Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: Going the distance




     Ed,
     
     I don't agree that the laser is always the "big ticket item in a 
     transceiver".  This may be true in long-haul telecom (and it may be 
     true for 10GbE serial), but in both 1000-SX and 1000-LX transceivers, 
     the laser (i.e. VCSEL or FP) accounts for a small fraction of the 
     overall transceiver cost. It is by no means obvious that a serial 10G 
     FP transceiver, will cost less than a 4x2.5G WWDM transceiver.  
     Regardless of this, I don't think it will hurt the process to have a 
     2km objective in addition to a 10km objective.
     
     -Brian Lemoff
      HP Labs


______________________________ Reply Separator _________________________________
Subject: RE: Going the distance
Author:  Non-HP-ecornejo (ecornejo@lucent.com) at HP-PaloAlto,mimegw2
Date:    7/1/99 7:33 AM


I believe 2km over SMF will cover a majority of Campus LAN applications, and
therefore a good place to start. I base the 2km distance on FDDI cabling
structure, GE survey presented by Chris D., and ISO's 2-3km spec. 

The serial proposal is two fold, one for shorter reach(2km), and one for
intermediate reach(15km). The 2km approach uses an uncooled unisolated
Fabry-Perot laser; this should be considerably less cost than 4 uncooled
unisolated DFBs(i.e. WWDM approach). I know there are other factors here
like packaging, and electronics, but I believe most folks would agree that
the laser is the biggest ticket item in a transceiver. If the minimum
distance is >2km, you will be excluding a potentially lower cost solution.
Unnecessarily IMHO because it covers a majority of your applications. 

For the same reasons I want 2km, I would not want to exclude anyone at the
longer distances. Therefore, I would support 2km, and 10km as the two
distances for SMF. Also, in my view the two distance proposals would be the
same footprint and electrical interface to the PCS or MAC, so it is not a
major hassle having seperate laser spec's for the PMD.

Going further distances beyond 10km, or 15km is never a problem; it is just
how much customers are willing to pay.

I concur with my colleague from HP that we should meet customer demands, but
also consider technology capabilities and costs. Lets not be too exclusive
at this early juncture. It is always easier to increase our distances than
to throttle them back in case we run into unforseen problems with MMF or SMF
at 10G.

Ed-Lucent
> ----------
> From:
> BRIAN_LEMOFF@HP-PaloAlto-om16.om.hp.com[SMTP:BRIAN_LEMOFF@HP-PaloAlto-om16
> .om.hp.com]
> Sent:         Wednesday, June 30, 1999 7:36 PM
> To:   pbottorf@nortelnetworks.com
> Cc:   Bruce_Tolley@3com.com; hfrazier@cisco.com;
> rtaborek@transcendata.com; stds-802-3-hssg@ieee.org
> Subject:      Re: Going the distance
> 
> 
>      Paul:
>      
>      I believe that the WWDM approach that we have presented, when used 
>      with uncooled, unisolated, reduced spec (on SMSR, linearity, RIN)
> DFBs 
>      will cost equal or less than the serial 10G FP laser approach that 
>      Lucent has proposed for 2km, and significantly less than the isolated
> 
>      DFB approach that they have proposed for 15km.  This would support 
>      ~300m on 62.5 micron fiber, and ~10km on SMF.  To push this approach 
>      to 15km may involve tightening the SMSR or RIN spec which will 
>      significantly increase the cost (since much of the low-cost nature of
> 
>      the approach depends upon using low-cost DFBs).  
>      
>      I realize that there is an ongoing debate about relative cost, but it
> 
>      would be a shame to set the objectives at 2km and 15km, when there is
> 
>      a potentially very low-cost solution that can go 10km. I have no 
>      problem with a 2km and 10km objective, since I believe we can satisfy
> 
>      both with a single cost-competitive solution.
>      
>      I realize that Rich Taborek and others have criticized PMD suppliers 
>      (such as HP and Lucent) on this reflector for suggesting objectives, 
>      based on what each of our solutions can handle, but I don't think we 
>      can afford to ignore it either.  I don't deny that I favor 300m on 
>      installed base MMF and 10km on SMF in large part because that is what
> 
>      our WWDM module can support.  I have no problem backing off on these 
>      (say to 200m and 5km) if it allows alternative PMDs to be considered.
> 
>      Lucent wants 2km (not 3km) to be an objective so that their serial FP
> 
>      laser module is not excluded.  I'd like 10km (and not 15km) for the 
>      same reason (although I believe we're still competitive at 2km).
>      
>      Distance objectives should reflect what the customers need, but
> should 
>      be influenced by what the available technologies can achieve.
>      
>      -Brian Lemoff
>       HP Labs
>      
>      
>      
>      
>      
> 
> 
> ______________________________ Reply Separator
> _________________________________
> Subject: Re: Going the distance
> Author:  Non-HP-pbottorf (pbottorf@nortelnetworks.com) at
> HP-PaloAlto,mimegw2
> Date:    6/30/99 3:24 PM
> 
> 
> 
> Bruce:
> 
> I support your outlook. We already have 5-10 km with gigabit. Supporting
> these distances is necessary to support the installed base. Since the
> technology cut-off point is around 20 km and since a standard MAN distance
> is 15 km I'd prefer a slightly more aggressive objective of 15 km. This
> would still give reasonable design margin. I also believe lower cost
> technology exits below 2.5 km motivating a 2 km specification. We could
> either specify a single 15 km objective which covers all the applications
> with a higher component cost or 2 km and 15 km each with a different price
> point.
> 
> I would like to see further extension into the MAN. GigE is selling day
> with reaches over 50 km. The next standard distance is 40 km. I've heard
> this distance was chosen to match the spacing of microwave towers. With
> the
> 40 km distance I believe we would have all that is necessary to build
> Metro
> networks. 
> 
> The three standard distances on SMF would be 2, 15, and 40 km. Each of
> these ranges I believe falls into a different technology with a different
> price point. If anyone has a technology which can do all three at the 2 km
> price then all the better.
> 
> Paul
> 
> At 09:13 AM 6/30/99 -0700, Bruce_Tolley@3com.com wrote:
> >
> >
> >
> >The point has been made before that today customers are already going 5
> to
> 10 Km
> >with 1000BASE-LX.  There should be no debate that it is a market
> requirement to
> >go 5 to 10 km with 10 GbE.
> >
> >While I am willing to consider accepting  a conservative 2 to 3 km goal
> as
> the
> >official goal of the project,  we need to acknowledge that this is a
> >conservative goal and, as we get on with the work of the project , we
> should
> >investigate whether we can stretch this goal..
> >
> >Bruce Tolley
> >3Com Corporation
> >
> >
> >
> >
> >Rich Taborek <rtaborek@transcendata.com> on 06/29/99 05:01:32 PM
> >
> >Please respond to rtaborek@transcendata.com
> >
> >Sent by:  Rich Taborek <rtaborek@transcendata.com>
> >
> >
> >To:   Howard Frazier <hfrazier@cisco.com>, HSSG
> <stds-802-3-hssg@ieee.org>
> >cc:    (Bruce Tolley/HQ/3Com)
> >Subject:  Re: Going the distance
> >
> >
> >
> >
> >
> >Howard,
> >
> >I will gladly accept your suggestion of removing the portion of the
> >motion in parenthesis as a  friendly amendment post-haste given your
> >support of this motion as a seconder.
> >
> >- Rich
> >
> >Howard Frazier wrote:
> >
> >> The more I think about it, the more I like the idea of adopting the
> >> 802.3z link distance objectives for 10 Gig. We should remember that
> >> we are still in the study group phase, trying to scope out a project.
> >> We can always adjust the objectives later.
> >>
> >> Let me therefore state my support for Rich's proposed motion:
> >>
> >> > "Support the premises cabling plant distances as specified in
> >> > ISO/IEC 11801"
> >> >
> >> > The distances supported in ISO/IEC 11801 are:
> >> >
> >> > 100 m for horizontal cabling (applicable to copper, MMF, SMF)
> >> >
> >> > 550 m for vertical cabling: (applicable to SMF and possibly MMF)
> >> >
> >> > 2-3 km for campus cabling: (applicable to SMF)
> >>
> >> Rich, I would encourage you to drop the parenthesis.
> >>
> >> If we adopt this objective, we can make progress on the rest of the
> >> work we need to do as a study group.  As was demonstrated in 802.3z,
> >> we will need to review the objectives periodically, and revise them
> >> if there is consensus to do so.
> >>
> >> Howard Frazier
> >> Cisco Sytems, Inc.
> >
> >-------------------------------------------------------------
> >Richard Taborek Sr.    Tel: 650 210 8800 x101 or 408 370 9233
> >Principal Architect         Fax: 650 940 1898 or 408 374 3645
> >Transcendata, Inc.           Email: rtaborek@transcendata.com
> >1029 Corporation Way              http://www.transcendata.com
> >Palo Alto, CA 94303-4305    Alt email: rtaborek@earthlink.net
> >
> >
> >
> >
> >
> >
> >
> >
> >
> Paul A. Bottorff, Director Switching Architecture
> Bay Architecture Laboratory
> Nortel Networks, Inc.
> 4401 Great America Parkway
> Santa Clara, CA 95052-8185
> Tel: 408 495 3365 Fax: 408 495 1299 ESN: 265 3365
> email: pbottorf@NortelNetworks.com
>