Re: [sami] Bringing new work into the IETF

[Melinda Shore <[e-mail]>]

On 8/22/11 5:52 AM, So, Ning wrote:
> It is important to have a common agreement on requirements.  I
 > stated my opinion and use case.  I will reach out to the carriers
 > I have been work with in other areas to see if they also have the
 > similar requirements.  If it is true, then IETF has to
 > play a key role in this.

I tend to think that this is basically of a piece with wishing for a
pony. That said, as important it is to reach an agreement on
requirements, it's even more important to reach an agreement on
what piece of work is to be done.

That said, the IETF is the Internet Engineering Task
Force.  Just because something moves across a network doesn't mean
that it's necessarily a network problem.  There's also the question
of past being prologue - in an odd feat of chartering rserpool
tackled moving live servers but not moving any of the state associated
with them, and the question of moving middlebox state didn't even
come up.  I tend to think that it's because people involved in the
chartering decision looked at the problem and decided that if the
working group took it on they'd never finish.  Anyway, to
circle back around it seems to me that it's probably a good idea to
stay focused on the network aspects of, erm, whatever it is you
think the IETF should do.  This is not just because of the
question of organizational scope, but also because of the
expertise within the group.

I've also thought about the question of running layer 2 subnets
over the transport layer but that's really not what has been
proposed so far.  I think that a good way to get a handle on
the answer to David's question, "If we build it, will they come?"
would be for one of the proponents of this work to try to
convince Huge Virtualization Vendor <x> to participate in
developing a mechanism that would allow their images to run on
Freeware Hypervisor <y>.

Melinda