[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Thoughts on IETF scope definition
- To: iesg@ietf.org
- Subject: Thoughts on IETF scope definition
- From: Alex Zinin <zinin@psg.com>
- Date: Fri, 17 Oct 2003 16:51:55 -0700
Thinking about the IETF scope definition...
It seems we can identify multiple dimensions of the scope
and try to set the bounds on each direction. See below.
1. Position in the TCP/IP stack or layered networking model.
E.g., it used to be layers 3 and 4. We're now certainly
lower and higher than that (E.g. SUB-IP and APPS areas)
This is where the general level of expertise is discussed
as well.
2. Technology life cycle.
One could generally identify the following stages:
1) research
2) architecture and design
3) protocol specification
4) implementation
5) interop testing
6) bug fixing
7) deployment
8) operations
Research is done in the IRTF. We do some of 2 and are hopefully
good at 3. We do NOT do 4, 5, or 6. We don't do actual deployment
and operations, but we have an area where related aspects are
discussed
3. Relevance to the Internet.
Ranging from "I want to send a message to my son's PC upstairs and
it could be over IP", to "we'll have to change the Internet
router's queuing algorithms to accommodate blah". Here we could
potentially differentiate between different types of devices
and networks technologies could be developed for and implemented
on.
For this, I like the statement from the small group meeting,
specifically the part that says:
expected to have
wide deployment on, or form part of, the Internet
infrastructure
4. Relevance to IETF standards
ranging from "we only bother with changes to IETF standards if
they are directly relevant to the Internet" to "all changes to
IETF standards should be handled with the IETF".
Of course, other key questions we need to ask ourselves are:
- is there something we regret the IETF took on
- is there something we regret the IETF did NOT take on
(note that these questions are different from whether or not we
regret a specific technology was invented/marketed/deployed, i.e.
some of us may not be totally in love with MPLS, but when the
ball started rolling, was the IETF the right place for it?)
We should then look at the examples and see if there's a generic
trend there that should result in moving some area in or out of
the IETF scope...
--
Alex
http://www.psg.com/~zinin/