What ever a site is defined to be one would have the following issues:
Distinct: not necessarily uniquely numbered but able to be differentiated from other sites with easily distinquishable routing boundry
Routable within: sub-networks to support hierchial structure
handle this in the following manner:
Distinct: bits 17-48 designate a separate site, can be set using bits 17-48 from globally routeable address assigned to one of the site's networks or by creating a nearly random number as previously proposed on this list.
Routeable: bits 49-64 are already designated for subnetting a site.
As a Network Administrator I can see the requirement to be able to set up my our intranet routing domain, and would like to know that I will not inadvertantly route out to another site (same designation). An easy way to avert this is to use bits 17-48 for site designation.
Beyond this, the loose definition of a site is adequate.
DG Ses
DALE G SESVOLD
Senior Network Engineer
MacAulay-Brown, Inc
> However, we could take anything that fit in a /48 and call this a site.
> I'm not sure that it would be a good idea...
I think it is a good idea, though it cannot be the _definition_ of the
"site".
IMHO, site-local scope will be used to protect against renumbering,
and for the networks not connected to the Internet.
For this purpose, some connected networks serapated within a single
geographic location _can_ be a <site>. But a <site> is not limited to
be an single location.
Unlike IPv4 usage of a private address, IPv6 site cannot be nested.
So I don't think it is a good idea that you assign a <site> for a too
small piece of networks only for the reason that they are geographically
isolted.
Atsushi
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
