On 2007-06-12 09:27, Paul Vixie wrote:
Not that it really matters, since ULAs never appear off-site anyway.
depending on what you mean by a site, both in relative and absolute terms
of space and of time, there might be general agreement on this point, or not.
hopefully the regress isn't infinite. care to take it a step and see?
The regression is easy: the site for a given ULA is the routing domain
within which that ULA is not filtered.
That is not intended to be frivolous. If two enterprise networks agree
to recognize each other's ULA, the ULAs can be routed over a link (or VPN)
between those two networks. But both enterprises' ISPs will filter ULAs.
I'm not making this up. Something very like it is current practice on
IPv4; it's just less well defined since we don't have ULA space.
Brian
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
