Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×
The Internet Government The Almighty Buck United Kingdom IT Technology

Vint Cerf Calls For IPv6 Incentives In UK 164

sweetpea23 writes "Vint Cerf, the 'godfather' of the web and Internet evangelist for Google, has highlighted the need for cash incentives to encourage ISPs and businesses in the UK to move to version six of the IP addressing scheme (IPv6). In response to the UK government's stance that its role in the transition will primarily be advisory, Cerf suggested a system of tax credits for upgrading equipment to v6 capability — similar to the 'cash for clunkers' scheme in the US. 'You'd have to do the math to see what impact it would have, but creating some business incentive might be helpful,' he said. His words echo those of Axel Pawlik, managing director of the RIPE NCC, who warned last month that that the IT industry is adding unnecessary risk and complexity to Internet architectures by ignoring the availability of IPv6 addresses. the Internet authority IANA is expected to assign its last batch of IPv4 addresses in June 2011."
This discussion has been archived. No new comments can be posted.

Vint Cerf Calls For IPv6 Incentives In UK

Comments Filter:
  • Less carrot more stick from the government. Companies get too many benefits as it is.
    • by metamatic ( 202216 ) on Sunday November 14, 2010 @09:47AM (#34222064) Homepage Journal
      Lots of government procurement already depends on your product being IPv6 capable. The problem is really the ISPs. I'm ready to switch on IPv6 tomorrow, but my ISP doesn't support it, so I'm stuck tunneling.
      • I know I am not using IPv6 at the moment. How can I test whether it is my setup that fails, or my ISP that fails (or any other part)? Also, what advantages do I get _now_ (while everybody(?) is still also on IPv4) when I have full IPv6 support?
        • I heard there are some IPV6-only torrent trackers setting up. Makes sense to reduce the numbers to leechers. Found only one so far though. http://ipv6.torrent.ubuntu.com/ [ubuntu.com]
      • Anyone know if all the millions of home cable and DSL modems are going to be compatible with native IPV6?
        • by gentry ( 17384 )
          Not without firmware upgrades which are so far not forthcoming. I currently terminate the ADSL link via PPPoE to Linux server because I was unable to find a reasonably priced ADSL modem/router combo for home use that support IPv6.
        • Re: (Score:3, Informative)

          by KiloByte ( 825081 )

          Then they'll just use IPv4. We're not talking about single-stack IPv6 for now, and not for many years from now as well.

          There were several criminally broken models of home routers that blackholed AAAA DNS requests causing long timeouts, but they are basically the only technical obstacle to giving customers native dual stack, at least where the last mile is concerned. And those can get their firmware upgraded.

          • by gentry ( 17384 )
            I don't get this - the obstacle to native dual stack on end users network that the routers do not support IPv6 at all. They have no ability to get IPv6 addresses or route IPv6 packets. While the end user can use 6to4 and other methods that's not native. Firmware upgrades may be an option, or may not if the routers are already short of free RAM/ROM.
            • Uhm, no. You can't force a client to use IPv6, just like you can't force them to upgrade from Win98.

              This is about a few bastards with broken routers blocking the upgrade for the rest of us. ISPs don't want to add IPv6 support because those few people with broken routers would cause an outcry that "this ISP sucks, their Internet is broken".

              To move forward, there is no need to force IPv6 onto everyone -- those routers don't need to be upgraded to support IPv6, merely to not break down in the presence of IPv

              • by gentry ( 17384 )

                I see your point now and don't agree that issues with AAAA records is one of the reasons things move slowly. If that were the case they'd be seeing problems now - Google, for instance, deal out AAAA records and broken AAAA lookup would hamper requests to them where the client OS thought it should try the IPv6 route.

                On the whole migration to IPv6 should be transparent to the end user. The firmware on the (admittedly ISP controlled) router is upgrade to support IPv6, the router than starts to emit and respond

        • put simply, there aren't any.

          Well, thats not strictly true. I know the D-link DIR-625 [dlink.co.uk] supports it (and is advertised as such), but that's the only one I ever found that does. No Netgear device does or Belkin that I found, and they don't even recognise the search term on their web site, just to show how clued up they are on the subject.

          Oh, I should say that no online computer shop seems to have one of these for sale. Ho hum.

          • by smale8 ( 1940262 )

            Also the DIR-825.

            And the DIR-615.

            And the Airport Extreme and Time Capsule (no PPPoE with these though).

            I don't live in the U.K., but I had no trouble finding any of these for sale in my area.

            (There's even a Wikipedia page [wikipedia.org].)

          • Linksys' newer routers do IPv6. There isn't a UI for it but if you plug it in it'll work.

    • Re: (Score:3, Insightful)

      by Bert64 ( 520050 )

      They should simply mandate that anyone providing internet connectivity (ie any isp or telco) MUST provide ipv6, either alongside or instead of ipv4.
      If every end user and every site they try to visit is dual stack, a lot of traffic will occur over ipv6 without users even realising it and ipv4 will gradually die out.

    • I think all you need is to distribute some goodies on IPV6-only sites.
    • Re: (Score:1, Informative)

      by Anonymous Coward

      Less carrot more stick from the government. Companies get too many benefits as it is.

      There's going to be plenty of stick (and other pain) once IANA runs out of IPv4 addresses, and gives the last blocks to the RIRs. If your organization (especially those of medium and large sizes) doesn't have at least a basic test bed for IPv6 connectivity you're going to be in a world of hurt. At the very least a few of your NetOps folks should be playing with IPv6 in their "spare time" to understand how it works.

      Just this week I purchased an Apple AirPort Extreme which has IPv6 functionality built-in, and

      • The parallel/black market for IPV4 addresses is already growing. In practice the price of these will start going up. Existing companies/services start running out of IPV4 addresses to hand out to clients of services with strong growth. New projects/companies start having trouble getting blocks of addresses large enough to attend all clients. Some resort to black market, some to NAT.
  • How about they switch over to IPv6 and then lease their existing v4 blocks to the highest bidder?
  • Dear citizens of UK
    to speed up IPv6 adoption we created backwarder tax.
    You pay 50 pounds per year for every IPv4 address you use that year.
    Payment is by wire order to account 2349564322/3432
    • This method would actually have worked if the people who made IPv6 hadn't decided to make the standard backwards incompatible.

      As it is, IPv6 must be run in dual-stack mode, which means that even if you've got an IPv6 address, you must also have an IPv4 address. The reason for this is that the people who designed IPv6 are effectively incompetent. They designed a standard yes, but gave no thought to a transition plan.

      • As it is, IPv6 must be run in dual-stack mode

        That was true a few years ago, but now we have DS-Lite so you can free up those IPv4 addresses.

      • by paul248 ( 536459 )

        This method would actually have worked if the people who made IPv6 hadn't decided to make the standard backwards incompatible.

        How would you propose to make IPv6 backwards compatible? How can a device that only understands 32-bit addresses send a packet to a 128-bit address, and what problem would it solve that NAT64 doesn't?

        As it is, IPv6 must be run in dual-stack mode, which means that even if you've got an IPv6 address, you must also have an IPv4 address.

        T-Mobile disagrees. They're deploying an IPv6-only NAT64 network as we speak:
        http://groups.google.com/group/tmoipv6beta/web/t-mobiles-technical-architecture [google.com]

  • A flock of sheep will begin moving as soon as they see the 'leader' move; the leader is often just whichever one decides to move first because it runs out of food underneath itself then sees some someplace else.

    What we have with IPv6 is a bunch of fat lazy sheep who decide they will get off their butts once they see the rest of the flock someplace else.
    • by houstonbofh ( 602064 ) on Sunday November 14, 2010 @10:26AM (#34222342)
      Or, a bunch of overworked networked admins that don't want to start a big project for a problem they are not feeling yet... Switching over my office would take a lot of time I do not have for absolutely no payoff. And my 10-12 hour days are full already.
      • Switching over my office would take a lot of time

        Really? Took me about 20 minutes to configure a tunnel to HE.net. Every host on the LAN that supports IPv6 automatically started using it, to the point that Windows machines update their hostnames on the domain controller to resolve via both IPv4 and IPv6. Everything Just Worked without any manual intervention. What sort of monumental problems are you anticipating?

    • by gmack ( 197796 )

      The leader in this case being Microsoft. Windows XP only partially supports IPv6 with some test drivers installed. Vista was the first actual IPv6 ready OS from Microsoft so before that there was a big question of: what is the point of going through the trouble of being IPv6 ready when 99% of your customers couldn't use it?

      I've seen a lot more action now that Win7's numbers are moving up.

  • by RPoet ( 20693 ) on Sunday November 14, 2010 @09:39AM (#34222010) Journal

    I wonder if Berners-Lee cringes when he sees Cerf described as "the Godfather of the web" :-)

    • by Gruturo ( 141223 )

      Yeah, probably he cringes at the confusion between "Web" and "Internet" when people report that.

    • Re: (Score:3, Insightful)

      by TheRaven64 ( 641858 )
      Probably not as much as Cerf cringes when he reads TBL described as 'the father of the Internet'.
  • by reiisi ( 1211052 ) on Sunday November 14, 2010 @09:44AM (#34222054) Homepage

    I still think they should have solved this in the early '90s by switching to a byte-extensible addressing scheme.

    Something like defining x.x.x.1-127 as four byte and x.x.x.128-250.y, y 128 as five byte, and so forth.

    • Bit by the markup.

    • Re: (Score:3, Informative)

      Er no, IPv4 headers have space for exactly 4 bytes of destination address information. You might be able to kludge the protocol to allow for a larger address space, but as a kludge it would be inefficient, and encountering extended packets would break the majority of existing IPv4 stacks. The solution was arrived at by some very smart people, and that's IPv6. We won't run out of addresses on IPv6 for a very long time indeed.

      • It's unlikely that we ever will run out of IPv6 addresses as there's enough addresses to give each person living on Earth roughly 5×10^28 addresses. Which is quite likely going to be enough for anybody that could possibly follow in the future. So, it's technically possible, but for reasons related to the speed of light, physical size of the Earth and solar system it would be very difficult to ever get to the point where you need an IPv7.

        At that point it would be more of a problem actually getting th
        • by GvG ( 776789 )
          My ISP gave me a /48 address block for my ADSL connection. Which means I, as a single person, have 2^80 (10^24 if I converted correctly) addresses at my disposal. Now, obviously this is still far less than your 5x10^28 number, but it is a hell of a lot closer than you would think on first sight. My point being, it's nice to have this humongous 128 bit address space, but if we're going to waste 99.9999999999999999999999% of it we'll run out sooner than you'd expect.
        • by paul248 ( 536459 )

          The actual number of addresses is a red herring, since every user will be assigned between 2^64 and 2^80 addresses.

          For all practical purposes, IPv6 supports about 64K times as many users, but each user can have an unlimited number of devices.

          • 68k times assuming everything gets a /48 (2^80 addresses), which is only for now AFAICT so that routing tables don't become ginormous and cause systems to break down and cry.

            Also, it will be possible to reclaim space and shove stuff onto smaller subnets as needed, as we won't have legacy blocks, class E, etc. taking up large sections of the address space.

      • We won't run out of addresses on IPv6 for a very long time indeed.

        Nobody will ever need more than 640k.

        • What about those who need 640k + 1? Maybe it's time to look to IPv7.

          In all seriousness, I don't know how they're going to deal with people with legacy devices, which may not be upgradable to IPv6 compatible.

      • Some very smart people produced the OSI networking stack doesn't mean it was the product that won out
    • Oversimplifiying... (Score:5, Interesting)

      by Junta ( 36770 ) on Sunday November 14, 2010 @10:05AM (#34222190)

      For one, the protocol defined address specifically as 32-bit. Functions processing IPv4 generally use unsigned integers for the address. Functions to do a variable length address would take longer to process/route. The routing tables would likely be atrocious even if it theoretically could work.

      In short, it only does better at backwards compatibility at the extremely superficial aspect of entering addresses textually looking more usual and making a more specific effort for an existing IP to map trivially to a new scheme. Existing IPv4 stacks would have had no easier time trying to talk to 192.168.2.250.2 than fd7e:691a:da42::1. Besides, having the high values magically become reserved on the host portion of existing networks would conflict with existing host addresses in use.

      IPv6 can work but has been subject to three major pitfals:
      -It looks scarily different. People treating addresses like phone-numbers and not doing DNS in a ubiquitous has exacerbated the problem.
      -They completely omitted a strategy for v4-only to v6-only communication until this year. For a long time they didn't want to endorse anything with the letters 'NAT' in them and delayed a sane interop strategy hoping the problem would magically disappear so the 'evil' NAT wouldn't become a pillar of v6. I'm optimistic that the results of this year paves the way for meaningful progress.
      -v6 and associated protocol largely chose to throw the baby out with the bathwater on many fronts. v6 for a long time declared DHCP dead, then when DHCP was revived for v4, they threw out the existing behavior and started from scratch, eliminating many option codes and changing client identifier behavior to be hard for existing DHCP admins to deal with. This has in some cases rendered workflows in IPv4 simply impossible and in many more exacerbates the first problem in that a *lot* of relearning and reworking is required to acheive the same results with IPv6 as in IPv4.

  • So I navigate around IPV6 sites. What is there different to see or do? Not much. Some IPV6 brownie points, carrots, or something is needed. Sixxs [sixxs.net] installed rather easily on Ubuntu, just had to issue a command line. The magical-gui installer almost did it though.
  • I'm sick of all the noise about IPv6. ISPs already have monetary incentives to switch to IPv6: If they don't adopt it, eventually they will fall behind their competitors and maybe run out of bussiness. Governments do not need to create a "bussines incentive" giving away even more money for free just to encourage bussiness do what they should be doing with their own money anyway. It's not like these companies are like the financial sector, which can bring down the economy when it fails. The IPv6 bussiness in

    • by mpe ( 36238 )
      Governments do not need to create a "bussines incentive" giving away even more money for free just to encourage bussiness do what they should be doing with their own money anyway.

      All too often what happens is that businesses either adapt or come into the industry to simply take the money and not deliver the goods. Governments tend to be better at handing out money than they are attaching enforceable conditions to its use.
  • Why all the fuss about getting the government involved and giving them handouts?

    If you don't switch to ipv6 soon, then your clients will either be disconnected from the internet, or unable to connect to ipv6 compliant websites. Do we NEED any more reason? You're going to piss off customers = loss of profits. Make the damn switch already.
  • by ckdake ( 577698 )

    Great, just what we need. More corn subsidies.

  • Why do they need incentives?

    How about the incentive not to lose their connection to the internet?

    Why does everyone 'deserve' something for conforming to a technology standard? Shit or get off the pot.

  • So what he is saying is that a country that is taking extreme austerity measures should find monies to pay businesses to do what they will have to do anyway. I suspect that google is going to get some of this feee money, n'est pas?

    I suspect that the UK has enough of free market so that if the established companies can't provide the service, others will step in. I also suspect that if established companies can't provide the services, it may very well be cheaper to nationalize them, pay for upgrades, and

  • Cash for clunkers scheme in the UK was part of a near-global initiative to save the auto industry at a time when it's collapse would have been it's most damaging to the economy and society - during a global economic crisis. The scheme did cost quite a lot of money, but did save industry and jobs that were viable in the long term*. It also helped move people towards greener and more fuel-efficient cars. Furthermore the money was basically going straight to consumers/tax payers.

    In TFA no such arguments are g

    • by mickwd ( 196449 )

      Sadly, the cash for clunkers scheme also meant that there are many thousands of perfectly good cars parked in fields that must legally be scrapped (i.e. they can't be resold). Technology does improve, but fuel-efficiency hasn't improved that much since 1999/2000.

      The money wasn't just going straight to taxpayers - the whole point of the scheme was to help the car industry through a severe drop in demand. It basically kept the flow of cash going to (almost wholly foreign-owned) car companies. That said, it pr

  • Does he have a daughter getting married soon? I have a wish I'd like granted.
  • So are we going to be drilling holes in the old routers to make sure they are never used again?

  • first, the government will offer limited money on a limited-time basis for this. the corporations will have the paperwork in months before it occurs. those companies - corporations and telcos, mostly - will use up said funds.

    then, the corporations and telcos will 'offer' consumers the opportunity to upgrade 'ahead of the curve' once their own infrastructure is on ipv6. they will, of course, 'pass the cost on to customers'.

    the smaller shops - the ones which don't qualify for the government assistance, don't

  • The problem is just not with ISP. It is the web at large that is not IPv6 ready and not connecting to the IPv6 network.

    Here is a good example.

    ping6 -c 4 slashdot.org
    unknown host

    In this case as so many. The motivator for IPv6 switch over is only going to happen after everything goes to hell IPv4 wise.

  • If somebody wants to get new IPv4 addresses and can't route IPv6 over at least 75% (95%?) of their network, they should be denied unless they can document a really really good reason why implementing IPv6 isn't appropriate.

    That, at least, should get ISPs looking really hard at what isn't IP6 capable at this late a point in time.

  • by Mathieu Lu ( 69 ) * on Sunday November 14, 2010 @05:13PM (#34225646) Homepage

    I always had a hard time understanding IPv6 until I read the Running IPv6 in practice [debian-adm...ration.org] howto on Debian-administration and tried it at home. The next move is configuring the office where I work to use such a tunnel, then a friend's colo server, then our hosting environment. It's really not hard. Get over the adressing scheme. IPv6 is much easier to manage than NAT.

    Tunnelbroker [tunnelbroker.net] by Hurricane-Electric also does a great job of making IPv6 easy to use and fun to learn (the "certification" games). They also throw in free DNS hosting, and announcing IPv6 addresses using BGP is possible.

    Now stop whining and bite the bullet :-)

  • by rs79 ( 71822 )

    "Godfather of the web" ? (facepalm)

    As for V6... I dunno. There's a non-zero chance something else will pop up and get used, we're in our second decade o this turkey and the only poeple that believe in it are the ones with a vested interest in its success.

    If noting else we can use the multicast space. There's a gazillion V4 addresses "reserved by IANA" that are never going to be used for multicast and can be recycled.

The computer is to the information industry roughly what the central power station is to the electrical industry. -- Peter Drucker

Working...