RE: [EFM] 10 km objective, Optical PMD perspective

["Vipul Bhatt" <vipul.bhatt@xxxxxxxxxxx>]

Rob,

I am afraid we don't have that luxury. What you call overspecifying
is the job we accepted. No, a laser doesn't have to be specified to
be FP, but we do have to specify the maximum RMS spectral width.

Our objectives include providing a family of physical layer
specifications, two of which must support at least 10 kilometers of
link length. That, in turn, implies defining a set of distinct PMD
types and at least identifying and referring to selected fiber
types. A set of specifications makes a PMD type distinct. RMS
spectral width, for example, is a required PMD specification because
we can't determine link distance without it.

We are also guided by the principle of choosing the fewest possible
number of distinct PMD and fiber types. It's a consequence of our
commitment to minimize equipment, operation and maintenance costs.
We made that commitment when we requested PAR approval.

So we have no choice but to operate within that domain and make
difficult decisions about what our MINIMUM set of PMD/fiber types
should be. Implementations and augmentations that are supersets of a
standard are likely and expected. For example, 1000BASE-LX links
that support 10 kilometers operation are widely implemented, after
802.3z specified an operating range of upto 5 kilometers on
singlemode fiber.

Regards,
Vipul

vipul.bhatt@xxxxxxxxxxx
408-542-4113

=====================

> -----Original Message-----
> From: owner-stds-802-3-efm@majordomo.ieee.org
> [mailto:owner-stds-802-3-efm@majordomo.ieee.org]On Behalf
> Of Carlisle,
> Robert S
> Sent: Wednesday, October 03, 2001 7:52 AM
> To: stds-802-3-efm@ieee.org; 'Vipul Bhatt'
> Cc: Du, Xueping; Feng, Weiwei; Jay, John; Kunzi, Anne;
> Musgrove, Kendall
> Subject: RE: [EFM] 10 km objective, Optical PMD perspective
>
>
>
> Vipul,
>
> I think you have uncovered a potential "rat hole" that we
> may me leading ourselves in and that is to overspecify
> the PMD in EFM.  Why not just specify the wavelength
> allocation over single mode fiber?  Why does it have to
> be an FP?  Using an alternative laser type not only
> extends the reach, but enables other fiber types besides
> G.652.  So why would I use a dispersion shifted fiber
> i.e. G.655?  Many reasons....
> *	Upgradeability for downstream data rates way above
> 1.25 Gbps...now we can think about all those HDTV
> channels in the future or downloading a 7 Gb movie in a
> matter of seconds.
> *	Allows the use of DML without dispersion
> compensation in the HE for analog video.
> *	Reduces customer fiber inventory between metro and
> access fiber types.
> *	DWDM possibilities for FTTB
>
> My point is simple...the standard should avoid
> overspecifying the PMD to allow network engineers the
> flexibility to use the right electronic, opto-electronic,
> and fiber mix to extend the data rate and/or distance to
> enable the applications of the future.
>
> Best regards,
>
> Rob Carlisle
>
>
> > ----------
> > From: 	Vipul Bhatt[SMTP:vbhatt@xxxxxxxxxxx]
> > Sent: 	Tuesday, September 25, 2001 8:01 PM
> > To: 	stds-802-3-efm@ieee.org
> > Subject: 	[EFM] 10 km objective, Optical PMD perspective
> >
> >
> > Dear colleagues,
> >
> > Our distance objective for optical links is ">= 10 km".
> We need to
> > now examine that in greater detail, and make a
> difficult decision.
> >
> > When converting this requirement into Optical PMD
> specifications, it
> > is becoming clear that Fabry Perot (FP) lasers will
> dispersion-limit
> > the distance to somewhere between 10 and 13 kilometers,
> depending on
> > how aggressive we get with the specs. FP laser is the
> workhorse of
> > nearly all 1000BASE-LX transceivers. It is
> cost-effective and field
> > proven. Unfortunately, it won't support 20 kilometers.
> Throwing a
> > bigger power budget at the problem won't help. (Details
> are being
> > discussed on the P2P reflector.)
> >
> > To support longer distances, we have to resort to
> either temperature
> > controlling the FP laser, or specifying an additional
> PMD type with
> > more expensive or less proven lasers.
> >
> > Temperature control is worth investigating, but it has
> its own set
> > of issues, and we end up cost-penalizing the majority of (short)
> > links.
> >
> > Adding a PMD type sounds easy, but we should think
> again. We already
> > have 5 PMD types on the table - at least one copper,
> one temperature
> > extended LX, and three single-fiber. Adding one PMD
> type effectively
> > adds two PMD types, because head-end and subscriber-end need
> > opposite wavelength plans.
> >
> > We can make all Optical PMDs capable of supporting 20 km by
> > specifying expensive or less proven lasers, but that
> would deny the
> > majority of links the benefit of using FP lasers.
> >
> > This begs two questions: Will a number between 10 km
> and 13 km be an
> > acceptable limit? What percentage of total deployments
> are likely to
> > exceed this limit?
> >
> > Using regenerators for longer distance links is
> expensive. But so is
> > the inclusion of 7 PMD types in a standard.
> >
> > Thanks,
> > Vipul
> >
> > vbhatt@xxxxxxxxxxx
> > 408-542-4113
> >
> >
> >