Paul,
> http://www.ietf.org/internet-drafts/draft-francis-ipngwg-site-def-00.txt
> This draft proposes the following rigorous definition of a site:
> A site is defined as the set of routers that fit inside a hexagonal
> shape with a distance of one kilometer between opposite corners.
> Any hosts sharing a link with a router in a given site are
> considered part of the site. The hexagonal shape is chosen because
> it can be tiled while still approximating a circle.
>
> The administrator is free to decide exactly where to lay the
> hexagons. However, they MUST not overlap, and every router MUST be
> within the boundary of one and only one hexagon. All stories of a
> multi-story building are considered to be in the site covered by the
> hexagon.
Won't work, can't work. An organisation that happens to span 1.1 of these
hexagons, or even 20 of them, is not going to contort itself to meet
such a definition. I can't imagine why any campus style organisation would
even dream of defining a single campus as more than one site, and what about
slightly split campuses? Or campuses with 5 km between their two halves, but
running a single network? I think your "north bank/south bank" example shows
exactly why any attempt to define a site rigorously is doomed.
I think a site is whatever a site network manager says is a site. The scope
of site-local is whatever the site network manager says it is.
Tell me a good reason why we couldn't run the whole of IBM world-wide
as a single site-local scope, if we wanted to?
Brian
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
