Re: 1310nm vs. 1550nm window for 10GbE

[David W Dolfi <dolfi@xxxxxxxxxxxxxxxx>]

Dear Brian,

I very much agree with your comments about the 1550 window.  From the
point of view of the transceiver, there are a lot of positives.  You
mentioned the bigger eye safety margin, which helps a lot with respect
to the power budget.  There is also the voltage requirement on the
laser drivers, which decreases with longer wavelength.

I think that the main issues with 1550 come from the fiber side of 
things.  Unless dispersion shifted fiber were used, dispersion would be
an issue at 1550, and dispersion related penalties (ISI, Mode Pertition
Noise) would limit the link lengths more severely than in the 1300 nm
window.  The other fiber related problem has to do with the use of multimode
fiber, which is still a significant issue for 10 Gb Ethernet, given both
the large amount of multimode fiber installed and the recent introduction
of newer, higher bandwidth multimode fiber by several vendors.  To my
knowledge, there are really no specs around for multimode fiber at 1550,
and I don't know of anyone who is designing multimode fiber specifically to 
work there.  I would be interested in comments from the fiber folks regarding
the use of multimode fiber at 1550 nm.


Dave Dolfi
Hewlett-Packard
email: dolfi@xxxxxxxxxx