From end2end at rsuc.gweep.net  Mon May  2 05:33:08 2011
From: end2end at rsuc.gweep.net (Joe Provo)
Date: Mon, 2 May 2011 08:33:08 -0400
Subject: [e2e] NDDI & OpenFlow
In-Reply-To: <00d501cc05da$5b797e40$126c7ac0$@edu.pk>
References: <BANLkTik6S3e+nQ2W887RppmDKWTOjavvKw@mail.gmail.com>
	<507D6DCB-5000-4E9D-BF52-2B7C9AD4D890@eecs.utk.edu>
	<BANLkTinRddNgbN8UQE3WXnScmGdb-r8acw@mail.gmail.com>
	<4DB996DF.1010908@isi.edu>
	<BANLkTim2Ls-dJ7Xvp6_xC+i_CZaQiBtC4w@mail.gmail.com>
	<00d501cc05da$5b797e40$126c7ac0$@edu.pk>
Message-ID: <20110502123308.GA86500@gweep.net>

On Fri, Apr 29, 2011 at 12:27:48AM +0500, Zartash Afzal Uzmi wrote:
[snip]
> While
> replacing FIB with a more complex flow table appears okay in enterprise
> networks, I still have to convince myself that it will be a feasible idea to
> replace FIB with a more complex flow table in the service provider core
> where FIB performance requirements are more stringent. Anyone more familiar
> with Open Flow can throw some light?

Unfortunately the session was not recorded, but NANOG50 had a track 
involving basics/motivations, campus/national(jp, us)/regional(eu)
deployments & usage, as well as a laundry list of folks doing things 
'now' (see end of Davy's slides). Track synopsis and presentation 
links:
http://www.nanog.org/meetings/nanog50/abstracts.php?pt=MTY2OSZuYW5vZzUw&nm=nanog50

To Noel's MPLS question, "An Open-Source Interoperable MPLS LSR" was
recorded at the same meeting shows deploying a stand-alone openflow
device programmed from label-switching data, not really "based on" it.
Easy to see this being expanded, and the presentation mentions future 
work of plugging into controllers:
http://www.nanog.org/meetings/nanog50/abstracts.php?pt=MTYzNSZuYW5vZzUw&nm=nanog50

To Randy's point, I'd say that most of the tools we've been using 
over the decades can be used to assemble reasonable networks or be 
thrown together incorrectly. To what degree one scales the cloud of 
openflow switches behind a single/common set of controllers might 
be answered similarly to how one decided between N-by-single-processor 
'expendable' routers or fewer 'hardened' multi-processor routers.

Cheers,

Joe

-- 
         RSUC / GweepNet / Spunk / FnB / Usenix / SAGE / NewNOG