[address-policy-wg] proposal 2011-04, proposal 2011-05, concept document: IPv6 PA/PI unification, intransparent ip-eaters

Hi!

I see there's broad support for the 2011-05 proposal - by an overwhelming majority of IXP people...
There also seems to be broad support for the 2011-04 proposal - I can't really say for sure, but probably by an overwhelming majority of 6RD people...

So my 2 cent as some non-IXP non-6RD guy:

- I think IXPs should get the IPs they need. That is, just like everyone else. No special treatment. So in 2011-05's sense: 'no'.
- I think people who want to use 6RD should get the IPs they need (that includes space for a reasonable 'decent setup' not just the 'minimum space to make it work'). That is, just like everyone else. No special treatment. As 2011-04 doesn't treat 6RD special, that would be ok in this sense - but also superfluous. If $organisation submits documentation reasonably justifying the need for a /29 (like "we do 6RD"), I don't see where the current version of 5.1.2 doesn't already allow what 2011-04 wants to achieve. That's why I'd say 'no' to 2011-04 while at the same time I'd appreciate giving /29 to people who need it for their 6RD setup. Also, I believe 5.2 already allows the same (/29 for 6RD), and I see no need to add a specific paragraph under 5.1 (while I think there might be potential (but not necessarily a need, imho) to make 5.2 more defined and concise).
- I am totally a fan of KISS (errm - and btw I expand that to "...small and simple", not "...simple, stupid" ;), and this explicitly and specifically includes policies. I'd therefore generally abstain from putting more specific cases or redundancy than absolutely necessary into policy.

In the thread up to now, I didn't see anything that made it clear to me why making IXPs (and/or 6RD) special cases is necessary (or even legitimate).
There have been arguments as to lots of IXPs are noncommercial, commercial and noncommercial IXPs cannot be distinguished from each other, and they're all 'good', so they all should be special. I think that's wrong in every way: most IXPs are commercial, it's easy to identify whether an IXP is commercial or noncommercial (on the given examples to illustrate how difficult it is: these three were clearly all commercial), and 'IXPs are good for everybody' is not much different to saying 'starting work at 6am is good for everybody'.

Let's not start making some things more equal than others.

Well and in case somebody wants to fill up the 'Arguments opposing' paragraph - wouldn't be wrong, I'd say...

Following the same rationale, on Gert's concept on "IPs are IPs, IP users are IP users.": while true; do echo -n +; done
I really like this! (Except maybe for the 'special case' range - I'd suggest to just skip that, too. Better let's just say "Every IP is special."...

There was an example mentioned like $usual_suspect wants another /2, as they do regularly, and as $rir hasn't got sufficient insight into their net design, they just get it. This might be a bit exaggerated ;) - but I think it describes a major existing problem, which really needs to be addressed. In the end it reads to me: The request isn't sufficiently justified. It just shouldn't be approved, in this case - until justified.
A huge part of the address space (v4) obviously suffered this kind of problem since ages - and on the so called IPv4-depletion: just by 'cleaning up' some of the existing insane allocations, the 'problem' would disintegrate into thin air (and it might be used as another means to promote v6, btw).
I even consider this a moral obligation for anything assuming the role of coordinating the IP commons. I believe we have a really pressing issue here.

Regards,

	Chris