[ppml] Policy Proposal: Definition of known ISP and changes to

[Leo Bicknell]

In a message written on Fri, Aug 31, 2007 at 01:34:18PM +0000, Paul Vixie wrote:
> the important observation is that ipv4 its start through a swamp and if we
> want ipv6 to get a start we might have to allow a swamp also.  remembering
> that the ipv4 swamp used to be a huge portion of the total ipv4 table, the
> most recent couple hundred thousand prefixes added have dwarfed the swamp.

Emphasis on the "might".

I've thought long and hard about the "recreate the swamp" issue.
At this point I think both sides of the coin have equal arguments.

In the "pro new swamp camp":

  - It gives early adopters a way to get in without having to get the
    big wheels of their company churning.  An individual or research
    group can get a small block to play with and not have to disturb
    marketing for a full roll out plan.
  - It provides a mechanism for the network to organically grow.
    Interesting parties will dual stack simply because they can.
  - At the end of the day it will end up being a small number of routes.

In the "con new swamp camp":

  - Organic growth is not the issue this time.  We're not trying to 
    create something new, but rather transition to something different.
  - Most businesses have been able to get a /32 for a few years for
    free and it hasn't made a difference, why will giving out /48's
    (/56's, /64's, whatever) make any difference?
  - We had the 6bone as a place for eary adopters to experiment.
    (Did it go away too soon?)
  - We're trying to do something better this time around than we did
    last time.  
  - Addresses are not part of the problem.

I also think there are two arguments that frequently get mixed up.
Several people have argued "anyone who has PI now should be able
to get PI in v6", and others have argued "we need to give out IPv6
to anyone who wants it."  The latter is "recreating the swamp" in
my opinion. The former does not have to have the same effect.
Somewhere between 25-50% of the "swamp" prefixes are not routed
today, depending on your view.  I think it's possible to give
everyone who has PI today PI in IPv6 without creating all of these
latent unused allocations which may pop up far in the future (e.g.
as may happen in IPv4, if they get value).

I think the last part in the con category is the most interesting
to me.  I'm not sure it's completely true, but there is some truth
about it.  There have been many long discussions on PPML of "my
router can't hold that many prefixes" or "we have no money to upgrade
gear to IPv6" or "I can't get the manpower to transition before
it's needed", etc.  All of these are business reasons people can't
get IPv6 out the door, and most of the people making these arguments
have IPv6 addressing!

-- 
       Leo Bicknell - bicknell at ufp.org - CCIE 3440
        PGP keys at http://www.ufp.org/~bicknell/
Read TMBG List - tmbg-list-request at tmbg.org, www.tmbg.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20070831/59dd22d0/attachment.sig>