Joe Touch wrote:
> Hi, Aaron,
>
> Aaron Falk wrote:
> ...
> > Regarding an RG charter, I'm not asserting that terms like 'slice' and
> > 'RSpec' should be included in the charter but I think there is a growing
> > community around a conceptual model that includes establishing virtual
> > environments that include link, computation, measurement, storage, and
> > other resources. This is clearly more general than Joe's scope of only
> > discussing the communications elements. However, IMO, this broader
> > scope is a useful area for collaboration and is appropriate for an RG.
> > I'd be interested in hearing whether others agree. If so, perhaps there
> > needs to be a name change for the RG ("Virtual Testbeds", maybe?)...
>
> It's interesting to talk about virtualizing everything at once. We have
> fairly well-defined terms for talking about OS virtualization, though -
> PlanetLab co-opted the term slice for that a long time ago, and it's not
> clear that GENI's redefinition of the term will change that.
Larry Peterson was part of the group developing the initial GENI
conceptual design and the PlanetLab folks are an active part of GENI so
it's not surprising that concepts developed on that project are present
in GENI. I wouldn't say that the concept of a 'slice' has been co-opted
but rather extended from referring to purely host-based resources to a
more diverse set that includes network resources.
>
> Regardless, OS stuff is somewhat outside the scope of the I in IRTF.
> IMO, that overlaps more where the Grid forum or various cloud computing
> forums sit.
Internet, IETF, and IRTF work reaches up beyond links in many areas.
Email, for instance. Or, if you prefer more current examples: SIP and
P2P. I agree that the GRID folks have some expertise in this area. I'm
hopeful some will come to play.
> Testbeds are useful important, but IMO many such issues
> could easily fit under the IETF operations area.
My assessment is that most of the attention in this group is coming from
folks working on developing testbeds which use virtualization for more
than just links, e.g., FIRE, GENI, JGN2Plus, Akari. As systems builders
(not operators, at this point) interesting research issues will come up.
An IRTF research group could be a forum for collaborating on them. And,
don't get me wrong, I think developing a sound conceptual basis for
talking about virtualization in networks will be helpful here.
> The Internet research
> here, AFAICT, is in network virtualization, which is where I thought
> this group was headed. I could be wrong, of course.
In my view, that is a subset of the interesting issues. But, as a
practical matter, the RG should be structure around the interests of the
participants. Hopefully others will speak up about their interests.
To be very explicit, I'd like to ask whether the folks on this list
who might participate in a research group feel that such an RG
should focus on a) the virtualization of networks or b)
virtualization of systems including networks and others things such
as hosts, services, storage, sensors/actuators.
regards,
--aaron
Note Well: Messages sent to this mailing list are the opinions of the senders and do not imply endorsement by the IETF.