Complex Problems

[john.loughney at nokia.com]

James,

> I'm sorry, I think this is directly relevent to the list's discussion topic. My
> message was in response to Marshall's proposal for the IETF to not do any
> architecture work and just do engineering for other SDOs that bring their
> architecture and requirements to IETF, but I believe this has in fact been
> happening for some time now with respect to 3GPP, and I believe it has led to
> some significant problems, and a monumental increase in work load for the ADs as
> they do the technical liason. 

Well, 3GPP interaction with the IETF is only really relevant for 2002.  There was
not much 3GPP-specific work prior to that.  I think the problems we are discussing
existed earlier than that, so I don't see that 3GPP is directly relevant.  In
fact, 3GPP has been working quite nicely with the IETF in the IPv6 and IPv6OPs
working groups, as well as others.  There has been some divergence, perhaps,
with the SIPPING working group - but I'd venture to say that SIP as a protocol
has required quite much work in order to be deployable on the Internet - not
to mention interworking with the PSTN (but that is much more of an IETF-ITU 
laison), so there would have been much SIP work without 3GPP last year in
any way.

For the record, I do think the IETF should be doing architecture work of some
sorts, and I'd say that the amount we are doing is about right.  However, 
better education and promotion of the architecture work is needed to be
done here inside of the IETF & the responsible parties to do this are probably 
the work group chairs.

> The ADs, WG chairs, and WG members not associated
> with the other SDO could simply roll over and not push back on requests that
> violate basic Internet architectural principles, but, to their credit, the folks
> involved in the 3GPP liason have not done that. Rather, they have attempted to
> educate 3GPP on why the Internet architecture is as it is and also tried to show
> 3GPP how their requirements could be met in a way that is consistent with the
> Internet architecuture. But this comes at a massive cost in time for the ADs.
> 
> So if IESG work overload is a problem, then this is one 
> contributing factor.

As John Klensin stated, perhaps this is more of an issue of management techniques
and style more than anything else.  I'd also suggest that part of any
percieved overload can be directly attibutable to the success of IP and IP
protocols.  It stands to reason that as IP is deployed in more places and
for more reasons then we will have more and more work for the IETF.  If
not managed well, work overload comes as a cost.

br,
John