Re: 3gpp-analysis: Recommendation on tunneling in the UE

[Tim Chown <tjc@ecs.soton.ac.uk>]

On Tue, Nov 18, 2003 at 01:02:12PM +0200, Pekka Savola wrote:
> 
> I'll respond, just in case you were serious.. :-)

I wasn't :)
 
> The point of ISATAP (as I see it) to run it as an intra-site tool.  It
> could be rather useful there, especially if there are VPNs and such where
> you can't use VLANs etc..  However, rather than specifying something
> similar that runs inter-domain, we should focus on determining what is
> actually needed when running inter-domain.

I agree.  VLANs require an amount of administrative configuration, and
is more aimed at enabling whole subnets/links, whereas ISATAP is more
"opportunistic" and aimed at sparse/scattered hosts in a site.
 
> Note: ISATAP however is less intrusive for the global architecture, as it
> uses the interface ID for its operations, does not require a prefix -- so
> it's transparent to the rest of the 'net users, rather than 6to4/Teredo.
> So, it's not IMHO totally fair to compare ISATAP to 6to4/Teredo w/ 
> interdomain architecture.

True.  But we should consider that for any proposal too.  I'll add it
to the tunneling considerations draft (I propose widening the scope of
that draft to general tunneling rather than unmanaged network-specific).

Tim