Re: [v6ops] I-D Action: draft-ietf-v6ops-ula-usage-recommendations-02.txt

[David Farmer <[e-mail]>]

On 2/19/14, 19:46 , Lorenzo Colitti wrote:
> On Thu, Feb 20, 2014 at 10:35 AM, Mark Andrews <[e-mail]
> <mailto:[e-mail]>> wrote:
>
>      > How are they punished? And what will they do when they are
>     punished? I
>      > guarantee that renumbering will not be the solution, NAT will be the
>      > solution.
>
>     Both sites just generate a NEW ULA prefixes.  They can continue to
>     use the collision prefix until the heat death of the universe if
>     they wish to or they can migrate to the new prefix.  IPv6 is NOT
>     IPv4.  Running parallel prefixes is *standard* proceedure.  They
>     just need to talk to each other using the NEW prefixes which should
>     be no more complicated than pushing new address selection tables.
>
>
> No, sorry. One of the main reasons people are advocating ULAs here is
> "because we'll have stable space and will never have to renumber!!11".
> Once you buy into that mantra, you'll be hardcoding IP addresses into
> configuration again, exactly like we do in IPv4 today. And exactly like
> in IPv4 today, renumbering will be prohibitively expensive.

I wouldn't say networks using ULAs will never need to be renumbered. 
But, they provide address stability independent of provider based GUA. 
You could get unlucky, you can screw up, but even if you don't, you will 
probably want to eventually consolidate your ULA prefixes just for 
simplicity of managing your networks.  If you acquire enough other 
companies there will eventually be a big enough win to justify the work 
of consolidating your ULA prefixes.

> As for multiple ULA prefixes... again, I think you're ignoring the
> realities of corporate IT staffing, corporate IT systems, and vendor
> capabilities.
>
> I think it's obvious that the path of least resistance (and thus, the
> solution that most admins would choose) will be NAT/NPT. After all, if
> you want to use ULAs to talk to the outside world (And why wouldn't you,
> right? It's what we do in IPv4, right?), you have to do NAT or NPT anyway.

The evolution to mobile devices and their transitory nature are
chaining the dynamics of most client network.  Fewer and fewer devices 
have even a semi-permanent relationships with an individual IP address. 
  Mobility and the consumerization of IT is driving more and more 
dynamic configuration of devices loosening the bonds between most 
devices and their IP addresses.

Further, the evolution toward cloud computing, elastic services models, 
and deployment orchestration are causing similar changes in server 
networks and data centers.  Old school corporate IT models are by no 
means dead.  But things are changing, you can't swing a dead cat without 
hitting a CIO who is talking about BYOD or Cloud Computing.

Renumber is never going to happen without some pain, but between these 
fundamental changes in the way we use networks and the capabilities of 
IPv6, renumbering is much more doable.

-- 
================================================
David Farmer               Email: [e-mail]
Office of Information Technology
University of Minnesota
2218 University Ave SE     Phone: [phone]
Minneapolis, MN 55414-3029  Cell: 1-612-812-9952
================================================