Re: [GLOBAL-V6] ... but important!

  • To: ipv6-wg at ripe dot net, "Wilfried Woeber, UniVie/ACOnet" <woeber at cc.univie dot ac dot at>
  • Subject: Re: [GLOBAL-V6] ... but important!
  • From: turchanyi.geza at ln dot matav dot hu
  • Date: Wed, 13 Feb 2002 19:10:26 +0100
  • Cc: global-v6 at lists dot apnic dot net
  • Sender:
      I would like to reply to your comments first, then add a few words to the 2000
      IPv6 (/32) LIR / Regions concept.
      1, 6NET
      Well, when 6NET had been designed there were no router vendor supporting IPv6 in
      the production release of their big enough router -- this was what I wanted to
      point out.
      On the other hand, now you would able to support IPv6 in the Geant core, because
      the Juniper boxes are supporting the basic functions needed since last November.
      Of course, as you pointed out, the Geant core management is not IPv6 based, and
      testing a management system on a production network would be too risky.
      This is what I said: there are many networking element of IPv6 only networking
      that is too risky to use in a production network.
      Probably there are still enough free bandwidh in the Geant core, therefore I do
      not realy understand your second argument. If 6NET just wanna to test IPv6 only
      features, a Layer-2 Ethernet VPN could have been created on the Juniper based
      MPLS core.
      Anyhow, I expect, that IPv6-only backbone service will be possible this year.
      What I do not expect, that it will be possible to convert a bigger percent of
      the existing Ipv4 sites to IPv6 only sites sooner than two years.
      There is a highly innovative and adaptive environment: the University of Vienna.
      Please tell me, when do you expect to get rid of the last IPv4 address used in
      your campus? Is it within two, or five years, or more?
      ...... Number of IPv6 LIRs
      If IPv4 networking should be supported for longer time, then why it is so urgent
      to speak about massive number of IPv6-only registry?
      This is what I can not understand.
      My first priority is the following:
      We should invent an IPv6 address allocation mechanism that allows to allocate
      IPv6 addresses by the LIRs already having experiences in address allocation and
      serious enough in networking;
      ... and only the second priority could be to invent a mechanism for establishing
      brand new LIRs jumping into the IPv6 address allocation business.
      Lets assume that 80% of the already experienced LIR could be qualified for IPv6
      allocation within two years, and an other 10% will get the needed experience in
      Counting with the estimation of Gert, that allowing 2000 IPv6 LIR per Region is
      not too risky, I do not see more place for those brand new IPv6 LIRs (not going
      through the normal IPv4 procedures), than 200 per Regions.
      Is there anybody who expect more than 10% for the "IPv6 only" ISPs compared to
      the "IPv4 and IPv6" in two years?
      Best regards,
      "Wilfried Woeber, UniVie/ACOnet" <woeber at cc.univie dot ac dot at> dátum: 2002.02.12
      Címzett:       global-v6 at lists dot apnic dot net
      Másolat:       woeber at cc.univie dot ac dot at (Vakmásolat: Turchányi Géza/PKI/HTC2)
      Tárgy:         [GLOBAL-V6] Correction, OT, but important!
        I haven't double-checked who actually submitted the folowing statement
      => 2, You might heard about big research initiatives for the introduction of
      => service. A good example is 6net.
      => 6NET will have initially its own infrastructure, and there is a good reason
      => behind: it would be too risky to use a production network.
        This is simply *wrong*!
        The reason for 6NET having a dedicated set of lines *initially* is
        - to show that a high-speed *native* v6-based network can be built,
          used and managed with *today's* technology and products (or to find
          out what is missing), and
        - the fact that the existing core of high-speed links, owned by and
          connecting European NRENS at speeds of 2.5..10Gbit (aka GEANT),
          can not easily (=cheaply ;-) be cut up into separate native paths
        Wilfried. (for 6NET Workpackage3: Basic Network Services)
        Wilfried Woeber                 : e-mail:
        UniVie Computer Center - ACOnet : Tel: +43 1 4277 - 140 33
        Universitaetsstrasse 7          : Fax: +43 1 4277 - 9 140
        A-1010 Vienna, Austria, Europe  : RIPE-DB: WW144, PGP keyID 0xF0ACB369
      - This list (global-v6) is handled by majordomo at lists dot apnic dot net