Re: 10xGbE on DWDM

[Paul Gunning <paulg@xxxxxxxxxxxx>]

Hi Bill,

I find that we are talking the same language. Now let 
me add a few comments and questions.


Bill St. Arnaud wrote:
> 
> Paul:
> 
> > Many industrial research teams: Alcatel, NTT, Nortel and
> > BT and university teams: TUD-Denmark, ETH-Zurich, HHI-Berlin,
> > have had active research programmes in the optical regeneration
> > area for the past decade. Prototypical devices have been built
> > and demonstrated, but a market hasn't yet emerged because
> > electronics has alays provided a cheaper alternative. I'm not
> > convinced that 10xGbE will kickstart the market for optical
> > regeneration---a pity. However I believe that 100xGbE WILL!
> 
> I think the major issue is cost.  There several trans-oceanic cables (e.g
> the new USA-China cable) up to 12,000 km in length that use 100% optical
> repeaters.  The technology is there, but as you point out electrical
> regeneration is a lot less expensive.  That is why currently CWDM technology
> is proving more attractive than DWDM solutions.  I am sure in time that
> situation will change.
> 
> I suspect in a few years most networks will be 100% optical and data
> transparent and many of the debates we are having about the type of coding,
> scrambling, etc will be largely irrelevant.  In fact some researchers are
> proposing that we use only SDL - simple date link protocols as a replacement
> for SONET and/or GbE.  SDL protocols only have a start of packet indicator -
> no CRC, no overhead bits, etc


I fully concur. I would welcome a pointer as to where I could
find out some more information about the SDL protocol(s). It(they)
sounds very, very interesting.


> > Moreover each network node can have an independent
> > local clock; every clock within the network being
> > nominally synchronous. Global synchronisation IS NOT
> > required much the same as with current Ethernet networks.
> 
> Packet over SONET routers and Ethernet networks do not require any form of
> global clock hierarchy like traditional SONET networks.  The clock is
> recovered internally in both cases.


I realise that. But is there a requirement for additional
frame overhead in that you need a contiguous sequence
of preamble "training" 'bits' at the front of the frame 
to phase sync. the Rx PLL?


> > Also why are Cisco moving away from full-blown SDH/SONET
> > functionality with their DPT technology?
> 
> I can't speak for CISCO, but DPT is a much more efficient use of bandwidth.
> Most SONET systems have an unused protection fiber which DPT can utilize
> simultaneously with the working fiber.


Agreed. But there ARE clock management issues between
adjacent nodes on a DPT ring in that you have to
explicitly nominate a master and slave relationship.


As the serial bit-rate increases clock synchronisation,
and the management of same in an arbitrary network
topology, becomes more and more crucial.

David Cotter and Andrew Ellis here at BT Labs have pointed
this out and suggested some solutions in:


	"Asynchronous digital optical regeneration and networks,"
	Cotter D., and Ellis, A.D.: Journal of Lightwave Technology, 
	vol.16, no.12, Dec. 1998, pp. 2068--2080.

For those with OSA access here is a link:

	http://epubs.osa.org/jlt/confirmed.cfm?doc=1298/16lt12-cotter.pdf


In the context of serial 100xGbE, Paul Prucnal's group at Princeton
have an intersting presentation in the same issue:

	 http://epubs.osa.org/infobase/jlt/htm/1298/2169_16lt12-toliver.html


If as you said above, SDL protocols are being proposed for 
higher bit-rates then spare a thought for an paper from
as far back as 1995 from our group:


	"Self-routing of 100 Gbit/s packets using 6 bit `keyword' 
	address recognition," D. Cotter, J.K. Lucek, M. Shabeer, 
	K. Smith, D.C. Rogers, D. Nesset and P. Gunning:
	Electronics Letters, vol. 31, no. 25, 7th December, 1995, 
	p. 2201-2202

For those with IEE/IEEE access


	http://ioj.iee.org.uk/pdf/el/1995/25/19950959.pdf


> > Note: Synchronisation between adjacent nodes is not effected
> > via SONET/SDH with DPT. What are the deployment plans
> > for 40Gbit/s SONET/SDH? Are solutions available now? Soon?
> 
> Nortel and NEC have both announced OC-768 SONET products 


OK. I wan't aware of that. I'll take a look. I wonder how
critical timing and timing management is at 40Gbit/s?


> > Ultimately the most compelling advantage of Ethernet lies
> > in its modularity and simplicity. Sure even I can plug
> > the RJ45 into the wall socket beside me and it works!!!!
> > I fear that some of the solutions proposed for 10xGbE so
> > far have lost sight of this....and the fact that 100xGbE,
> > ...1xTbE will inevitably follow.
> >
> >       Keep it simple, keep it serial!
> 
> Agree 100%


Great.


> > Finally, a plea for tolerance and respect. I think 10xGbE
> > is where two communities: Datacomms and Photonics are
> > meeting for the first time we really have a lot to
> > learn from each other.
> 
> Agree 100%


Even greater.


> > Paul.
> >
> > p.s.  The EDFA window is not narrow. It's 25THz from
> > 1535 -> 1565 nm! Indeed there is an extra 25THz available
> > from 1580 -> 1610 nm using next generation EDFAs.
> > Plenty of room for expansion!
> 
> But it is a lot narrower than CWDM.


Bill, I'm not absolutely sure what CWDM is. I think 
it is Coarse Wavelength Division Multiplexing? If so, 
then CWDM  is not really appropriate at 1300nm/1550nm 
because  deployed TELCO fibre is expensive as you 
appreciate (not in terms of Silica, but rather 
access rights of way, deployment costs, maintenance 
and so forth.) Consequently you really want to 
squeeze every last drop of bandwidth out of it.
To do this its best to keep the wavelengths as close 
together as possible hence the necessity for "expensive"
temperature control for Tx Lasers and Rx demuxes.

To this end, might it be possible to "borrow" the ITU
wavelength grid for SONET/SDH? Then leverage the "mature"
single wavelength laser diodes that have been produced
to the ITU-T spec for appropriation in 100xGbE Gigachannel? 
(That is assuming that serial 10xGbE is DWDM muxed to 100GbE as 
an interim solution before "full-blown" serial 100xGbE.)

Unfortunately I can't attend the meeting in Idaho
but I would be very interested to learn of any proposals
for 100xGBe. Finally, I think the Lucent team who did 
the GigaChannel demo at Networld+Interop last month 
have really excelled:


	http://www.bell-labs.com/project/gigachannel/

 
> > The filters and optical sources used for DWDM based
> > on Arrayed Waveguide Grating (AWG) technology will
> > cost next to nothing when Moore Law economics prevails.
> > Just like IC's before them, its only "processed sand"
> > festooned with a few choice elements.
> 
> Agreed but the same economics will also affect CWDM


Yes. But deployment economics would tend to prefer DWDM
so as to maximise bandwidth per physical fibre. DWDM allows 
you to have many additional "virtual" fibres per physical fibre.

Finally I think the physical layer should be hidden from
the user as much as possible. In fact its presence should
never "be felt" hence it is best practice to set the BER 
as low as is possible bearing in mind the economic and
physical constraints. It should be defined in the standard
and then written in silica, GaAs etc.


> Bill

Paul.
-- 
Paul Gunning
Futures Lab
BT Laboratories

Phone:	+44 1473 647049
Fax:	+44 1473 646885
http://www.labs.bt.com/people/gunninp