On Mon, 17 Jan 2005 18:33:08 -0600
"Stephen Sprunk" <[EMAIL PROTECTED]> wrote:
>
> Part of the problem is "site" has been overloaded to mean "administrative
> domain", whereas the common meaning is a single physical location. Neither
> makes sense to me in the 09 text, though.
>
Maybe it would be better to be even more generic and abstract about the
scope of these addresses, and have some text similar to the following.
This would eliminate specific reference to certain technologies and the
different scenarios and execeptions that can be created with them :
"Addresses and prefixes allocated from within one or more /48 ULA
address spaces are only to be visible within the administrative domain
of the network in question. Traffic with ULA sources and / or
destinations and route advertisements containing ULA prefixes should
(SHOULD(?)), by default, be prevented from leaving or entering the
administrative domain."
The edge of the administrative domain may coincide with the edge of a
BGP AS, an IGP domain, or the boundary between a service provider and
customer, in the case where static routing is being used."
And then the text referring /48 or longer exceptions such as the VPN
case, merging of organisations etc.
As much as it would be good to stay generic, identifying BGP as a
special case with a default block on the ULA /7 prefix would probably be
best as non-intentional leaking of ULA address spaces could have global
consequences. Other than that though, I think it would be better not to
identify or suggesting boundaries to constrain ULA prefixes to, such as
IGP areas.
Regards,
Mark.
--
"This signature intentionally left blank."
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
