On 2/23/13 12:23 AM, Hannes Tschofenig wrote:
> In a nutshell NSIS was really developed to be fairly generic.

I think nsis is a truly heinous protocol, but that's not really
important.  We needed to document relevant protocols because
people kept asking if we'd thought about this one or that one.
I've done a pile of middlebox protocols myself, and brought both
the initial midcom work and the stuff that metastasized into
nsis into the IETF to start with.

But, what we're really interested in doing is taking a look at
the overall picture and identifying where there's no existing
tool (or solution, if you will), since the IETF has tended to
punt on some of the more challenging problems around topology
and whatnot.  We've got a great heap of protocols to communicate
with middleboxes but we're missing a lot of the structure that
we'd need to move flow-associated state to a new middlebox as
a network flow has endpoints migrating.


Melinda

_______________________________________________
nvo3 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nvo3

Reply via email to