The MPLS WG Archive

["Shun Yao" <shyao@ece.ucdavis.edu>]

I believe Darren was referring to all-optical ~circuit~ switch, not packet
switch. Up to now we haven't seen one working ~fully-optical~ (no
transponders, strictly non-OEO data plane) switch in the market, although
many companies are claiming to be building one. In-band signaling is not as
simple as split/amplify. First of all cascaded amplification add noise and
other physical impairments, then when you push the data rate higher, all
sorts of things will happen to the signal. If out-of-band signaling is used,
performance monitoring of the data path will be difficult. I would think
twice even if a vendor throws me a definite answer, because there is no
working product yet.

Shun

> -----Original Message-----
> From: owner-mpls@UU.NET [mailto:owner-mpls@UU.NET]On Behalf Of Alec
> Myers
> Sent: Tuesday, February 06, 2001 11:34 AM
> To: 'mpls@uu.net'
> Cc: Darren Freeland
> Subject: Re: G-MPLS on all-optical switches?
>
>
> Bearing in mind:
>
> 1. The crudity of pure optical memory (fibre delay lines)
> 2. The lack of pure-optical logic
> 3. The lack of a fast (packet rate) wavelength switching technology
> 4. The delay in line clock-recovery between differing
> transmitter/receivers
> pairs at the start of each packet
> 5. The difficulty in dynamically recovering from non-linear effects in
> optical paths whose length varies per-packet
>
> I wouldn't expect to see a real FULLY optical switch any time soon.
>
> There are a few papers around that describe possible control
> schemes, but
> they are all rather crude, and make only passing reference to MPLS. If
> you're interested I can forward the references.
>
> -A
>
>
> ----- Original Message -----
> From: "Darren Freeland" <darren.freeland@alten.co.uk>
> To: <mpls@UU.NET>
> Cc: <ip-optical@lists.bell-labs.com>
> Sent: Tuesday, February 06, 2001 2:24 PM
> Subject: G-MPLS on all-optical switches?
>
>
> > Hi,
> >
> > Can any vendor give me some information on exactly how they
> intend to
> > run G-MPLS on FULLY optical switches?  For the control
> signals, will you
> > use optical splitters or the like to tap off some of the lightpath
> > inband (amplifying to compensate for the power loss), or
> will you use an
> > outband network?
> >
> > Any info much appreciated ....
> >
> > Darren.
> >
> >
>