| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Howard,
Information I have on oscillators/tolerances & pricing aligns well with your
findings. Although there may be up to a 2-3x price factor between 100 & 20
ppm parts, the absolute difference is in the 'used hubcap' realm. Something
that could be washed out over time with the large volumes of Ethernet ports.
Cross-linking to Henning's thread on reference clocks, I could then envisage
a 10GE LAN PHY card with a 156.25M +/-100ppm crystal or a 10GE WAN PHY
card with a 155.52M +/-20ppm crystal. I'm wondering what your view is for a
dual-mode UniPHY card. Obviously a 20ppm crystal also satisfies the LAN spec,
but a 730kHz swing is more than a crystal can be tuned, so it sounds like both
oscillators would be needed. That's about an extra square inch of board space, etc.
Thoughts?
...Dave
-----Original Message-----
From: Howard Frazier [SMTP:hfrazier@xxxxxxxxx]
Sent: Monday, June 05, 2000 6:26 PM
To: stds-802-3-hssg@xxxxxxxx
Subject: Re: 20 ppm clock tolerance for WIS
Praveen,
I appologize for the delay in responding to your question. I wanted to
be absolutely sure of the facts before I replied.
I have sent out 5 requests for quotes for oscillators, and thus far I
have received three responses.
The first two responses indicated a less than 10% price premium for a
20 ppm 155.52 MHz oscillator versus a 100 ppm oscillator with otherwise
identical specs. I can't talk about absolute price on this reflector,
but I can say that the price difference is very small. Truly, it is
lost in the noise as far as the overall cost for a 10 Gigabit interface
is concerned.
The third quote I received indicated a more substantial cost
difference. The 20 ppm oscillator cost twice as much as the 100 ppm.
However, the lead time for the 100 ppm oscillator was 20 weeks, whereas
the 20 ppm oscillator is available immediately. Once again, the
absolute price difference is a very small percentage of the overall
cost for a 10 Gigabit adapter.
My conclusion remains the same. Specifying 20 ppm is the right way to
go. The cost difference does not justify changing to 100 ppm, because
this change probably will compromise compatibility with existing OC-192
transponders and regenerators.
As to the jitter specifications, it is my understanding that the WAN
PHY proponents want to use all of the same optical components that are
being specified for the LAN applications of 10 Gigabit Ethernet. I
believe that the SONET jitter specifications do impose additional cost
on the optical components. It is unreasonable to burden the LAN
applications with the cost of meeting the SONET jitter specifications.
Therefore, while I am keeping an open mind on the topic, and I am
willing to consider other points of view, I believe that we should not
impose the SONET jitter specifications on the 802.3ae PMDs, and I don't
think that this will compromise compatibility with OC-192 transponders
and regenerators.
If I get any additional responses to my requests for quotes, I will
be happy to share them on the reflector.
Howard Frazier
Cisco Systems, Inc.
----- Begin Included Message -----
From praveen@xxxxxxxxxxx Tue May 30 15:38:01 2000
X-Sender: praveen@mailhost-sv
Date: Tue, 30 May 2000 15:56:05 -0700
To: gnicholl@xxxxxxxxx, hfrazier@xxxxxxxxx, stds-802-3-hssg@xxxxxxxx
From: Praveen Kumar <praveen@xxxxxxxxxxx>
Subject: 20 ppm clock tolerance for WIS
Mime-Version: 1.0
X-SMTP-HELO: vitesse.com
X-SMTP-MAIL-FROM: praveen@xxxxxxxxxxx
X-SMAP-Received-From: outside
X-SMTP-PEER-INFO: email.vitesse.com [206.216.176.6]
Gary, Howard:
For the benefit of those who could not make it to your presentation in
Ottawa, could you clarify some of the issues that you raise.
http://grouper.ieee.org/groups/802/3/ae/public/may00/nicholl_1_0500.pdf
You mention that the "cost difference between +/-20ppm and +/-100ppm
oscillators is a tiny fraction of the total cost of a 10GigE interface".
Could you perhaps substantiate this statement with quantitative input (some
real numbers). My understanding is that the +-100ppm tolerance was
specified only to keep the cost down (as the cost differential between a
20ppm solution and 100ppm solution is perceived to be significant).
You recommend using "LAN PHY jitter specs". This makes the WIS
incompatible with installed base SONET . This doesn't seem to meet your
goal of being compatible with installed OC-192 SONET
infrastructure. Please clarify.
-Praveen
----- End Included Message -----