RE: 1310nm vs. 1550nm window for 10GbE
- To: <bgregory@xxxxxxxxx>, "David W Dolfi" <dolfi@xxxxxxxxxxxxxxxx>
- Subject: RE: 1310nm vs. 1550nm window for 10GbE
- From: "Bill St. Arnaud" <bill.st.arnaud@xxxxxxxxxx>
- Date: Thu, 6 May 1999 10:38:50 -0400
- Cc: <stds-802-3-hssg@xxxxxxxx>, <dolfi@xxxxxxxxxx>
- Importance: Normal
- In-Reply-To: <000AC176.C21031@xxxxxxxxx>
- Reply-To: <bill.st.arnaud@xxxxxxxxxx>
- Sender: owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx
Hmmm. I just assumed that 802.3 HSSG would be looking at 1550 solutions as
well as 1310 and 850
I agree with you on longer haul links it makes a lot more sense to operate
at 1550
I am not a big fan of EFFA pumping. It significantly raises the overall
system cost. It only makes sense in very dense wave long haul systems
typically deployed by carriers.
CWDM with 10xGbE transcivers should be significantly cheaper. That is
another reason why I think there will be a big market for 10xGbE with all
those transceivers every 30-80km on a CWDM system. However there is a
tradeoff. There is greater probablity of laser failure with many
transceivers and the need for many spares. I figure somewhere between 4-8
wavelengths on a CWDM and transceivers is the breakpoint where it is
probably more economical to go to DWDM with EDFA. Also EDFA is protocol and
bit rate transparent.
An EDFA will cost about $60k. But EDFA window is very small, so wavelength
spacing is very tight requiring expensive filters and very stable,
temperature compensated lasers at each repeater site. Also laser power has
to be carefully maintained within 1 db otherwise you will get gain tilt in
EDFAs. A loss of a signal laser can throw the whole system off, that is why
you need SONET protection swicthing. But companies are developing feedback
techniques to adjust power on remaining lasers to solve this problem.
A single 10xGbE transceiver will cost $10K??? Probably less. So 6 10xGbE
transceivers will equal one EDFA. No problems with gain tilt. If you lose
one laser you only lose that channel, not the whole system. Protection
switching not as critical, etc
Bill
-------------------------------------------
Bill St Arnaud
Director Network Projects
CANARIE
bill.st.arnaud@xxxxxxxxxx
http://tweetie.canarie.ca/~bstarn
> -----Original Message-----
> From: bgregory@xxxxxxxxx [mailto:bgregory@xxxxxxxxx]
> Sent: Thursday, May 06, 1999 9:57 AM
> To: David W Dolfi; Bill St. Arnaud
> Cc: stds-802-3-hssg@xxxxxxxx; dolfi@xxxxxxxxxx
> Subject: 1310nm vs. 1550nm window for 10GbE
>
>
> There has been a lot of talk about CWDM (WWDM) in the 1310nm window
> and uncooled DFBs in the 1310nm window. From my experience,
> there is
> nothing really preventing this from being CWDM in 1550nm
> window or an
> uncooled DFB at 1550nm. There a sources for both 1310nm and 1550nm
> uncooled DFB technologies. Fabry-Perot lasers (mounted using some
> sort of external or integrated wavelength selection technology) are
> more common at 1310nm, but are still widely available in the 1550nm
> window. A wavelength window centered on 1550nm would offer several
> advantages:
>
> ## Higher output power while still meeting Class I eye safety
> requirements (a significant issue!)
> ## Less fiber attenuation
> ## Possibility of EDFA pumping
>
> At the meeting in Austin, I don't recall any company suggesting a
> 1550nm window. I agree with David's goals of using this
> technology in
> longer distance links. A 1550nm window appears to offer a LOT more
> flexibility for both short and long links.
>
> Comments?
>
> Regards,
> Bryan Gregory
> bgregory@xxxxxxxxx
> 630/512-8520
>
>
> ______________________________ Reply Separator
> _________________________________
> Subject: RE: Why 10xGbE may be a bigger market than 1xGbE - a custome
> Author: "Bill St. Arnaud" <bill.st.arnaud@xxxxxxxxxx> at INTERNET
> Date: 5/6/99 8:35 AM
>
>
> David:
>
> I am only speculating that WWDM may not work on long distance with ZDSF.
> Because ZDSF fiber has steep dispersion slopes away from the zero point
> wavelengths that are located the furthest away from the zero point would
> suffer significantly more dispersion than those closer to the zero point.
>
> The dispersion is non linear on ZDSF fiber so the dispersion generally
> cannot be compensated with dispersion compensating devices. Most
> ZDSF fiber
> is 1550, and so in the 1310 range the dispersion would be linear.
>
> On 10 Km distances this may not impact WWDM particularly if you
> use narrow
> spacing. I was thinking more in terms of 20-80 km distances and CWDM
> spacing.
>
> The multi-level analog signaling is conceptually appealing. But you are
> correct that it may not be practical in the field. Multi-level signaling
> requires very linear lasers and detectors. Does anybody know who laser
> linearity degrades with age and temperature?? How susceptible is
> the system
> to optical noise, etc etc
>
> Bill
>
>
> -------------------------------------------
> Bill St Arnaud
> Director Network Projects
> CANARIE
> bill.st.arnaud@xxxxxxxxxx
> http/tweetie.canarie.ca/~bstarn
>
>
>
>
>
>
>
> > -----Original Message-----
> > From: David W Dolfi [mailto:dolfi@xxxxxxxxxxxxxxxx]
> > Sent: Wednesday, May 05, 1999 9:29 PM
> > To: bill.st.arnaud@xxxxxxxxxx
> > Cc: stds-802-3-hssg@xxxxxxxx; dolfi@xxxxxxxxxx
> > Subject: FW: Why 10xGbE may be a bigger market than 1xGbE - a
> customer's
> > perspective
> >
> >
> >
> > To: Bill ST Arnaud/CANARIE
> >
> > From: David Dolfi/Hewlett-Packard
> >
> > Subject: DWDM in "gopher" fiber
> >
> > Date: 5/5/99
> >
> > Your message to the GbE high speed study group (below) was forwarded
> > to me by one of my engineers. In it you make the point that wide WDM
> > (WWDM) might not work on "gopher bait" fiber. As an advocate of WWDM
> > for 10 Gb Ethernet, I would like to understand why this is the case.
> > I would think that fiber which is damaged or overspliced would work
> > better at the lower baud rate (3.125 GBd) of 4 channel WWDM than at
> > the 4X higher baud rate (12.5 GBd) of serial 10 Gb Ethernet.
> >
> > Please remember when responding that the 4 channel WWDM we are
> > advocating
> > for 10 Gb Ethernet has the following properties:
> >
> > 1. The targeted link lengths are 10 km or less, so no EDFA's are
> > required.
> > 2. The 4 wavelengths we are using are in the 1300 nm region of the
> > spectrum (NOT 1550) and have channel spacings of 15-20 nm.
> >
> > Finally, I would think that fiber which is overspliced and/or
> > damaged would have a high loss, and I'm puzzled why a multi-level
> > coding scheme (which typically requires a high S/N ratio) would be
> > suitable for such fiber while WWDM would not.
> >
> > Am I missing the point here? Please respond!
> >
> > Dave Dolfi
> > email: dolfi@xxxxxxxxxx
> >
> >
> > ------------- Begin Forwarded Message -------------
> >
> > From lbuckman@xxxxxxxxxxxxxxx Wed May 5 16:01:23 PDT 1999
> > Received: from hplms2.hpl.hp.com (hplms2.hpl.hp.com [15.0.152.33])
> > by saiph.hpl.hp.com (8.8.6 (PHNE_17190)/8.8.6 HPLabs
> > Workstation) with ESMTP id QAA29223;
> > Wed, 5 May 1999 16:01:22 -0700 (PDT)
> > Received: from hplms26.hpl.hp.com (hplms26.hpl.hp.com [15.255.168.31])
> > by hplms2.hpl.hp.com (8.8.6/8.8.6 HPLabs Hub) with ESMTP id
> > QAA27579;
> > Wed, 5 May 1999 16:01:21 -0700 (PDT)
> > Received: from hplex3.hpl.hp.com (hplex3.hpl.hp.com [15.0.152.178])
> > by hplms26.hpl.hp.com (8.9.1a/HPL-PA Relay) with ESMTP id
> QAA02543;
> > Wed, 5 May 1999 16:01:20 -0700 (PDT)
> > Received: by hplex3.hpl.hp.com with Internet Mail Service (5.5.2232.9)
> > id <K1ZGX8C1>; Wed, 5 May 1999 16:02:47 -0700
> > Message-ID:
> <3BB8E8BD7405D211AE0C00A0C9B6B0BF01BF3098@xxxxxxxxxxxxxxxxx>
> > From: "Buckman, Lisa" <lbuckman@xxxxxxxxxxxxxxx>
> > To: "DOLFI, DAVE (HP-PaloAlto,unix3)" <dolfi@xxxxxxxxxx>,
> > "LEMOFF, BRIAN (HP-PaloAlto,unix3)" <lemoff@xxxxxxxxxx>
> > Subject: FW: Why 10xGbE may be a bigger market than 1xGbE - a
> customer's p
> > erspective
> > Date: Wed, 5 May 1999 16:02:43 -0700
> > MIME-Version: 1.0
> > X-Mailer: Internet Mail Service (5.5.2232.9)
> > Content-Type: text/plain;
> > charset="iso-8859-1"
> > Content-Transfer-Encoding: 8bit
> > X-MIME-Autoconverted: from quoted-printable to 8bit by
> > saiph.hpl.hp.com id QAA29223
> > Content-Length: 5031
> > Status: RO
> >
> > Dave and Brian,
> > Here is a message against WWDM...
> >
> > Lisa
> >
> >
> > -----Original Message-----
> > From: Bill St. Arnaud [mailto:bill.st.arnaud@xxxxxxxxxx]
> > Sent: Wednesday, May 05, 1999 2:44 PM
> > To: David_Law@xxxxxxxx; stds-802-3-hssg@xxxxxxxx
> > Subject: Why 10xGbE may be a bigger market than 1xGbE - a customer's
> > perspective
> >
> >
> > All:
> >
> > I have been watching the 802.3 high speed study group developments with
> > interest. This is a technology development we have been pushing
> > the vendors
> > for almost 2 years now. We are aware of a number of
> proprietary solutions
> > that have been promised to us later this year. So I hope the
> study group
> > moves quickly to develop a common standard.
> >
> > We are deploying an optical Internet across Canada
> www.canet3.net. We are
> > currently using POS, but we have identified a lot of
> deficiencies with POS
> > and we believe that "simplex, auto-negotiation 10xGbE" might be a better
> > solution for medium/long range WANs up to 1000 km. Here are
> the reasons:
> >
> > 1. There is a lot of "gopher bait" fiber in the ground that
> the carriers
> > can't use for traditional SONET networks. Gopher bait fiber is
> > usually fiber
> > that has been damaged, or has too many splices and will not
> > support OC-48 or
> > OC-192 clock rates. As well, a lot of carriers in the 80s deployed ZDSF
> > fiber that was optimized for TDM traffic and will not support
> > WDM. So these
> > carriers are leaving this stuff in the ground and are now
> deploying NZDSF
> > fiber that can support DWDM
> >
> > Ad hoc estimates we have heard is that about 30% of the installed
> > fiber base
> > is gopher bait.
> >
> > We believe that auto -negotiation 10xGbE would be the perfect
> protocol for
> > such a fiber, even if we use inefficient 8b/10b coding.
> >
> > 2. With auto-negotiation we could use the gopher bait fiber with long
> > repeater spacing, perhaps starting off with clock rates at less
> > than 1xGbE.
> > As demand for bandwidth increases we could install intermediary
> > transceivers
> > (and/or EDFA or PDFA) to support a faster clock rate. It allows
> > us to have
> > small upfront capital cost and incrementally grow the network
> bandwidth as
> > demand warrants rather than installing an expensive up front
> OC-192 system
> > and highly conditioned fiber.
> >
> > Auto-negotiation might also compensate for PMD changes as it varies
> > throughout the day.
> >
> > 3. A practical as possible we believe that 10xGbE should be
> made a simplex
> > protocol. The Internet, itself is fundamentally a
> > unicast/simplex network.
> > In the Internet today there is significant asymmetric traffic
> > loads between
> > the Tx/Rx paths -in some case the asymmetric variance can be
> > 16:1. As such
> > we are deploying links where they may be more Tx paths than Rx
> > paths. Since
> > there is no guarantee that there will an Rx for every Tx, as much as
> > possible auto-negotiation protocols, laser safety signaling
> should be done
> > at the IP layer. Maybe the PHY should be IP addressable and
> > managed at the
> > IP layer?
> >
> > 4. We would hope that the study group would also look at some of the
> > proposed SDL protocols as increasingly many optical links will
> run on data
> > and bit rate transparent networks. The overhead associated with
> > CRC, etc is
> > not necessary any more.
> >
> > 5. We believe that either a serial TDM, or multi-level signal
> > protocol would
> > be the best as WWDM might not work very well on gopher bait fiber
> >
> > Looking forward to see the results of your deliberations. We are in the
> > process of setting up a 10xGbe 50 km and 500km test facility
> for those who
> > might be interested in experimenting with some of the 10xGbE concepts
> >
> > Also, we have some background papers for those who are
> interested in our
> > program on optical Internet technologies
> >
> >
> > Bill
> >
> > -------------------------------------------
> > Bill St Arnaud
> > Director Network Projects
> > CANARIE
> > bill.st.arnaud@xxxxxxxxxx
> > http://tweetie.canarie.ca/~bstarn
> >
> >
> >
> >
> >
> >
> >
> > > -----Original Message-----
> > > From: owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx
> > > [mailto:owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx]On Behalf Of
> > > David_Law@xxxxxxxx
> > > Sent: Wednesday, May 05, 1999 11:28 AM
> > > To: stds-802-3-hssg@xxxxxxxx
> > > Subject: Higher Speed Study Group e-mail reflector archive
> > >
> > >
> > >
> > >
> > > All,
> > >
> > > After discussions with the IEEE it has been agreed that the
> > > reflector archive no
> > > longer needs to be in a password protected area of the web site.
> > > Due to this the
> > > archive has been moved to a new URL:-
> > >
> > > http://grouper.ieee.org/groups/802/3/10G_study/email/thrd1.html
> > >
> > > Please update any boookmarks you may have of the old location.
> > >
> > > Thanks,
> > > David Law
> > >
> > > |=========================================|
> > > | David Law |
> > > | Vice-Chair IEEE 802.3 |
> > > | 3Com |
> > > | 3Com Centre, |
> > > | Boundary Way, |
> > > | Hemel Hempstead, |
> > > | Hertfordshire, HP2 7YU, |
> > > | United Kingdom |
> > > | Phone: +44 1442 438060 |
> > > | Fax: +44 1442 438333 |
> > > | E-Mail: David_Law@xxxxxxxx |
> > > |=========================================|
> > >
> > >
> > >
> > >
> >
> > ------------- End Forwarded Message -------------
> >
>
>