The MPLS WG Archive

[Brian E Carpenter <brian@hursley.ibm.com>]

Kimmo.Raatikainen@nokia.com wrote:
> 
> > -----Original Message-----
> > From: EXT Brian E Carpenter [mailto:brian@hursley.ibm.com]
> > Sent: 02. June 2000 0:39
> > To: Bora Akyol
> > Cc: diffserv@ietf.org; 'mpls@uu.net'
> > Subject: Re: "This is a gripe" [was Re: [Diffserv] MPLS Diffserv
> > Extensions related questions/comments
> >
> >
> > Bora,
> >
> > I don't think we will see reports of Diffserv deployment
> > experience until
> > production releases from major router vendors support Diffserv.
> >
> > The edge boxes that I'm thinking of will need to do MF
> > classification and shaping
> > at LAN speeds (whatever that means in a particular year).
> > Other boxes only need
> > to do BA classification and policing.
> >
> > But this is old ground that we settled in the early days of
> > diffserv. And
> > I don't see what it has to do with MPLS.
> >
> >   Brian
> >
> Please keep in mind that gigabit cards are already available. In our
> experiments we have found out that a high-end PC running Linux can send at a
> rate of 250-300 megabits per second using TCP/UDP.

Of course. Edge boxes will need to support a gigabit. The server itself may
function as the diffserv edge box; this is already possible at 100 Mbit/s.
I would expect to see MF classifiers running at a gigabit soon enough.

My concern is to keep the job of core boxes that do IP direct over photons
as simple as possible, and I don't think we should ask them to do more than
BA classification.

  Brian

  Brian