On 9/16/15 9:10 PM, Panos Kampanakis (pkampana) wrote:
> Also I am not sure about what the "bickering on the topic" refers to.

At one time this was an opsawg deliverable but was dropped
because of the inability to come to consensus on some basic
questions about the document, including the question of what
problem it's intended to solve.  I don't particularly look
forward to another slugfest or the inevitable involvement of
ideologues.

That said, I do think there's value in such a document, for
several reasons.  One is that over the years there have been
a number of efforts to abstract firewall behavior as input
to the design of IETF protocols.  It's a challenge because
firewall behavior does tend to be highly vendor-specific, and
we've published several specifications that try to do it
(nsis's NAT and firewall layer, midcom, etc.).  It would be
(and would have been useful) to have a document describing
where we can reasonably expect to have firewalls in the network
and what we can reasonably expect from their behavior, to
be able to make better protocol design decisions.  Note that
this is distinct from a document making deployment
recommendations (or at least explicit ones).

I also think that there's value in reasoning about architecture,
and publishing a document describing that reasoning.  I suspect
that having it come from the IAB might reduce some of the
friction in moving the document along, but it might not reduce
it enough.  But it's quite clear to me that because nearly any
discussion related to middleboxes has erupted into ideological
warfare we've been hampered in producing particularly thoughtful
architectural work on what it means to have them in the network,
instead knocking out stopgap workaround protocols here and there,
and that's unfortunate.

This should particularly be a concern given the possibility
that i2nsf is going to be chartered, where there will be some
serious issues around trust, authority, and delegation.  It
very likely would have been useful to have this document as
those discussions progressed.

That said, I expect that if this is adopted again by opsawg,
it will not be a fun ride.  But it might be worthwhile to do it
anyway.

Melinda

_______________________________________________
OPSAWG mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsawg

Reply via email to