[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RRG] PI as a non-requirement



Ran,

I like how you did not separate service providers from large enterprises.  That can be taken further.  I would characterize our preferences differently.  All other things being equal, we all (enterprise, SP, etc) have the following preferences:
  1. Don't renumber at all, and incur no costs.
  2. Renumber exactly one more time, and incur a one time charge.
  3. Invest in tools and processes that make renumbering easier, and incur those charges and whatever associated with the recurring event.
Let's talk about (2) for just a minute more.  Many organizational IT managers believe that there is just such a renumbering in their future and it is IPv6.  Hence any solution that requires a site to renumber may well bind itself to that event.  Worse (but far less likely), if IT managers as a whole do believe that there is some new routing architecture down the road that will solve problems relevant to their business and require renumbering, they may further delay their own migration to IPv6.

Your email was more about (3), only we haven't really discussed who pays for it.  In the end the end user will pay, and the costs are not small, and neither the tools are processes are simple.[1]  The question is simply which will cost more, and will those costs be visible in some fashion to those people.  CIDR was implemented and deployed with the tacit approval of end user organizations because nobody saw an alternative, even though we (at the time) understood that there would be costs associated with renumbering (PIER, et. al).  If there is an alternative and it is visible, and the costs are clear, then the end users will have more of a say this time around.

This is not to say that one should oppose (3).  Renumbering occurs for many different reasons, not all of which have to do with this argument (I'd argue in fact that most don't).  Rather we should all probably pursue (1), (2), and (3).

Finally, as I have mentioned previously, quite a lot of work has already gone into looking at the renumbering problem for IPv6.  Again, please see RFC 4192.  Beyond that perhaps it would be useful for people to carefully listen to what Tim Chown has to say, because he did EU-funded work on this very subject.[2]

Eliot

[1]  A nod to Bill Manning, who desired a world whereby one could renumber large swathes of network in five minutes.  It's a great ideal we should STILL continue to pursue.
[2] See http://www.6net.org/publications/deliverables/D3.6.1.pdf and http://www.6net.org/publications/deliverables/D3.6.2.pdf.  This work was jointly funded by the EU and Cisco.  See also http://www.mail-archive.com/ipv6@ietf.org/msg07648.html and related messages.