| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
Dear
802.3ba, Here is recommended
characteristics and our input on
nAUI: The XLAUI / CAUI team
would appreciate any feedback in this area. To start the discussion Chris
Cole has already submitted the following input in Comment
295: Max module trace length 3” -- We recommend 1.5" for
QSFP, this 3" distance seems to only apply to the CFP. The following limits
are fine...
Thanks, Scott
Kipp From: Jeffery Maki [mailto:jmaki@xxxxxxxxxxx] Sent: Thursday, January 08, 2009 9:05 PM To: STDS-802-3-HSSG@xxxxxxxxxxxxxxxxx Subject: Re: [802.3BA] XLAUI / CAUI Ad Hoc To those concerned about
XLAUI/CAUI, In general, we are supportive of the
Piers Dawe presentation (dawe_03_0109.pdf) entitled “Compliance points for
XLAUI/CAUI with connector” to be given next week in We
believe the portion of the loss
budget for the system host board should be as large as possible, since the
overall solution includes the system host board and the module. In consideration of cost, the system
host board will have far greater incremental cost than the module when higher
quality PCB material is used. Cost
sensitive applications using 40G interfaces will employ many such interfaces,
which require fanning out signals to an array of 40G interfaces. This fanning out also places a
requirement for longer system host-board trace lengths or higher system
host-board loss, which again prompts the need to use higher-cost
lower-loss-rate PCB and/or
higher-cost manufacturing methods. We
wish to see the system host-board loss be 8.2 dB. However, we would continue to be
supportive of the XLAUI/CAUI specification down to 7.5 dB for the system
host-board loss. In addition, since the Tx and Rx traces are segregated, FEXT
is more important
than NEXT. Thus, we would
like to see the specification cover FEXT. Frequency dependent masks
should actually be specified for
both FEXT and NEXT. Sincerely, Jeffery ——————————————————————— Jeffery J. Maki,
Ph.D. Principal Optical
Engineer Juniper
Networks, Inc. Desk
+1-408-936-8575 Lab
+1-408-936-1169 (Please leave
messages by email.) FAX
+1-408-936-3025 www.juniper.net ——————————————————————— IEEE
802.3 voter, OIF voter, & EA alternate voter Member
of OSA, LEOS, & IEEE ——————————————————————— From: Ryan Latchman
[mailto:Ryan.Latchman@xxxxxxxxxx] 802.3ba
Colleagues, In order to include a
chip to retimed module specification in 802.3ba, application level input is
required on expected module trace lengths, host trace lengths, numbers of via,
etc. for 40GbE and 100GbE. It would be extremely useful if this information can
be provided in terms of loss. The XLAUI / CAUI team
would appreciate any feedback in this area. To start the discussion Chris
Cole has already submitted the following input in Comment
295: Max module trace length 3” Best
Regards, Ryan From: Ryan Latchman
Hi Ali,
Chris, This is an important
discussion which needs to get resolved quickly. I would like to ensure
that XLAUI / CAUI maintains its broad market applicability as a simple retimed
interface. I don’t think the current specification methodology prevents it
from being leveraged to build retimed modules. I’ve put together the
attached material to show how retimed interfaces were specified in the past
(namely XFI). In XFI, you’ll notice that the Before Connector and After
Connector specs are similar. 40/100GbE modules may have an analogous
situation, depending on their size and electrical characteristics.
If we need to change
the XLAUI / CAUI specification, we need solid contributions on what needs to
change. Happy
Holidays, Ryan From: Ali Ghiasi
[mailto:aghiasi@xxxxxxxxxxxx] Chris |