Re: I-D ACTION:draft-ietf-ipv6-unique-local-addr-08.txt

[Bob Hinden <bob.hinden@xxxxxxxxx>]

Scott,

> > There are other things we are doing (renumbering procedures, multi6, the
> > NAP draft) to try and deflect this danger, but ULAs don't increase it.
>
> we disagree - I think they are an "attractive nuisance" (to use a
> pseudo-legal term) and we (the IETF) will rue the day that we approve
> this idea

I don't agree.

One reason I think this is not a significant concern is that the ULA 
prefixes are too long.  /48's aren't big enough for large 
organizations.  They will want a shorter prefix if they are going to go to 
the trouble of "convincing" their ISPs to route them for general usage 
(instead of using a PA prefix).  Instead of getting their ISPs to route a 
/48 ULA prefix, I think it is much more likely that they will do what 
people do today.  They will become a member of an RIR, and then become an 
LIR and get a PI prefix.  This is fairly common as I understand it.

If there is a need in the community for PI addresses that can be used by 
large organizations, then the RIRs should address that need directly and 
provide them.  It's not that hard for organizations to become an LIR today, 
so the RIRs might as well make the rules clearer and allow organizations to 
get these assignments without having to become a member of the RIR.

The underlying concern with this is, of course, the size of the default 
free routing tables and the stability of BGP.  ULA's aren't going to make 
this worse (or better).  Organizations that really need PI addresses will 
get them and then get ISPs to advertise them (as they do today).  It 
doesn't matter if the entry in routing table is a RIR assigned prefix, a 
ULA prefix, or something else.  It's still one entry in the routing table 
and routing updates.  The problem is a real one and is not created by the 
IETF standardizing ULA addresses.  Their intended usage if for local 
addresses (hence their name) and most people think they are an improvement 
over site-local addresses.

Bob





--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@xxxxxxxx
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------