world is clinging stubbornly to IPv4"

            >That's coming real soon now and only with IPv6 and HNCP (Home Networking Control Protocol).

            Sorry if I'm a little restrained in my excitement - I worked on various Home Networking Standards in the 1980's when home networking was also supposed to be "coming real soon now"...

        5. Roland6 Silver badge

          Re: "the world is clinging stubbornly to IPv4"

          >So in principle, there should be little reason why your typical home ... shouldn't be using IPv6 for many functions - with the users being largely unaware.

          To illustrate this point, in investigating a Windows printer communication problem, I realised that my recently purchased printer was dual stack, given my iPads are dual stacks and my Win7 and Win10 laptops are also dual stack and all this was running straight out of the box, without further investigation, I have little idea which protocol stack these systems are using to communicate with my printer. Yes, my router might only be talking IPv4 to the world, but this doesn't impact the home LAN...

      3. Joe Gurman

        Re: "the world is clinging stubbornly to IPv4"

        Erm, is there any reason all those shoddily (or not at all) secured IoT devices should be directly on the Net, rather than behind a NAT gateway?

        1. Nanashi

          Re: "the world is clinging stubbornly to IPv4"

          Nobody is suggesting that your IoT stuff should be directly on the internet. We're just saying that NAT is an unnecessary headache, that it breaks too much stuff and requires too many compromises, and that you should be avoiding it.

          You do not need NAT in order to connect your devices via a router.

          1. Anonymous Coward
            Anonymous Coward

            Re: "the world is clinging stubbornly to IPv4"

            You do not need NAT in order to connect your devices via a router.

            On the flipside, you don't need to persuade people to give up NAT to persuade them to move to IPv6. I don't understand why so many people make it harder by insisting that they must do both at the same time?

            1. Yes Me Silver badge

              Re: "the world is clinging stubbornly to IPv4"

              "you don't need to persuade people to give up NAT to persuade them to move to IPv6."

              People, in the sense of domestic or cell phone subscribers, don't need to know anything about it.

              If you mean professionals running small office networks, they might need to know, but not necessarily.

              If you mean professionals running enterprise or campus networks, they mostly know already. Including the knowledge that you can run your IPv4 connection through NAT+firewall, your IPv6 connection through firewall only, and get exactly the same security protection.

              1. Anonymous Coward
                Anonymous Coward

                Re: "the world is clinging stubbornly to IPv4"

                People, in the sense of domestic or cell phone subscribers, don't need to know anything about it.

                But they will.

                If the anti-natters were willing to wait their turn we could have the world on IPv6 tomorrow. Sadly they've none of the emotional maturity required to realise when their pet quibble needs to be put on a back burner, so they become one of the big obstacles that IPv6 supporters have to deal with.

                1. Nanashi

                  Re: "the world is clinging stubbornly to IPv4"

                  You can do NAT with v6 already. Nothing stops you apart from the sheer silliness of it.

                  It's been four days since your post, and now that Google's stats (which lag behind by a couple of days) have updated, they're showing that deployment is still at the 20%/24% week/weekend cycle it's been at for a while, and hasn't jumped up to 100%.

                  So no, it seems like the ability to do NAT on v6 didn't lead to the world being on v6 a day after your post. This suggests that the main blockers are probably elsewhere.

          2. Doctor Syntax Silver badge

            Re: "the world is clinging stubbornly to IPv4"

            "We're just saying that NAT is an unnecessary headache, that it breaks too much stuff and requires too many compromises, and that you should be avoiding it."

            Millions of people are, however, running NAT and haven't even noticed that their stuff is broken.

          3. JohnFen

            Re: "the world is clinging stubbornly to IPv4"

            "You do not need NAT in order to connect your devices via a router."

            I don't think that anybody is arguing that you do. However, using IPv6 doesn't mean that NAT provides no value and should be discarded.

        2. Alan Brown Silver badge

          Re: "the world is clinging stubbornly to IPv4"

          "Erm, is there any reason all those shoddily (or not at all) secured IoT devices should be directly on the Net, rather than behind a NAT gateway?"

          Actually what happens is that because of CGNAT, they tunnel out to servers in other parts of the world to get a stable address and that's where a lot of the problems start rearing their ugly head (essetnially creating a VPN backdoor into your NATed home network.

      4. JohnFen

        Re: "the world is clinging stubbornly to IPv4"

        "An IoT solution, using simple strain gauges and movement sensors, would be both cheaper, provide more information i.e. the exact location and scale of the slide, be more reliable and require less maintenance. "

        That's true, but is not an argument for using the internet for this sort of thing. This is exactly the sort of thing that belongs on a private network.

    5. gnarlymarley

      Re: "the world is clinging stubbornly to IPv4"

      When a business feels it is pressured enough to have an IPv6 website, that business will ensure that it can still get money from the IPv4 holdouts.

      It very well could be that the holdouts are avoiding IPv6 due to its IPv6 built in rotation of addresses, which can make it harder to track who is who than just normal NAT.

  5. Chronos
    FAIL

    Second class netizen

    CG-NAT cannot be considered "Internet." At best, it's a glorified proxy. A true Internet connection is the full 65535 ports bi-directional.

    We've only ourselves to blame. We've had two decades to address (sorry) this yet we're still arguing about what colour the bike shed should be¹.

    ¹ NAT vs a few lines of iptables code is starting to look a bit like barrel scraping for excuses not to deploy.

    1. Phil O'Sophical Silver badge

      Re: Second class netizen

      CG-NAT cannot be considered "Internet." At best, it's a glorified proxy. A true Internet connection is the full 65535 ports bi-directional.

      But since few machines ever need more than a few dozen of those ports they can effectively be considered as adding 16 bits to an IPv4 address field, which is what NAT does. It might be ugly for the purists, but it works.

      We've only ourselves to blame. We've had two decades to address (sorry) this yet we're still arguing about what colour the bike shed should be

      Well, perhaps the people who designed the bike shed should have asked themselves whether we needed a new bike shed in the first place, or whether a bigger garage would have been better. IPv6 was proposed at a time when completely changing the "whole internet" was a smaller and still possibly manageable task. It's a lot harder today.

      1. Chronos

        Re: Second class netizen

        But since few machines ever need more than a few dozen of those ports they can effectively be considered as adding 16 bits to an IPv4 address field, which is what NAT does. It might be ugly for the purists, but it works.

        FSVO "works." It doesn't play particularly well with SIP. FTP active mode is right out. Admittedly these are things that the average cat video consumer isn't likely to be worried about right now.

        Yet there's the equivalent of the slowly boiling frog going on here. You don't really need an understanding of nuclear fission yet I bet you have it. Bringing us all down to the lowest common denominator is dangerous. Ignore the loss of functionality at your peril while it stifles future innovation and freedom of choice under the duvet of NAT.

        I remain amazed that the folks on here can't see the potential problems of these second class connections.

        1. This post has been deleted by its author

          1. Anonymous Coward
            Anonymous Coward

            Re: Second class netizen

            Downvoted for using "FTP active" as a reason for anything...

        2. Doctor Syntax Silver badge

          Re: Second class netizen

          "I remain amazed that the folks on here can't see the potential problems of these second class connections."

          Nevertheless if you talk to those who are using such connections and tell them about their potential problems they'll look at you blankly and say "but it works".

          1. Alan Brown Silver badge

            Re: Second class netizen

            "Nevertheless if you talk to those who are using such connections and tell them about their potential problems they'll look at you blankly and say "but it works"."

            So did people who had the most amazingly shitty TV pictures until the day they saw a decent display on their neighbour's set thanks to a decent antenna

      2. Roland6 Silver badge

        Re: Second class netizen

        > IPv6 was proposed at a time when completely changing the "whole internet" was a smaller and still possibly manageable task.

        IPv6 was the result of giving 'academics' (ie. those with a more purist and theoretical bend) the opportunity to completely change the Internet at a time when it would have been possible to do so, unfortunately, they didn't realise the door for change was only open for a very brief period of time - with the door effectively being closed on the release of Windows 95...

        Looking back it is obvious that a much more pragmatic approach should have been taken that allowed for a rapid move to a more extendable IPv4 like protocol which would have allowed yet to be defined features such as a larger address space to be added. ie. take the lessons learnt from ISO OSI network addressing and do better. However, to give the authors of IPv6 their due, no one at the time really appreciated just how quickly the Internet would take off, plus as they in the main hadn't been involved in IEEE/ISO OSI/MAP/TOP/GOSIP had very little idea just how long getting technical design and agreement took.

        1. Anonymous Coward
          Anonymous Coward

          Re: Second class netizen

          IPv6 was the result of giving 'academics' (ie. those with a more purist and theoretical bend) the opportunity to completely change the Internet

          And therein lies the problem. It appears that the engineers were not consulted and the IPv6 protocol was designed to be incompatible with IPv4 - not a good idea.

          1. Jason Bloomberg Silver badge
            FAIL

            Re: Second class netizen

            It appears that the engineers were not consulted and the IPv6 protocol was designed to be incompatible with IPv4 - not a good idea.

            A very bad idea and why IPv6 has mostly been ignored.

            People will move with the times, adapt to new features being added, but most are reluctant to take a huge jump sideways into new territory. Expecting anyone to put aside all they have learned and begin again from scratch is a huge ask, and especially when the need cannot be demonstrated.

            I would probably be using IPv4+ now if it existed. I have barely even looked at using IPv6.

          2. Gerhard Mack

            Re: Second class netizen

            "And therein lies the problem. It appears that the engineers were not consulted and the IPv6 protocol was designed to be incompatible with IPv4 - not a good idea."

            Are you honestly saying that when AT&T, Microsoft, DEC, Cisco and NTT all participated in the working group that helped decide on IPv6, they didn't send engineers?

            1. Anonymous Coward
              Anonymous Coward

              Re: Second class netizen

              "Are you honestly saying that when AT&T, Microsoft, DEC, Cisco and NTT all participated in the working group that helped decide on IPv6, they didn't send engineers?"

              Hard to say without being there. We can only judge by the outcome.

              By comparison (contrast?) how many engineers has ICANN got working on the "whois problem"? Or anything else? From the outside it seems like mostly business people, lawyer types and bureaucrats. Turns out many committees fit that pattern.

              Just sayin ....

          3. SImon Hobson Bronze badge

            Re: Second class netizen

            It appears that the engineers were not consulted and the IPv6 protocol was designed to be incompatible with IPv4 - not a good idea

            And pray tell, how would you propose to create something with more address space that IS compatible with IPv4 equipment ?

            Not a single piece of IPv4 kit would be capable or sending or receiving packets with (say) a 64 bit address, or a 48 bit address, or indeed any packet with other than a 32 bit address in the 32 bit field where the existing code expects to put/find 32 bits. So no, it's not going to work to say "lets fix this problem but maintain compatibility".

            Once you accept that bit of reality - that whatever you do will require new code along with everything that goes with it - then most of the arguments against IPv6 really start looking rather lame. Frankly, if we're going to have to have a bunch of new code, then lets at least make a big jump so that by the time we've adopted it we aren't already having to plan the next jump !

            And then a lot fo the perceived difficulty with IPv6 is actually fixing some significant problems that existed with IPv4 - but which most people, even the majority of network savvy people, don't realise are there.

            And don't get me started on NAT - the millions, or even billions of man hours wasted on working around that breakage (not to mention the hardware investment in things like proxy servers) doesn't bear thinking about. Most people don't see NAT as a problem because of all those man hours spent making things work - the reason they (most of the time) don't see problems isn't because there aren't problems to be seen.

            1. Roland6 Silver badge

              Re: Second class netizen

              >And pray tell, how would you propose to create something with more address space that IS compatible with IPv4 equipment ?

              One of the solutions proposed at the time (early 90's) was for a simple address enlargement, leaving everything else the same. By combining this with a 1-to-1 IPv4 to IPv6 address mapping - hence the IPV4 address space within the IPv6 address scheme, a relatively efficient protocol gateway can be built. Obviously, for the transition period this placed a restriction on the IPv6 address blocks that could be used.

              I think you can see that whilst it might have been a feasible transition strategy - in a private network, it has too many negatives for it to be feasible in a public network. I seem to remember this was a problem common to many proposals. Additionally, I suspect many of the proposed solutions would have worked on the early 1990's Internet, but would be unworkable on today's larger and more complex Internet.

  6. Herby

    Obvious need for..

    Nice home routers that work well with IPv6. Most of the ones I see are woefully inadequate in this regard. We also need to have the DSL & Cable modems work with IPv6 addresses.

    Then there is the addressing of IPv6, which allows for an IP address for every grain of sand the world has known (*SIGH*).

    I suspect that IPv7 (whatever that will be) might be more to the point, and possibly backward compatible with IPv4 with little effort.

    Then again wishing and hoping will only get you so far.......

    1. TRT Silver badge

      Re: Obvious need for..

      IPv8. Even numbers are the release versions.

    2. Wellyboot Silver badge

      Re: Obvious need for..

      >>I suspect that IPv7 (whatever that will be) might be more to the point, and possibly backward compatible with IPv4 with little effort.<<

      It would need to be compatible with v4 AND v6, good luck getting that built by a committee.

      v6 started out as only 'We'll need more address space fairly soon'

      1. Len

        Re: Obvious need for..

        A backward compatible future technology will not help. The problem is that IPv4 is not forward compatible and you can't solve that without replacing IPv4. If replacing IPv4 devices is inevitable one might as well go the whole hog and fix a list of issues beyond lack of addresses.

        1. JohnFen

          Re: Obvious need for..

          "If replacing IPv4 devices is inevitable one might as well go the whole hog and fix a list of issues beyond lack of addresses."

          It was exactly this thinking that introduces so much friction to changing to IPv6. If they hadn't gone whole-hog all at once, then everyone would have already made the change. In terms of encouraging adoption, it's better to have a series of relatively small changes over a single ginormous change.

          1. Charles 9

            Re: Obvious need for..

            UNLESS even the smallest change breaks the mold. That's the problem with IPv4. It was designed (DECADES ago) for fast processing on limited memory and computing power. Ergo, it was designed with a fixed (meaning impossible to adjust) packet structure. It wasn't designed with forward compatibility in mind, and without that, there's no way you can make an extended IPv4 work without breaking something. Otherwise, how does one cram 13 eggs in a carton only built for 12 without breaking something along the way? And no, something like EzIP won't work very well without the same kind of infrastructures you would need in IPv6 anyway (special edge routers et al--guess what NAT46 and NAT64 routers do).

      2. Anonymous Coward
        Anonymous Coward

        Re: Obvious need for..

        v6 started out as only 'We'll need more address space fairly soon'

        A bit like "we'll need more lanes on this road to handle traffic". The pragmatic solution is to widen the road, but all too often the idealists decide to build a railway instead "because it's better" and then they blame us when the traffic jams are huge & the trains empty.

      3. JohnFen

        Re: Obvious need for..

        "v6 started out as only 'We'll need more address space fairly soon'"

        Yep, and if it had stuck to solving just that problem, we'd all be using it right now. Instead, it got feauture-creeped to the point where changing to it is a much more substantial task than it had to be to solve the original issue.

    3. BanburyBill

      Re: Obvious need for..

      OK, so explain to me. In detail. How do you increase the size of an IPv4 address *without* changing IPv4 packet structure in a way that will require updating the software (and very possibly hardware) on every router on the planet? Hint: extending the header, or changing field usage therein fails that test.

      The designers of IPv6 weren't (and aren't) idiots, and they weren't (as often suggested) detached academics either, unless you think the likes of Cisco, Google, Nokia and Ericsson are detached academe.

      1. Anonymous Coward
        Anonymous Coward

        Re: Obvious need for..

        How do you increase the size of an IPv4 address *without* changing IPv4 packet structure in a way that will require updating the software (and very possibly hardware) on every router on the planet?

        You don't, obviously.

        What you do is create a compatible protocol so that:

        - New routers can recognize and handle both formats

        - Old routers can at least identify enough of a new-format packet to route it to some catchall system that can handle it properly.

        IPv6/IPv4 dual stack handles the first issue, but IPv4 systems just barf when given an IPv6 packet & throw it away, thus preventing any interim solution. You can't connect an IPv6 device to anything but an IPv6-capable network.

        Telcos understand this, which is why incremental phone system & number upgrades, although often inconvenient, still work. It's a pity that the internet academics placed purity of design that requires a forklift upgrade ahead of a potentially-ugly but workable incremental model.

        1. BanburyBill

          Re: Obvious need for..

          So, old routers need to handle particular IPv4 packets in a special way. So you need to update them. And how exactly does the catchall system magically deliver packets half your network doesn't recognise?

          1. Anonymous Coward
            Anonymous Coward

            Re: Obvious need for..

            So, old routers need to handle particular IPv4 packets in a special way. So you need to update them.

            No, you specifically don't want to upgrade them.

            And how exactly does the catchall system magically deliver packets half your network doesn't recognise?

            It does recognise them, it has to for the kluge to work, it just thinks they are IPv4 packets and routes them as such. Perhaps with special addresses which mean "this is an IPvX packet, the real address is somewhere else". IPvX finds the address, existing IPv4 just routes them based on the special address which is guaranteed to terminate on a dual-stack system which can process them. Or maybe by using the options field, etc.

            There are plenty of ways to kluge a protocol so that old systems think it means one thing, and new systems can find the extra info and do something else with it. Phone companies have been doing it for years.

            1. Gerhard Mack

              Re: Obvious need for..

              "There are plenty of ways to kluge a protocol so that old systems think it means one thing, and new systems can find the extra info and do something else with it. Phone companies have been doing it for years."

              Not the best example. Phone numbers are really just the equivalent of DNS at this point. And thanks to the fact that there is no central authority, the the phone companies ended up with a completely insecure setup that has allowed attackers to reroute calls and SMS. The vulnerabilities are well known and have been used in the past to intercept bank's 2 factor authentication SMS and break into people's accounts. Also, people would be royally pissed if it took as long to connect to a website as it does to establish a phone call.

              Also, the rest of your argument is nonsensical. You can't route packets with extended features without upgrading the core routers that would need to route based on those very extended headers.

        2. Baldrickk

          Re: Obvious need for..

          What you do is create a compatible protocol so that:

          https://xkcd.com/927/

      2. Anonymous Coward
        Anonymous Coward

        Re: Obvious need for..

        The solution was to take IPV4 and make the packet size and address field larger, make IPV4 address part of the V6 address space ( ie: 0.0.0.0.1.2.3.4 == 1.2.3.4 ), name it IPV6 and go to the pub.

        1. SkippyBing

          Re: Obvious need for..

          'The solution was to take IPV4 and make the packet size and address field larger, make IPV4 address part of the V6 address space'

          Something like this. I'm not up to speed on the technicalities of IPv4 vs v6, but if you want widespread adoption of a new technology it needs to be backwards compatible, because people. As a prior example see the roll-out of colour television, where the signal was backwards compatible with black and white sets, that made it a painless upgrade when you came to replace your set* because that's all you had to change, not even a new aerial.

          Now if I wanted to go IPv6 god knows how many not working boxes I might have in my house because they only talk IPv4, and as a consumer I have better things to do than deal with that by buying more boxes or babel boxes that convert between the two.

          *Probably in about 20 years because these things weren't disposable at the time.

          1. Len

            Re: Obvious need for..

            'The solution was to take IPV4 and make the packet size and address field larger, make IPV4 address part of the V6 address space'

            That is precisely the problem with the lack of forward compatibility in the IPv4 standard. Any minor change to IPv4, like adding an extra octet or even an extra header, would mean that existing IPv4 devices wouldn't understand it.* That lead to people, decades ago, realising that if you need to replace or reprogram existing devices you might as well do it properly and solve a whole bunch of other things too.

            * Hell, there are even ranges in IPv4 that were 'reserved' for future use that are now lying unused because over the years quite a few manufacturers decided to cut corners and not implement them. Nobody will want to get those IP addresses assigned to them now as you never know if they get dropped somewhere along the way. Fixing that will, again, require a large scale replacement so you might as well go straight to IPv6.

          2. Anonymous Coward
            Anonymous Coward

            Re: Obvious need for..

            > if you want widespread adoption of a new technology it needs to be backwards compatible, because people. As a prior example see the roll-out of colour television, where the signal was backwards compatible with black and white sets...

            For most 525-line and 625-line systems, backwards compatibility wasn't difficult. In those systems, adding color was nothing more than adding supplemental information within the payload. Almost everything else about the broadcast remained the same. For terrestrial broadcasts, you also had the benefit of there being no intermediary devices between the transmitter and receiver that had to be color-aware.

            But for a backwards compatible IPv6 system, things become very difficult because you need end-to-end awareness for it to work. What were to happen if the extra address length was added as an IPv4 header option and it passed across a router that wasn't IPv6-aware? That router could strip the information, resulting in a misdirected packet.

            Sometimes, passive backwards compatibility just isn't possible. Just look at the "dual-stack" period that Ireland and the UK had while it supported both the newer UHF 625-line and older VHF 405-line systems. Between the incompatible frequencies, channel width, and timing, the only way to bridge the two standards was through the use of translator stations that converted 625-line content down to 405-line signals for older televisions.

            In the case of IPv6, proxies and network address translation devices act as a equivalent to those translator stations. Sure, it adds complexity in some ways, but it also simplifies things in other ways.

          3. Alan Brown Silver badge

            Re: Obvious need for..

            "I'm not up to speed on the technicalities of IPv4 vs v6"

            That much is obvious. Various ways of trying to but out compatible ways forward were discussed and discarded because it was clear they would end up as a Rube-Goldberg (Heath-Robinson for the ukites) mess from the outset and things would go rapidly downhill from there.

            IPv4 was not designed with expandability in mind. It was intended to be a temporary solution to a pressing problem with a 5-year life expectancy whilst a "real" internet protocol (IPv5) was under development. That "solution" turned out to be IPX, which turned out to be completely unusable because Novell forgot that it needed to be widely routable.

            At its core, IPv6 _IS_ just IPv4 with extra bytes tacked on. The way its written is different but you can use octets if you really want to (the format change is to keep it manageable) and it still has 65536 ports, etc. There are some extensions to add functionality and extra undefined type fields to handle future growth in types, because IPv4 demonstrated that what was projected wasn't enough.

            The single biggest problem is that there is NO WAY WHATSOEVER that an IPv4 host can initiate a connection to an IPv6 one. It would be like a 2 dimensional being trying to connect to a 3 dimensional one. Or to put it another way, it would be like trying to communicate from our universe to other bubbles in the multiverse. We simply can't address those spaces whilst sitting in this frame of reference.

            IPv4 is full. There are no more addresses being issued - ever. There's some horse trading of existing ones around the edges which is simply making the existing routing quagmire even worse whilst people are kidding themselves that "Ip addresses" are valuable. meantime, there are quadzillions going begging if you look in the other direction, to the tune of a few million "IPv4" IP ranges person.

            Whoever pays out millions for a few IPv4 subnets and then has to explain to the investors that they're essentially worthless is going to be an unhappy camper. There are already fairly large tracts of netspace that are unreachable from IPv4-space. Even if Joe-average consumer can't see that those of use who need to deal with the US, EU or Chinese high speed academic networks are already feeling it and IPv6 is gaining popularity across asia thanks to the effects of multiple levels of IPv4 CG-NAT.

            If you want to beat on the Brexit drum, live in your little IPv4 walled garden and dream of the days of empire and a smaller network world, then feel free, but bear in mind that the last outfits to leave IPv4 space will be the marketers - who love the stick-in-the-mud reluctant shifters, so you can look forward to an ever-increasing proportion of spam as the rest of the world moves to ipv6-only.

        2. Jellied Eel Silver badge

          Re: Obvious need for..

          The solution was to take IPV4 and make the packet size and address field larger, make IPV4 address part of the V6 address space ( ie: 0.0.0.0.1.2.3.4 == 1.2.3.4 ), name it IPV6 and go to the pub.

          That was a proposal. Telco world has managed this (and lived with the pain) by simply adding digits. So add another octet to v4 and assign based on ITU (cough) country codes. UK gets to be 44.0.0.0.0-44.255.255.255.255. As a bonus, it'd make routing more efficient, and assuming allocation rules were followed, geolocation more reliable.

          Problem was IPv6 was designed by committee, so the usual problems of feature creep. Namely MAC-IP address mapping.. Which overlooks some.. issues with MAC address allocation, overlap etc. But resulted in application layer stuff being pushed onto the network layer, which is much happier when it's just routing/forwarding/switching.

          1. Roland6 Silver badge

            Re: Obvious need for..

            >Problem was IPv6 was designed by committee, so the usual problems of feature creep.

            Also most of the committee were purists and so rather than be pragmatic and agree a solution that everyone was prepared to implement in a very short space of time (the Internet then was both very small and maintained by a very small community), they wanted to solve a whole host of other issues...

            1. Jellied Eel Silver badge

              Re: Obvious need for..

              Also mobile operators. They were largely behind the increase pushing for an IP address for every phone. Some time later, they still have IMSI/IMEI and keep their walled gardens to make it harder for anyone but themselves to make money off mobile users. Like you say, challenges began when non-ISP's got more involved in the decisions. It's kind of the downside to democratic decision making processes. Everyone gets a vote, then the ISPs get to implement it.

              But then the ISPs or telcos typically focus more on MPLS because switching is faster than routing. And the header's even smaller. Operators kinda won on that one, which resulted in some sour grapes after people noticed V4 ToS bits map nicely onto MPLS labels, meaning DSCP can be quietly ignored. Not like customers really need 64 levels of granularity in deciding which packets to drop, and given 'Net Neutrality', it's ignored anyway. Sort of.. :)

        3. Doctor Syntax Silver badge

          Re: Obvious need for..

          "The solution was to take IPV4 and make the packet size and address field larger"

          The tricky bit of this is doing it in a way that lives with a lot of kit that doesn't know what to do with such a packet. There needs to be a way for your extended protocol to allow a node using it to be able to recognise that the node it's trying to talk to is plain old IPv4 only

          1. Anonymous Coward
            Anonymous Coward

            Re: Obvious need for..

            Obviously you'd have to upgrade the routers and servers to IPv6, but the main thing is that you wouldn't have to reconfigure them.

            Listen 1.2.3.4:80

            becomes interpreted as:

            Listen 0.0.0.0.1.2.3.4:80

            Network settings on the server remain the same:

            Subnet: 1.2.3.4/24

            Is interpreted as 0.0.0.0.1.2.3.4/24

            etc.

            Companies would have sold IPv6/4 proxies for servers which couldn't be upgraded, which could have kept track of the v6 address of the source and used a fake address when talking to the server.

            We'd have been on v6 a decade ago under this cunning plan.

    4. Len
      Happy

      Re: Obvious need for..

      In hindsight 2012 turned out to be 'The Year of IPv6 on Consumer Equipment'. That was the year when a dozen or so most popular router brands added IPv6 support en masse. When Comcast started its mass roll-out of IPv6 in 2012, millions of their users became guinea pigs to find the teething problems with these early consumer implementations.

      Six years later I think it's safe to say that most people wouldn't even notice if their ISP would switch to dual stack (IPv4+IPv6) because their modem or router will handle that just fine. My parents have had a dual stack connection at home for years, they have no idea that when they are watching something on YouTube, Facebook, their news website or Googling something that it is delivered over IPv6. It just works.

    5. Nanashi

      Re: Obvious need for..

      I see there's quite a bit of "v6 should've been backwards-compatible" circle-jerking here... do you guys really not realize that v6 is backwards compatible? You can connect from v6 to v4 via NAT64, and you can run v6 islands accessible over the v4 internet with 6to4.

      Those are roughly the only forms of backwards compatibility that are possible with the design of v4, and v6 has them. There are some other forms of backwards compatibility that are impossible, and v6 doesn't have those because those are impossible.

      What else could we possibly have done??

  7. Joe Montana

    Glorified proxy

    A few years ago you had ISPs which advertised that you were "part of the internet" because you got a dedicated ipv4 address, as opposed to some lesser providers which put you behind some kind of proxy service...

    There are plenty of routers that work fine with ipv6, providers like sky and bt are now providing ipv6 by default with the routers they supply for instance, most users don't notice the difference.

    What will spur ipv6 adoption is creating demand for it - offer services that either require ipv6, or work better with it. Microsoft do this to some degree with the xbox one which declares that it works better if it has ipv6 connectivity, but providers could do more to encourage this. Many providers offer beta access to various services to a limited audience for instance, why not provide these services only over ipv6 for the beta phase?

    1. nematoad

      Re: Glorified proxy

      "...offer services that either require ipv6, or work better with it."

      I take your point, but when does such a decision morph from a carrot to a stick?

      Much like MS denying the latest and greatest to anyone not running Windows 10 I feel that this is a very tricky line to walk, and beating someone over the head just because you want to "move on" risks alienating the very people you need to have a business.

    2. Mike@home

      Re: Glorified proxy

      "What will spur ipv6 adoption is creating demand for it - offer services that either require ipv6, or work better with it. "

      I wonder what would happen if Google were to start giving lower rankings to sites which didn't provide an IPV6 address. They certainly hastened the transition from HTTP to HTTPS

  8. mark l 2 Silver badge

    My home broadband used GG-NAT and it can be a pain with sites that use you IP address to identify uses and for Geo-location, as all the public IP addresses my ISP uses are based in London but I am based 200 miles away.

    Also a couple of times I have found that the public IP address has been blacklisted when trying to post on forums etc. as it is shared with 100s or maybe even 1000s of other ISP customers if one of them starts spamming then it can block everyone else using that public gateway.

    1. TRT Silver badge

      Geo-loc fail.

      Apparently when I got my new iPhone and was putting in my Apple identity details, a new device was signing in to my iCloud account in Manchester. Which is a neat trick, as I was in London at the time.

    2. Chemist

      "and it can be a pain with sites that use you IP address to identify uses and for Geo-location"

      Well I've got a fixed IP address but usually Google Maps puts me in West Yorkshire or near Warwick - neither of these is true

      1. rnturn

        My fixed addresses's physical locations are humorously mis-identified when visiting certain stores' web sites as being in as being in the south suburbs of Chicago or even in Texas. My guess is an outdated database. (I'm actually in far northern Illinois.)

        1. onefang

          "My guess is an outdated database. (I'm actually in far northern Illinois.)"

          Generally the only people that know your IP address and your physical address is your ISP, since they supply the first, and need to supply your Internet service to the second. So we can blame them for GeoIP not being accurate a lot of the time, the source info comes from ISPs.

          My last home ISP just GeoIP placed my static IPs in their NOC in Sydney, which was good for privacy at least (I was no where near Sydney). I've only just moved home, and due to reasons had to switch ISPs, dunno yet where they put us, but I think they use CG-NAT, I may be living in a NOC again. For some odd reason, I haven't asked them why yet, my Amsterdam server has recently "moved" to Ukraine, it's GeoIP used to be accurate.

          OK, I just checked, my Amsterdam server is back in Netherlands, and my home is two suburbs away, instead of the capital city of some other state. Not sure what is out there, but likely a Telstra exchange, I know it's not the main Telstra NOC for this city, that's in the same suburb as my home.

    3. JohnFen

      "all the public IP addresses my ISP uses are based in London but I am based 200 miles away."

      I'm in the US, but my ISP has something similar going on. If you geolocate me by my IP address, you're going to be off by about 100 miles.

      However, I consider this to be a very, very good thing, not a bad thing.

    4. eldakka

      sites that use you IP address to identify uses and for Geo-location,

      Well those are 2 problems right there.

      IP addresses are neither identifiers or location beacons. And they never should be.

  9. Alan Brown Silver badge

    Ipv4 origins

    "When IPv4 was under development during the 1970s, it must have seemed reasonable to specify 32-bit source and destination fields that dictate approximately 4.3 billion possible addresses"

    Just because you can HAVE 4.3 billion IP addresses means you SHOULD do it.

    "After all, the entire internet at the time comprised a relatively small population of government and academic sites. Then came the World Wide Web in the 1990s, and global subscriber numbers went stratospheric."

    The first octet was supposed to be routing information, the second octet, routing information within the organisation. (a bit like country codes and area codes)

    That went out the window long before the WWW was an idea in Tim B's head, mainly because IPv4 was a "hacky kludge" with a projected lifespan of maybe 5 years whilst the real Internet Protocol was developed (Didn't you ever wonder what happened to IPv5?)

    A lot of the first few bytes of IPv6 is intended to provide exactly that routing detail. Yes it COULD provide quadzillions of individual IP addresses, but that's not the actual design purpose and focussing on that misses the entire fucking point (In fact, if you did start filling up the space, you'd wreck the easy routability in the same way that IPv4 routing tables have become a clusterfuck in the core levels)

    As others have said: CG-NAT is not "Internet", it's a window out of a walled garden.

    There is a possible way forward on this (for UK readers, if not YMMV): The Internet Engineering task Force declared IPv4 to be in official "sunset period" back in 2017 and that it is to be replaced by IPv6.

    This gives force to misleading advertising complaints to the ASA and your local Trading Standards office against any provider which claims to offer "Unlimited" access without IPv6 (that's a very real limitation right there - effectively a walled garden with decreasing horizons over time), or "Internet access" without IPv6 - again for the same reaon (Ipv6 is an essential part of the Internet, therefore not providing it means that what's on offer is no longer an Internet service, but some kind of cut down version.)

    Casting back to the mid 2000s when complaints were upheld about 3's Walled Garden web-only service being called "Internet", complaints like this might have legs - especially if lots of complaints are sent in.

    1. Doctor Syntax Silver badge

      Re: Ipv4 origins

      "A lot of the first few bytes of IPv6 is intended to provide exactly that routing detail."

      How does that work with portable devices?

      1. Nanashi

        Re: Ipv4 origins

        They get an IP from the network they're connected to. The prefix for the network is allocated from its parent allocation, and so on up to the root. All allocations are large and sparse so as to reduce routing fragmentation.

        Basically it's no different to v4 in method, it's just that the address space is large enough that aggregation can work.

        1. Charles 9

          Re: Ipv4 origins

          In addition, IINM there are special allocations set aside for setups that don't easily fit into a geographic or otherwise sane structure (like perhaps a pan-global company with multiple external gateways in multiple countries). IOW, part of the reason for 128 bits is to (unlike last time) try to take everything into consideration.

  10. Flywheel
    Black Helicopters

    Privacy implications

    Ignore my tinfoil hat, but if every Internet/network-capable device I possess has a unique IPV6 address, it's going to make it very difficult for me to deny any alleged wrongdoings when the Authorities gather my location metadata together.

    "Mr Flywheel, you claim you were in Manchester but your IP-toaster, toilet, and toothbrush all confirm you were in Neasden..."

    With my currently heavily-NAT'd setup I can at least relax safely in the knowledge that my location is often wildly inaccurate!

    1. Len
      Holmes

      Re: Privacy implications

      Fortunately some people thought about that and drafted RFC 4941, 'Privacy Extensions' https://tools.ietf.org/html/rfc4941

      This means that the externally visible addresses of your individual devices (the suffixes) are changing all the time. The only thing that isn't changing is the prefix your ISP gives you. That produces a similar privacy situation as with IPv4. Quite a few ISPs don't give out fixed IPv4 address or IPv6 prefixes so even that changes once in a while for many people (or when you restart your modem).

      One could even argue that, without a global database of the blocksize that each ISP on the planet hands out, it is impossible to know for certain which part of the IPv6 address is the prefix and which the suffix. Without knowing the size of the block that the ISP hands out you can't know for certain whether 2001:db8:85a3:0:0:8a2e:370:7334 and 2001:db8:85a3:0:206:9b1e:900:6231 are the same user or not. That makes an IPv6 address with Privacy Extensions a better guarantee of your privacy than an IPv4 address.

      1. Jellied Eel Silver badge

        Re: Privacy implications

        One could even argue that, without a global database of the blocksize that each ISP on the planet hands out,

        Theoretically that would be IANA, then delegations via the RIR. So for example-

        https://www.ripe.net/publications/docs/ripe-655#5

        Which explains RIPE's allocations and assignments policies, so a /29-/32 to LIRs and utilisation based on /56s.

      2. JohnFen

        Re: Privacy implications

        Those privacy extensions are better than nothing, but they're pretty poor.

    2. Nanashi

      Re: Privacy implications

      That's not how IPs work. You're thinking of MAC addresses, not IPs.

      When we say "unique v6 address", what we mean is that only one machine has that IP at a time. That doesn't mean that the IP is in any way tied to the machine, or that it can be used to identify which machine was using the IP. To do that, you'd need some sort of IP<->MAC mapping. Since MACs don't make it through routers, the only place that mapping can be made is on the router to your own network.

      I'm not sure why you mentioned NAT here, unless it was just simply because you misunderstood what NAT was doing for you. MAC addresses don't go through routers whether or not you're using NAT.

      1. Doctor Syntax Silver badge

        Re: Privacy implications

        "I'm not sure why you mentioned NAT here, unless it was just simply because you misunderstood what NAT was doing for you."

        I'd guess that what the OP was alluding to was that the NATed device can't be routed to so that anybody running round knocking on random IP addresses to see if they can find an open port won't be able to see such a device. If all addresses are routeable then that level of protection doesn't exist and you're left depending on the firewall; let's hope that that's properly configured at source and stays that way.

      2. JohnFen

        Re: Privacy implications

        "That doesn't mean that the IP is in any way tied to the machine, or that it can be used to identify which machine was using the IP"

        It does if you don't want to be constantly changing the IP address for the machine.

      3. eldakka

        Re: Privacy implications

        Since MACs don't make it through routers,

        Apart from those implementations that by default use the MAC address for the IPv6 suffix you mean.

      4. onefang

        Re: Privacy implications

        "To do that, you'd need some sort of IP<->MAC mapping. Since MACs don't make it through routers, the only place that mapping can be made is on the router to your own network."

        Or using some sort of protocol like that used for Second Life / OpenSim, where the login protocol includes your MAC and HD serial number, to make it easier for Second Life to ban you. Faking your MAC and HD serial number is a T&C offense in Second Life (if I remember correctly, been a long time since I've been there), OpenSim just replicated the protocol.

    3. JohnFen

      Re: Privacy implications

      "if every Internet/network-capable device I possess has a unique IPV6 address, it's going to make it very difficult for me to deny any alleged wrongdoings when the Authorities gather my location metadata together."

      More importantly, it makes it more difficult to evade tracking from the bad guys like Google, Facebook, etc.

  11. steelpillow Silver badge
    Childcatcher

    Big deal, huh

    At a technical level, IP 4 and 6 are incompatible. Yet on a functional level they are: both do the same job in allowing my web client to visit sites with both v4 and v6 addresses, and for those sites to serve to both kinds of client.

    It does not seem rocket science for a v6 configured NIC to run a parallel v4 emulator and direct traffic internally to the appropriate process. You know, ARM and x86 processors are incompatible but you can still emulate the one with the other and run its code, just do the same with IP. If such a need ever develops, I am sure someone will start meeting it. But at the moment, with NAT being so mature and useful, there is not yet a need. What v6 advocates really want is a killer upgrade that irretrievably borks NAT for all time. Good luck with that.

    Meanwhile, we have multiple processor architectures, we speak multiple languages ourselves, we fuel our vehicles with petrol, diesel or electric, we will just run two protocols on our networks. Big deal, huh.

    Icon for the children who will live with the reality.

    1. Gerhard Mack

      Re: Big deal, huh

      "It does not seem rocket science for a v6 configured NIC to run a parallel v4 emulator and direct traffic internally to the appropriate process. "

      We have that now, it's called dual stack and it's pretty much the default. Very few devices are IPv6 only for the time being and that's unlikely to change until IPv6 is supported by a strong majority of endpoints.

  12. Andy Mac

    I dread the day I ask someone for their IP and they give me an IPv6 address.

    1. Lee D Silver badge

      Why are you using IP addresses in that manner?

      I know precisely two IP addresses (and one mask) for my entire site.

      One internal (gateway and DNS). One external. Everything else is listed in the DNS and or literally doesn't matter so long as it's in the DHCP ranges (never make security access decisions based solely on membership of a subnet, it's trivially easy to discover the subnet and deploy a device with an unused address in that range).

      Why would any outside entity need to know the IP of an internal machine that doesn't already have access to the internal DNS server?

      Why would any internal entity need to know the IP of anything internal, they just query DNS.

      I literally keep a spreadsheet for a single range of statics, that I manage - and I couldn't tell you what any one of them related to without checking the spreadsheet. But for sure "ping servername" will give you the IP and check it's up in one move.

      The problem here is places referring to IPs, not that the format of IPs may be changing.

      Honestly, if worst came to the worst and DNS died and DHCP died, I could just give any machine any address in the subnet and it would work, or deploy literally any DHCP server over that subnet and have it work. I could run that from my mobile phone today (provided I remembered to put it into the switch management as an allowed DHCP server).

      Stop referring to IPs and suddenly IPv6 makes no difference to your working practice at all. What machine? OFFICEPC-0054. What gateway? The .1 of the range (or .254 or whatever you want to use). What server? SERVER1.domain.com. What UNC path? Hang them all off "storage.domain.com" using DFS or something.

      That people build systems where ANYTHING other than the gateway and primary DNS (and maybe secondary DNS) is on a fixed and well-known IP-address really worries me. Everything from IP migration to server replacement, to what happens if the DHCP lease file disappears tomorrow, all kinds of issues resolved by just referring to things by name.

      I literally couldn't tell you the IP of a printer, a network switch, a telephone (or even the telephony SIP master box), a PC, a server or anything off the top of my head. It literally doesn't matter. Much more important that they are on the right VLAN (e.g. telephony, printing, CCTV, access control, etc.) and know how to talk to other machines (e.g. default gateway) than anything else. I mean, there's a list somewhere for anything important, but for sure it wouldn't make an ounce of difference if I just picked any IP except gateway and DNS and just changed it to anything else because "I'd forgotten what that VM used to use". Everything would still authenticate where necessary, pick up settings allocated to it, be accessible to other machines, etc.

      1. eldakka

        Why would any outside entity need to know the IP of an internal machine that doesn't already have access to the internal DNS server?

        Why would any internal entity need to know the IP of anything internal, they just query DNS.

        Right, so to make IPv6 useable I must also:

        1) buy my own domain (or subdomain);

        2) set up and maintain my own DNS server;

        3) make it externally accessable?

        I thought IPv6 proponents say that IPv6 is easier than IPv4?

  13. Doctor Syntax Silver badge

    "When IPv4 was under development during the 1970s, it must have seemed reasonable to specify 32-bit source and destination fields that dictate approximately 4.3 billion possible addresses."

    And now that those 4.3 billion addresses are moe or less all allocated and, via NAT and CGNAT, over-allocated that's many billions of nodes, some with IPv4-only baked in via firmware which need to be migrated if you want an IPv6-only world. That's a seriously non-trivial task. And it's no good saying, as one commentard more or less did some time ago, if you don't understand hex you shouldn't be using it because most of the owners of those billions of nodes don't, never will and shouldn't need to.

    Unless a scheme had been devised to seamlessly - let me go full Bob here: SEAMLESSLY - migrate those it was always going to end up like this.

    Would it have been feasible to devise a protocol which accepted IPv4 as a fully accepted subset? I don't know, but if it would then anything else would have been a serious mistake.

    1. Alan Brown Silver badge

      "When IPv4 was under development during the 1970s, it must have seemed reasonable to specify 32-bit source and destination fields that dictate approximately 4.3 billion possible addresses."

      Actually, Vint Cerf was castigated for using 128-bit addressing and told to cut it down to 32-bits.

      640k ought to be enough for anyone.

      (Actually, as I've mentioned before, IPv4 was intended to be a temporary solution - which has lived more than 30 years past its use-by date, thanks to IPv5 being DOA.)

      "Would it have been feasible to devise a protocol which accepted IPv4 as a fully accepted subset?"

      No - and one of the driving reasons for making IPv6 address space so large is to ensure that needing to go through this exercise again in future is "unlikely"

    2. Nanashi

      Would it have been feasible to devise a protocol which accepted IPv4 as a fully accepted subset? I don't know, but if it would then anything else would have been a serious mistake.

      No, it wouldn't. There's no way to do what you're imagining. That's why we didn't do it.

      Note that it is possible to connect from v6 to v4 via NAT64, and there's a standard range for that (64:ff9b::0:0/96). For example, in the process of making this post, my browser is talking to 64:ff9b::104.18.227.129, since for the fun of it I removed v4 from my desktop and am relying purely on v6.

      Is that close enough to what you wanted?

      It's impossible to do a completely seamless transition, because v4 is just not designed in a way that supports that, and obviously you can't change v4 to support a seamless transition because that wouldn't be seamless. v6's design makes the transition as seamless as possible given the constraints it's working under.

      1. Doctor Syntax Silver badge

        "It's impossible to do a completely seamless transition, because v4 is just not designed in a way that supports that,"

        And given that there are several billion nodes out there on v4 don't be surprised that uptake of v6 isn't happening at all fast. As other comments have said, the time when the change should have been made was a long time ago. There are a huge number of people who have bought, or been sent by their ISP, a box, plugged it in and had everything just work because every single thing they have that connects to that box will use v4. Replace that new box for v6 and something will break - sod's law will ensure that. Nobody is going to want to handle all those support calls.

      2. Alan Brown Silver badge

        " in the process of making this post, my browser is talking to 64:ff9b::104.18.227.129"

        For the stick-in-the-muds, that can also be written as

        0.100.255.155.0.0.0.0.0.0.0.0.104.18.227.129 - but it's a bit bloody longwinded isn't it?

        THAT is why the new formatting was adopted.

    3. Chronos

      And it's no good saying, as one commentard more or less did some time ago, if you don't understand hex you shouldn't be using it because most of the owners of those billions of nodes don't, never will and shouldn't need to.

      That was, probably unsurprisingly, me. What I meant by that was that the internet protocol version should be utterly meaningless to the end user. If, on the other hand, you are in infrastructure, networking for a large organisation, router design or a hosting provider you really should have a grasp of hex words and how to split 128 bit addresses up by netmask. That wasn't a cue for everyone and her uncle Willie to go out and get CCNA qualified, it was a reflection on network people bleating about 128 bits being too hard when it really isn't, which is starting to sound like excuses.

      Apologies if that wasn't made clear at the time.

    4. david 12 Silver badge

      "4.3 billion addresses are moe or less all allocated"

      A quick look tells me that 1 billion addresses are more or less allocated. The other 3 billion are more or less not allocated.

      1. Nanashi

        Re: "4.3 billion addresses are moe or less all allocated"

        No, they're pretty much all allocated. You can see the high-level allocations at https://www.iana.org/assignments/ipv4-address-space/ipv4-address-space.xhtml; of the three quarters of the address space that's available for unicast allocations, all of it has been allocated either directly or to RIRs.

        Of the space managed by RIRs, the only remaining unallocated parts are summarized at https://ipv4.potaroo.net/, and they basically boil down to a total of about one /8, most of which is reserved for transitional purposes.

        1. Roland6 Silver badge

          Re: "4.3 billion addresses are moe or less all allocated"

          >No, they're pretty much all allocated.

          The question is to whom does this really matter?

          Fundamentally, I suggest the only people directly impacted by the near exhaustion of IPv4 addresses are those who need to obtain large numbers of IP addresses at low cost. which would seem to be mobile operators and cloud operators. Established ISPs with sufficient addresses to cover their existing customers and some growth aren't directly impacted.

          Interestingly, I suggest that IoT vendors can get away with things because DHCP moves the onus of having a spare IP address to allocate to their gadget on to the user. So what is going on currently, is a bit of market awareness and preparation: I expect in the next few years residential ISP's to offer IPv6 as a costed optional service enhancement - why should ISPs pay to provide a service others such as the IoT and home automation vendors will benefit from - the customer wants, the customer pays.

        2. david 12 Silver badge

          Re: "4.3 billion addresses are moe or less all allocated"

          You're not paying attention son.

          The other 3.3 billion more or less allocated addresses aren't actually allocated to anything. That is, there is no Thing they are allocated to. They are allocated to no Thing.

          1. Nanashi

            Re: "4.3 billion addresses are moe or less all allocated"

            There are no "other 3.3 billion addresses". Of the addresses available for unicast allocation, almost all have been allocated to somebody.

            Unless you're trying to say that there's a lot of addresses that haven't been assigned to a machine? Obviously there are, because that's how IPs work. Unlike MACs, they're assigned hierarchically in blocks of 2^n addresses. Of course that's going to lead to a lot of addresses not being assigned to a machine.

            Those "unused" addresses aren't actually unused, they're performing the very important role of making the internet possible at scale. They won't save you from v4 exhaustion.

  14. lsces

    Not even available ...

    Just received a modem from Vodafone for a second fibre broadband only to find it does not even understand IPv6 ... I did not even think to ask about it as the BT circuit already has it set up. I just assumed since the mobile phone has an IPv6 address that the broadband would as well ...

    It would perhaps be useful if we had an actual list of providers who do support an IPv6 connection?

  15. poohbear

    Specifics

    Okay, I'll bite:

    "340 undecillion"

    Can you please be more specific? :-)

    1. Anonymous Coward
      Anonymous Coward

      Re: Specifics

      Can you please be more specific

      2¹²⁸

  16. Warm Braw

    These organisations will no longer have great need for their IPv4 address

    There are a few organisations that have large allocations of private IPv4 address space they could be using for their internal networks, but they're a pretty small number compared to those using 10.x.x.x and 192.168.x.x addresses and they're not going to make a huge overall difference to IPv4 address availability now that all the regional registries have run out.

    Organisations that have smaller allocations of IPv4 addresses continue to need them if they want to communicate beyond the roughly 25% of the Internet that currently supports IPv6. As long as there are IPv4 hosts you might conceivably want to speak to, you have to retain an IPv4 address.

    You can devise schemes that permit IPv6 only and IPv4 only hosts to communicate but the argument has been they're unnecessary: the transition schemes all assumed that people would migrate to IPv6 before the IPv4 addresses ran out - and they still basically do (with NAT basically making the pool of IPv4 addresses go further). It's probably a bit late to revisit that now.

  17. simon_c

    When I started my IT career back in the mid 90s, the old, grizzled network engineer predicted he'd not need to worry about IPv6 as he'd be retured before it was needed.

    Now, 25+ years under by belt, I'm wondering if that wil be true for me too.....

  18. Pangasinan Philippines

    Unique Content

    'What will spur ipv6 adoption is creating demand for it - offer services that either require ipv6, or work better with it.'

    Just like the new fangled TVs in the '60s

    They had 625 lines instead of 405. But required purchase of new TV and aerial. How much???

    But BBC2 launched on 625 lines and had content that people wanted to watch.

    Joan Bakewell for example. Also Barry Norman with film (movie) reviews.

    Then colour came along and the nation responded with their wallets and credit cards.

    1. Herring`

      Re: Unique Content

      Well, the thing that struck me: what really drove demand for things like the Internet and VCRs was filth. When there is a pr0n application that required IPv6, then you will see things change.

      Mind you, I believe that there is far more grot on the internet than one person can ever get through. I'll update you with my progress on this as it happens.

      1. Anonymous Coward
        Anonymous Coward

        Re: Unique Content

        You have something there. Strictly regulate everything on IPv4 and allow a free and open internet on IPv6 and the people will migrate while IPv4 networks become childsafe.

        1. Herring`

          Re: Unique Content

          Well content filtering would be a lot harder with 2^128 addresses.

          We could be onto something here.

    2. Doctor Syntax Silver badge

      Re: Unique Content

      "They had 625 lines instead of 405. But required purchase of new TV and aerial."

      The killer there was that UHF sets were on sale for a long time before VHF was turned off. That meant that there was a long time when the broadcasters were supporting two systems. The switch-over to digital was handled by having digital to analogue adapters available and also dual-standard receivers so that the analogue only sets could die a natural death without being unusable.

      The TV analogy is also limited in application because on the whole in a household with multiple sets each one could go irretrievably TITSUP with out affecting the rest. But throw in a mixture of IPv4 printers in a domestic situation and the occasional box that absolutely has to run XP in a commercial situation and the TV analogy doesn't apply.

  19. DonL

    Mobile devices / 4G networks

    As far as I know of, none of the mobile networks support IPv6.

    Even if I just wanted to use IPv6 for my home network (to extremely simply the example), I wouldn't have any advantage of it just because of lack of support on 4G networks alone. Dual stack basically means double work and no advantages.

    I used to be very enthusiastic about IPv6 back in the days, but there are so many showstoppers it's not funny.

    In retrospect, it would've been better to just cram the additional adress space somewhere in some(rarely) used IPv4 field or something (with an prepend part and append part), so IPv4 carrier routers could just route it. Then support between the client and endpoint router would be sufficient to bring advantages. Then later, the prepended adress space could be used by the carriers when they're ready.

    Sure the protocol would look ugly on paper, but in the end it wouldn't really matter.

    1. Alan Brown Silver badge

      Re: Mobile devices / 4G networks

      "As far as I know of, none of the mobile networks support IPv6."

      Which is odd, because one of the driving forces of IPv6 design was mobile networks.

      1. Roland6 Silver badge

        Re: Mobile devices / 4G networks

        >Which is odd, because one of the driving forces of IPv6 design was mobile networks.

        And from my time working with 3GPP, circa 2000 it was also one of the drivers wanting the IPv6 working group to hurry up and deliver an agreed specification...

    2. Lee D Silver badge

      Re: Mobile devices / 4G networks

      I believe both 4G and DOCSIS have IPv6 support as a strict requirement.

      So they do have IPv6 (in fact, Google says 25% of their searches come from IPv6 IPs, most of them mobile carriers), and your phone and network has to support it. But maybe they aren't using it "by default".

      However, in the same vein, though DOCSIS specifies IPv6 as a requirement, Virgin Media still don't offer IPv6 to their customers.

    3. Anonymous Coward
      Anonymous Coward

      Re: Mobile devices / 4G networks

      > As far as I know of, none of the mobile networks support IPv6.

      I'm on EE any only have and IPv6 address on my mobile.

      They did a talk about it at UKNOF - https://www.youtube.com/watch?v=lKyuQ8mb_GE

      1. Roland6 Silver badge

        Re: Mobile devices / 4G networks

        An interesting contribution which shows that EE is serious about migrating their public internet services to IPv6 and doing so in a way that will be practically invisible to most of their users.

    4. Gerhard Mack

      Re: Mobile devices / 4G networks

      "In retrospect, it would've been better to just cram the additional adress space somewhere in some(rarely) used IPv4 field or something (with an prepend part and append part), so IPv4 carrier routers could just route it. Then support between the client and endpoint router would be sufficient to bring advantages. Then later, the prepended adress space could be used by the carriers when they're ready."

      And you are left with something that routes far slower than IPv4 since every router on the internet would need to have to check a flag followed by the prefix, and then the prefix and would still have all of the same compatibility issues of IPv6. In short: All of the disadvantages of IPv6 with none of the advantages.

      1. Phil O'Sophical Silver badge

        Re: Mobile devices / 4G networks

        And you are left with something that routes far slower than IPv4 since every router on the internet would need to have to check a flag followed by the prefix,

        Is that likely to be significantly different from having to check 128 bits for an address instead of 32?

        1. Gerhard Mack

          Re: Mobile devices / 4G networks

          "Is that likely to be significantly different from having to check 128 bits for an address instead of 32?"

          64 continuous bits for actual routing. In practice, IPv6 is sometimes faster than IPv4. Think of it from a CPU perspective: You are taking action based on a header flag (if x ,read y memory space), slowing down the CPU's ability to process the packet. That's not going to be as big of a deal on endpoints, but on core ISP routers, the timing difference would be noticeable.

    5. Anonymous Coward
      Anonymous Coward

      Re: Mobile devices / 4G networks

      Most actually do, at least in the US. AT&T does, Verizon Wireless does. Sprint doesn't seem to. I'm not sure about T-Mobile..

      1. Charles 9

        Re: Mobile devices / 4G networks

        Confirmed. I'm on T-Mobile and The IPv6 Test Site says it's native IPv6. And that was at least two years ago.

  20. Andraž 'ruskie' Levstik

    Asked my ISP when I'm getting IPv6 - "It's being worked on, we should be starting initial roll out next year."

    That was 5 years ago... I still don't have IPv6 from my ISP. And yes I do have a public IP with no CGNAT stuff infront.

    1. casperghst42

      I have a static IPv4 which blows the boat out of the water, as my IPS can't figure out how to allow my modem to have a dual stack ... my router (pfSense) can do it just fine.

    2. Alan Brown Silver badge

      > "It's being worked on, we should be starting initial roll out next year."

      > That was 5 years ago

      Ditto.

      Which is why I filed an ASA complaint about what they're calling Unlimited Broadband and Internet.

      There is ZERO mention of IPv6 on their website. I've pointed ou tthat some other ISPs are specifically saying they can't provide IPv6 whilst others (A&A) are pointing out that they can.

  21. phuzz Silver badge

    Article request

    Dear el'Reg

    You know what would be really helpful? If you rang up all of the main UK ISPs and asked them what they're doing about IPv6.

    The latest information I've heard about Virgin is this.

  22. Anonymous Coward
    Anonymous Coward

    These types of brow-beaty articles used to be fairly spaced out. Now we get two in two weeks?

    El Reg, you're venturing close to my troublesome user list and its consequences, whereby every repeated unreasonable demand extends the solution delivery time by three months. At this rate, I'll be migrating directly to IPv7.

    Jokes aside, let's make a deal, world + dog. I'll continue to engineer and administer my corporate networks in a manner that is consistent with business requirements and best practice security. I encourage everyone else to do the same.

  23. Anonymous Coward
    Anonymous Coward

    Like the alien said in Buckaroo Banzai: "It's a really bad design".

  24. LeahroyNake

    Take a SIP

    'such as those for IP telephony, will not operate because private IP addresses are not routable over the public internet, and therefore a node outside the network cannot initiate contact.'

    I wouldn't want any random spam phone call system phoning users ip phones without some kind of control.

    That is why you have a PBX and SIP trunks and / or PSTN or BRI gateway. It's not difficult.

  25. tentimes

    It's hideous for home users to set up on routers.

    I consider myself reasonably clued in, but when it came time to configure my TPLINK router (archer v900 v2) I found it really confusing. ISP's are not clear enough on how they use IPv6 which then makes it very difficult for me to configure the router. I eventually got an address system working through trial and error, though it is dynamically assigned behind the router when what I really wanted was some permanent IP addresses (IPv6) but I could not work out how to make that happen. So I have a static IPv4 address but I can't get a static IPv6 address, only a range, and I can't configure the range behind the firewall as I just can't get it working. The ISP has been cryptic and, frankly, I find the whole IPv6 scenario cryptic and obfuscated. If people want it adopted faster then they should explain it better.

    1. Gerhard Mack

      Re: It's hideous for home users to set up on routers.

      The first question you would need to ask your ISP is if the IPv6 prefix is static. After that, you can actually just assign the IP either on the router if you are using DHCPv6 or on the client machine if not. Static IPs are done the same way as IPv4, it's only the dynamically assigned addresses that are optionally different (Router Advertisement vs DHCPv6)

      1. tentimes

        Re: It's hideous for home users to set up on routers.

        Well I have easily set up reverese DNS for IPv4 but I can only get RADVD delegated to work on the router, which doesn't allow for static IP's and therefore I cannot assign a reverse DNS for my network. Ipv4 reverse DNS works perfectly. I have been trying for over a year but cannot figure it out. The IPv6 docs available are appalling and really hard to understand. I have two degrees, my first one being Electronic engineering. So I know how to read a spec.

        1. Gerhard Mack

          Re: It's hideous for home users to set up on routers.

          You are correct, RADVD does not support static IPs, however, you can set your IPv6 address directly on the endpoint as a static address the same way you would for IPv4.

  26. DCFusor

    Much talk about a committee

    That was made up of *vendors* who had no reason whatever to try for any kind of compatibility - they had reason to drive sales of new kit, though. No, they weren't stupid, but they do have rice bowls to keep full.

  27. Claptrap314 Silver badge

    It's not just FB

    Google has run out of 10.x.x.x internally by now. When they projected that they were 18 months out, they started the migration. When I was at IBM (early 2000's), we had 9.x.x.x & 10.x.x.x. I don't know if IBM had given 9 back by then or not...

  28. JohnFen

    Is that a problem, though?

    "the problem is that "inevitable" is starting to look a longer way off than expected"

    Why is that a problem?

  29. Anonymous Coward
    Anonymous Coward

    Trolling

    I'm starting to think that El Reg only runs these IPv6 stories as a way to troll for more clickbait :(.

  30. Reginald Onway
    FAIL

    If it works don't fix it...

    My ISP doesn't support IPV6. Instead IPV6 connections are routed through some kind of DNS conversion tables which SLOWS DOWN connections hugely.

    It's only been the last few years that any website in the world could do a simple IPV6 lookup. Try it sometime.

    And, I admit, the naming system is quite confusing to me.

    Firewalling IPV6 is hard and apps few and far between. It's exceptionally hard to filter OUTBOUND Ipv6. (To it's credit, the MS windows firewall does a pretty job at it.)

    However, in general IPV6 spam and crooks can blast right through today's router and firewall apps.

    Frankly my cyber life is better off without IPV6. Seems it's flawed solution for a non-problem, for most people and even tech coprs.

    1. Alan Brown Silver badge

      Re: If it works don't fix it...

      "Firewalling IPV6 is hard "

      In summary: "Bullshit"

      It's no harder than firewalling IPv4.

  31. Joseba4242

    IPv4 Forever

    This article, like others before, seems to misunderstand the incentives for IPv6 adoptions.

    If the number of IPv6-only services grows from 1 to 10 you might call it a tenfold increase, but overall it's still next to nothing. No reputable services will be IPv6-only unless practically all clients are IPv6 enabled. Please name just one significant service that's IPv6 only to support your claim of "more websites and online services will begin to only be available via IPv6".

    Some access and corporate networks have good reasons to enable IPv6. However there will always be a many that don't and that will only change if significant services were IPv6-only.

    Many IPv4-only services have little incentive to move to IPv4+IPv6 unless significant number of clients are IPv6-only. ElReg is a great example of this; even years of mockery of its user base were not sufficient and it continues to operate on IPv4-only just fine.

    Crucially a critical mass of adoption is not enough to break this stable cycle. You can have 95% of clients IPv6 enabled yet still need to provide service on IPv4. You can have 95% of services IPv6 enabled yet still need to provide IPv4 connectivity to your network.

    The only way would be to if either practically all access and corporate networks become IPv6 enabled, or significant services become IPv6 only. There is next to no chance of either of this happening.

  32. Geofb

    IPv6 Transition Technologies

    The UK IPv6 Council has organised a free half-day workshop on "IPv6 Transition Technologies", September 25 in London. A great chance to hear about transition technologies, and meet with council members and others who have wide experience of implementing IPv6 in educational establishments, business, and ISPs. Always friendly and eager to chat, share and learn. I've been attending their meetings since early 2015 and have learned much. Usually they retire to a local hostelry after the meeting proper, for more IPv6 chat and networking.

    Home page:

    http://www.ipv6.org.uk/

    Registration for the event (linked on home page):

    https://www.eventbrite.co.uk/e/ipv6-transition-technologies-workshop-tickets-48642326508

  33. John Sanders
    Mushroom

    "the world is clinging stubbornly to IPv4"

    This is because when IPv6 was first published it was DELIBERATELY made incompatible with IPv4 a little fact that was pointed out to the committee who came with IPv6 and not only ignored but smug at.

    It never ceases to amaze me to see that the IPv6 creators are the ones stubbornly rejecting reality and somehow if it is not adopted is because the rest of the internet can't see the brilliance of its godly design.

    IPv6 is and always have been garbage, it forces you to have a DNS and DHCP servers in place for it to work correctly, and does too many automatic things that that makes it a pain to deal with. And lest not forget the many, many times it has been changed, it has taken them 19 years to come with the final RFC spec.

    All that was needed was an extra 2 octets on the addressing, that's all. (This is roughly speaking for the technical pedants out there)

    Instead we got an over-engineered mess that is backwards incompatible and somehow it is everybody's fault is not adopted.

    Does anybody think that in December 1998 the people who designed it had the required expertise to do so*? I doubt it, the internet had only been running at large for three years at the time.

    (*I'm not talking about the bits and bolts, talking about what was that people and markets wanted)

    Add backwards compatibility to IPv6 and everyone jumps on board next day SIMPLE.

    1. Charles 9
      FAIL

      Re: "the world is clinging stubbornly to IPv4"

      "This is because when IPv6 was first published it was DELIBERATELY made incompatible with IPv4 a little fact that was pointed out to the committee who came with IPv6 and not only ignored but smug at."

      Do we have PROOF of this...or proof of the fact that because IPv4 is not forward-compatible (it was designed DECADES ago in an age when RAM and CPU power were at a premium) that if you're going to break the mold, might as well start fresh?

      "All that was needed was an extra 2 octets on the addressing, that's all."

      Riddle me this, Batman. How do you transport 18 eggs in an egg carton only build for 12 without breaking something? Because that's basically what you're proposing. You CANNOT add octets to IPv4 (which has a FIXED 32-bit address limt--it's in the spec) without breaking something. PROVE us wrong WITHOUT adding anything in between or changing likely-unchangable firmware.

    2. Nanashi

      Re: "the world is clinging stubbornly to IPv4"

      Yes, let's have an explanation please. If you think it's so simple to just add a few octets, tell us how to do it.

      There are good fundamental reasons why it's not possible to do, which I, Charles 9, and that "smug" "IPv6 committee" all understood. But if you've somehow worked out how we're wrong, then put your money where your mouth is and just tell us how! We'd really love to know!

      I've seen a lot of people say "just add some octets", and I've asked a fair few of them to explain how to go about doing that without ending up with something that has the same compatibility issues that v6 has, and nobody has managed to do so. Perhaps you will be the one to finally share the solution?

      (For the avoidance of doubt: whichever method you suggest has to actually work. This ought to be obvious, but something that doesn't work is not going to work.)

      1. Jellied Eel Silver badge

        Re: "the world is clinging stubbornly to IPv4"

        I'll have a go. And for simplicity, it assumes the problem we're trying to solve is increasing address space, not feature creep.

        So a v4 header starts with a 4-bit Version field. So that could signal '4' as it does now, '6' for v6 or something else. Then there's the IHL and other function fields, then 12 octets in, you get the source/destination bytes. So you could set '5' or '7' in the version field and expand the address space into the Options, because they're typically not used.

        For compatibility, devices would then just have to look at the version. If it's v4 in a v5 domain, address fields could be padded (eg 0.0.xxxx.xxxx.xxxx.xxxx) using a defined network to signal it's a v4 packet.

        And if people still think there's no simple solution.. Look at BGP. That went from 16 to 32bit ASNs starting in 2007 without breaking the Internet (much, because that's reserved for Kardashians).

        1. Nanashi

          Re: "the world is clinging stubbornly to IPv4"

          "Use the version field" is the approach v6 took. How does doing the same thing with a 5 instead of a 6 change anything?

          > For compatibility, devices would then just have to look at the version.

          How do you expect existing devices to know what to do with a version of 5? The whole problem is that existing devices don't know how to handle the new stuff.

          1. Jellied Eel Silver badge

            Re: "the world is clinging stubbornly to IPv4"

            Changing the version number just singles it's not plain'ol IPv4. For existing devices, response would be much the same if an IPv4 box recieved a v6 packet. If it's dual stack, it can process it, if it isn't.. It can be an interesting exercise to see how correctly an IP stack has been implemented. Or firewalls, which can be why simply blocking all ICMP is sometimes a bad thing.

            But again it's about deciding what the objective for reinventing the Internet is. Just expanding address space from 32 to 64bit is one thing, but IPv6 went so much further in it's 'design by committee' approach to include a slew of extra features, often unwanted and problematic. But existing v4 devices carry on working as normal because they see 4 in the version field. If they see 5, they could reply with an ICMP type 3 code 2 packet.. But that's really host/application behaviour.

            But it's also a very old proposal dating back to the 'Simple Internet Protocol' which was discussed here:-

            https://www.rfc-editor.org/rfc/rfc1454.txt

            SIP is simply IP with larger addresses and fewer options. Its main advantage is that it is even simpler that IPv4 to process.

            from 1993 where SIP and other proposals were compared because address space depletion was recognised very early on. Some years later, we ended up with v6 and a different SIP..

            1. Charles 9

              Re: "the world is clinging stubbornly to IPv4"

              Because ANOTHER problem hit: with the IPv4 address drought, routing tables got complicated, and no matter how much you try, complicated routing tables introduce latency, especially at the upper-tier routers that have to churn through a ton of this stuff every second. The reason IPv6 went so big was so it could go sparse...and simplify the routing problem WITH plenty of room to grow. Now instead of complicated lookup tables you can work with much simpler binary rules.

            2. Nanashi

              Re: "the world is clinging stubbornly to IPv4"

              Alright... so basically, what you're suggesting is no more backwards compatible than v6 is. Most of the rest of your post is just a rant about feature creep, which is itself interesting given that v6 didn't really add much in the way of new features.

              What new stuff is there in v6? Routing and subnetting work the same way as in v4. So do firewalls, and TCP/UDP/DNS if they count. IPsec? That was made optional, and it was backported to v4 anyway so apparently people wanted it. RAs? v4 got those in 1991. SLAAC is new, but SLAAC is super simple (and you can't really say "DHCP is ubiquitous, why not just use that, it's the obvious choice" because it wasn't in 1993. RFC 1531 itself is from late 1993, and there were tons of config protocols in common use for at least a few years after that.). NDP instead of ARP is new, but it does the same thing ARP does. It does go over multicast, but multicast isn't new either, and this allows removing broadcast which is the opposite of feature creep. Mobile IPv6? That came 8 years after Mobile IPv4. SEND? That's new, but nobody uses it. We did need a new socket type, and a new DNS API for handling multiple address families, but that was completely unavoidable. Any new address family would've needed that.

              And... I'm out of ideas. I'm sure I've missed something or another, but I'm not really seeing the creep. We could've added variable length addresses, or split the concept of an IP into separate location and identification IDs, or any number of other useful redesigns of L3, but we didn't. v6 is mostly just v4 with longer addresses.

  34. casperghst42

    Overly complicated

    I still remember when IPv4 started becoming popular back in the early/mid 90's, back then it was wooooo complicated - especially for the people who came from DEC, IPX or what ever was used back then (IPX over routers was basically crap), but I guess we all more or less mastered IPv4 reasonable fast.

    Over the years I've been trying to ger my head around IPv6, and I am still not able to find a IPv6 for Dummies which can explain me the basics in a reasonable easy and understandable way - I do this for fun, not for work...

    They have made IPv6 overly complicated, it's like you have to know more or less everything, before you start with the stuff, otherwise you'll end up having a network which might work, but you don't know why - and if you connect it to the internet you can't hide behind a NAT'd firewall - which is really really bad.

    And don't get me started on the notation, I've entered thousands of IPv4 address on my keyboard, on a numeric keyboard everything is there 0..9, and a '.' - everything which is needed. Try that with an IPv6 address ... right... see, got my point.

    I guess the universe is waiting for the next IP stack, which probaly will be an IPv16, which will be fully automatically discoverable, and have a functional (and understandable) security build in. Well I can always hope.

    IPv6 is there, but how many people know enough to implement it securily...

    1. Charles 9

      Re: Overly complicated

      "They have made IPv6 overly complicated, it's like you have to know more or less everything, before you start with the stuff, otherwise you'll end up having a network which might work, but you don't know why - and if you connect it to the internet you can't hide behind a NAT'd firewall - which is really really bad."

      Well...there's a problem right there. Why does it have to be a NAT'd firewall. Why can't it be a straight-up firewall?

      "And don't get me started on the notation, I've entered thousands of IPv4 address on my keyboard, on a numeric keyboard everything is there 0..9, and a '.' - everything which is needed."

      IPv6 addresses CAN be entered in decimal, too, WITH the standard period notation. That's why the bracket notation: it helps to distinguish which is which. Hex notation is simply easier to read for the longer addresses and allows for the double-colon shorthand, but the spec DOES accommodate.

  35. Tromos

    Adoption of IPv6 is inevitable.

    So is death, but I'm in no particular hurry.

    1. Charles 9

      Re: Adoption of IPv6 is inevitable.

      What about taxes? Sure, you'd like to hold them off, too, but it tends to be more trouble than it's worth most of the time.

  36. Donkey Molestor X

    guess who has fifteen psuedopodia and got 17/20 on ipv6-test.com?

    funny that folks are getting riled up about IPv6 again because i just got an IPv6 address on my home router after my last reboot. i'm getting 17/20 on ipv6-test.com. i can see regmedia.co.uk coming over IPv6 in IPvFoo. Guess I don't need my HE tunnel anymore. And my mobile phones have been getting IPv6 addresses over LTE for at least a year now.

    It's not all fun and games being behind CGNAT. Just wait until you get blacklisted from a service because of something that someone else did who shares or shared your public IPv4 address did. You won't be praising IPv4 after that, I'll bet.

    I got banned from 4chan once completely innocently because some other miscreant behind my public IPv4 was a bad boy/girl. And then I couldn't post memes and that made me sad. Yes, I'm too cheap to get a VPN.

  37. Sam Liddicott

    My printers "google-print" module mysteriously wouldn't work until I disabled ipv6 on the printer.

  38. guy_m

    Multicast

    "Other advantages of IPv6 are said to be support for multicast transmissions, allowing bandwidth-intensive traffic like multimedia streams to be sent to multiple destinations simultaneously rather than as multiple individual connections, thus saving on network bandwidth.”

    Multicast is also available when using ipv4, see RFC 1112. The subtlety is that there are no broadcasts in ipv6. Traditional broadcast functions from ipv4 such as Address Resolution Protocol (ARP) have equivalence in ipv6, using local pre-defined multicast groups. In the case of ARP that would be Neighbor Discovery Protocol (NDP). NDP is more efficient than ARP. For example, there is a multicast group for routers on a segment. To learn a router’s MAC address, a host need only send a query to that group. As most nodes will not subscribe to it, they will not receive traffic in which they have no interest.

  39. NiceCuppaTea

    Why does anyone care what going on in my internal network?

    I can use IPV4 on my internal networks then use a 6to4 enabled router at the gateway. This solves the Internets problem of me hoarding public facing ipv4's and also means I don't have a shit load of work to do testing all internal software!

    Let the internal network make its own way to IPV6 as and when the software running on it is ready. The gateway should be able to handle the transition.

    Same goes for my home network and that of the average Joe. 6to4 enabled routers and everyones happy. Cheap and nasty printers that dont support 6 still works as local network is 4 and 6, internet is happy because only 6 is presented to outside world.

    The only people that need to do any work from a consumer perspective are the ISP's delivering the correct spec / configuration routers.

    The only device that needs to be changed straight away from a corporate perspective is the gateway / edge systems. Everything else can wait. Don't see why everyone's getting their nickers in a twist about it.

    1. Nanashi

      Re: Why does anyone care what going on in my internal network?

      ...you do understand that 6to4 requires deploying v6 to your LAN, right?

  40. Anonymous Coward
    Anonymous Coward

    A view from bloggo towers

    We are doing upgrade to ipv6 at bloggo systems so we can sell our sizable ipv4 private address ranges while they still have some value.

    We have managed to sell them 57 times so far!

  41. David Crowe

    Businesses that switch to IPv6 are still dual-stack, so they can't give up IPv4 addresses. Basically IPv6 has NO ADVANTAGES until everyone converts. Which will never happen. So anybody who implements IPv6 is wasting time and money. What we need is IPv7. Admit that IPv6 was a failure and do it right (i.e. as an extension to IPv4 that gradually sucks the life out of its host). IPv6's independence from IPv4 was its fatal flaw.

    1. Nanashi

      That's an odd way to write "lots of advantages". It's not like you need to completely give up using v4 to use v6.

      >What we need is IPv7. Admit that IPv6 was a failure and do it right (i.e. as an extension to IPv4 that gradually sucks the life out of its host).

      You say this as if you think it were possible. It's not; there's no way to extend the amount of IPs in v4 without doing something that looks like v6.

      I'll ask you the same question I've asked other people: if you think it is, in fact, possible, then explain how to do it.

    2. Roland6 Silver badge

      Re: What we need is IPv7.

      The trouble is that this will suffer from all the same problems as IPv6.

      Yes, the IPv4 specification wasn't designed to futureproof, in the same way as say ISO OSI CLNS and it's inbuilt support for multiple address formats (both present and future) was.

      However, if the world had deployed OSI CLNS it would most probably have looked a lot like MAP/TOP/GOSIP, where only one address format was selected for use and thus implemented. Which would mean systems implementing this profile would not be able to communicate with systems implementing other address formats permitted in the full Standard. Ie. a CLNS implementation that only supported IPv4 addresses would be unable to directly communicate with a system using IPv6 addresses.

      To allow such systems to communicate would require the existing system to have a new address handling module installed ie. a networking software and/or firmware update. (I'm ignoring potential application issues here); something only devices in current support would receive. However, as we see with mobile phones, just because your device is new doesn't mean it will actually get updates. So there would be much uncertainty about whether such an update would actually result in a situation any better than where we are today with dual stacks, which actually makes things both a lot clearer and interop a lot easier.

      So, yes, IPv4/IPv6 interworking is a mess, however the real problem is the seeming lack of progress on a global public IPv6 Internet to encourage people to take the migration seriously.

      1. Charles 9

        Re: What we need is IPv7.

        "So, yes, IPv4/IPv6 interworking is a mess, however the real problem is the seeming lack of progress on a global public IPv6 Internet to encourage people to take the migration seriously."

        Problem is, a lot of people are in "If it ain't broke..." mode, and anything that's broke right now under the current regime, they probably don't need. Since people need an IPv4 address to stay visible on the Internet in general, people just keep trading in them. Who cares about the new market when everyone's still in the old one, IOW? Unless and until something essential seriously breaks stone cold dead, IPv4 will continue due to sheer momentum.

  42. AbeChen

    IPv4 Address Pool Has Been Expanded Significantly

    The main reason that IPv6 has not been rolling out smoothly is because it ignored the first rule of engineering in upgrading a working product / system, i.e., the backward compatibility to IPv4. Had it done so, the transition would have been completed a long time ago without even being notices. It is great that NAT came to the rescue by extending the service life of the Internet. First, the RG (Routing / Residential Gateway) version made the private networks possible. Then, the CG (Carrier Grade) version relieved the public address pool shortage, even though the extension may be finite. However, both of them have the basic restriction of not only limited choose of combination, but also allowing session setup only by their respective clients. The RG-NAT turns out to be desirable, while the CG-NAT prevents the Internet from providing the end-to-end connectivity that a full-fledged communication system should capable of.

    Our background in telephony enabled us to approach this Internet challenge from the knowledge of PSTN (Public Switched Telephone Network) that developed practices to expand the assignable telephone numbers through PABX (Private Automatic Branch eXchange) and less known CENTREX (CENTRal office EXchange) technologies.

    Instead of digging into the telephony details, we have submitted to IETF a proposal called EzIP (phonetic for Easy IPv4) about the solution from the networking perspectives:

    https://tools.ietf.org/html/draft-chen-ati-adaptive-ipv4-address-space-03

    Essentially, EzIP utilizes the very original IPv4 standard RFC791 and the long-reserved yet hardly-used 240/4 address block to expand each IPv4 public address by 255M (Million) fold. This is capable of serving an area with population up to about 39M which is larger than the largest city (Tokyo metro) and 75% of countries on earth. This capability not only enables governments, but also individuals to offer local sub-Internet services parallel to the current global version. These render IPv6 unnecessary.

    What relates to the incident article is that the RG-NAT will be preserved, while the CG-NAT will be utilized for the duration while EzIP-unaware (existing) IoTs are still in use. It should fade out once the subscribers recognize the benefits of the EzIP-capable IoTs. This general scheme will provide a smooth transition to a end-to-end connectivity Internet as originally envisioned when Internet was started with IPv4, yet without CG-NAT, nor IPv6.

    Thoughts and comments will be much appreciated.

    Abe (2018-08-28 18:53)

  43. bjorntoulouse

    When I was Fortinet literate NAT was the most vile invention - 1 address, 65000 users? Anything that makes that easier is fine by me.

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