On Dec 5, 2007 2:39 PM, Iljitsch van Beijnum <[EMAIL PROTECTED]> wrote: > ULA is LOCAL. > > It has nothing to do with PI. >
sort of correct... I believe the fear here is that if you are in a world of provider-assigned ip space without any simple hope for renumbering you will look for ULA-x as a 'no renumbering' solution and try to find some form of NAT initially that will make this all work for you with minimal work. If NAT is no longer avaialble and you have a /48 in ULA-x space that is 'yours' why can't you ask your provider to route that for you? and all of his peers/customers to do the same? This could lead to a very large pool/use of ULA-x in 'public' places in a short period of time. I'm not sure that the Internet today is ready to accommodate that sort of route growth. (this goes somewhat to the point below as well) > People need address space to number the links between their SQL and > web servers. This is completely orthogonal to address space used on > the internet. > this is also sort of correct. I think that ULA and the like come from the mentality that in the ipv4 world people numbered lots of 'internal' things out of 1918 space, following that line of logic people will want to use 'private' space on internal networks in the ipv6 world as well. They'll also want to, in most cases, have that space NAT'd to the real world... perhaps not NAT, perhaps some other equivalent technology, but still a 'private' address in the end. When I last lobbied for a ULA-x type of option it was for this very sort of thing, 'internal networks' that might later be connected to other 'internal networks' and thus I thought that ULA-C might make the most sense. I think that this all skips over a very real issue which is 'how does your host/router/blah handle having more than one address and how does it decide which to use when?' coupled with that is the 'do users want to manage more than one address going forward and all of the associated foo with that? (dhcp/dns/service-binding/firewalls/acls/policy-foo/etc)' I think that somewhere we ought to find the requirement for ULA before creating something like this. Additionally, without a reasonable solution for how enterprises deal with their current numbering issues (multihoming, internal address+NAT) the ULA-x discussion is going to be mostly headless and directionless... -Chris -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------