[arin-ppml] ULA-C

[Chris Engel]

David Farmer wrote:

> I believe, the stewardship the RIRs provide is necessary, and
> the RIRs'
> policy processes can properly manage the risk.  As for cost,
> while not a
> policy matter, in the shorter-term ULA-C is not going to be as cost
> effective as some would hope.  However, if in the longer-term
> the risks
> can be managed, then I have to believe that the RIRs will do
> the right
> thing.

David,

Speaking only for myself, I don't think the fee's I've heard thrown around (I think it was something like $1200 per year?) would present much of a stumbling block for most of those on the Enterprise who's addressing needs would actually justify ULA-C. Realistically, that's less then the coffee budget. No significant Enterprise should even blink at that. As long as there is no fee or justification for ULA-R (which I'm assuming because there is no registration) then everything should be fine. If you can't justify $1200 per year, then the burden imposed by using something which is not guarantied unique but still statistically alot bigger then RFC1918 is not that significant. Frankly I would think that most smaller Enterprises/startups will probably opt for ULA-R anyway. That would be our plan right now (assuming we can find NAT support in IPv6).

However, I can see the value in ULA-C for some Enterprises...and you'll definitely get some buy in there..even with significant fee's. I wouldn't have any problem fee/justification wise if ULA-C was treated identically to PI. The only caveat should be that justification for ULA-C shouldn't count against justification for PA/PI... you should be able to get both types of addresses for the same device. The real value of ULA-C (IMO) is the understanding that it's not supposed to be globally routed. Yes, that's enforced by convention only...but that's no different then pretty much anything else on the internet.


Christopher Engel