[Sdnp] Proposed SDN BOF charter

Robert Raszuk robert at raszuk.net
Tue Sep 20 09:09:58 EDT 2011


Hey Tom,

>>> I think this WG would be most useful to determine the language
>>> application or host talks to the network as today there is very
>>> little done in this area. The closest is abandoned RSVP Intserv
>>> due to the state overload of micro-flow reservations.
>
> Cool. We are getting somewhere because this is indeed one of the
> goals of the proposed WG. 8)
>
> --Tom

I am 100% supportive for that.

And the use case would be for example Hadoop getting ready to start a 
job for time duration N and asking "network" please connect this list of 
VMs for 40GB for N sec. Or better some grid interface doing it 
transparently for Hadoop.

However for other types of transports for given applications I would 
very much prefer not to require any change in the application. Network 
could/should be smart enough to recognize what is being send to it and 
treat it accordingly.

On this last point I think charter prohibiting enhancing data plane or 
even excluding option to recommend data plane enhancements (for example 
to the extend of at least recognizing the flows in a standard way) looks 
to me like a rock on the road.

Quote:

"This WG will focus solely on the communication protocol between
applications and SDN Conductor(s), SDN Conductor(s) and control software 
"plug ins"."

Cheers,
R.






More information about the SDNP mailing list