Re: Going forward with zero-config tunneling requirement

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

On Wed, Sep 29, 2004 at 07:38:02PM +0300, Pekka Savola wrote:
> 
> (i.e., draft-nielsen-v6ops-zeroconf-goals-01.txt + 

Karen's draft doesn't specifically mention the 3GPP specific assumptions
in the draft in Section 4.2 (i.e. TEP in the provider network, with or
without private IPv4 addresses used, no NAT traversal required, etc).
However, she does mention "no-NAT" in section 4.1, and the assumptions 
in 4.1 are really kind of defining the 3GPP scenario... so the text of 4.2
seems to become a little redundant.

Essentially the assumptions of 4.1 define the scenario.  Thus one solution
is to rework section 4 to define the scenarios, as others have hinted or
suggested.  So what are those scenarios?  Using the four v6ops transition 
scenarios of 3gpp, unmanaged, enterprise and isp seems sensible.

For 3gpp, we already have the scenario for zct within the 3gpp provider

For enterprise, can we assume the zct is also within the enterprise?  My
feeling is yes.   So the scenario is very similar.

For umanaged (=soho largely), the zct would not be within the soho network,
rather from a host in the soho network to some external service.  This
probably brings in NAT traversal, etc.   In this case, perhaps we should
only be considering assisted tunnelling.

For isp, I'm not sure zct has any real meaning.

Thus (unless we have other broad scenarios?) we should push Karen's draft
for 3gpp and enterprise, and assisting tunnelling for unmanaged.   This
could be documented in a reworked section 4 of Karen's draft?

Let's not try to dive down unnecessary ratholes here :)

Tim

PS. If there are other broad scenarios, we should do the v6ops analysis 
    for them for transition...(!)