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

RE: Chapter 46: preamble length




No Brad, I don't think this is accurate.  In the currently defined PCS's for
10GE we only adjust IPG for clock compensation.  It is true that adjusting
preamble was another option, but in every instance we chose adjusting IPG
and specified it that way so the option no longer exists.  The argument for
still having the XGMII/RS/MAC be tolerant of a reduced IPG was because:
a) historically they always have been,
b) some PHYs have taken advantage of that, even in FDX mode (e.g.
1000BASE-X),
c) we frequently add new PHYs over time we wanted to leave the option
available in case it would be important for some future PHY technology. 

Judging from the email stream I would say this is still controversial.  All
it takes to change it is a comment and a 75% majority.  

Personally I don't think it really matters one way or the other in theory.
In practice I think this could be characterized as "bug bait" or an
interoperability time bomb.  The fact that preambles could be shortened is
easy to miss.  Implementations that are not tolerant of it could be in the
field for years before someone else comes out with an implementation that
does shorten the preamble.  Then you've got a real mess.

Steve

-----Original Message-----
From: Booth, Bradley [mailto:bradley.booth@intel.com]
Sent: Wednesday, March 28, 2001 12:04 PM
To: stds-802-3-hssg@ieee.org
Subject: RE: Chapter 46: preamble length



There are a number of places within 802.3ae that synchronization between
clock domains may require some form of compensation.  The WIS was just an
example of a part of the structure that may require clock compensation.
Generally, we permit shortening of the IPG and preamble to assist with this
compensation so that the data remains untouched.

I'm not sure if that's helping you understand it any better.

Cheers,
Brad

		-----Original Message-----
		From:	Sanjeev Mahalawat [mailto:sanjeev@cisco.com]
		Sent:	Wednesday, March 28, 2001 1:46 PM
		To:	Booth, Bradley; stds-802-3-hssg@ieee.org
		Subject:	RE: Chapter 46: preamble length

		Booth,

		At 10:08 AM 03/28/2001 -0800, Booth, Bradley wrote:

		>In 10GbE, truncation of the preamble can occur due to
		>the asynchronous timing associated with the WAN PHY.

		PCS + PMA = LAN PHY

		PCS + WIS + PMA = WAN PHY

		The PCS and PMA processes are common in both LAN and WAN
Phy.
		The only difference is extra WIS. And all WIS is supposed to
do is 
		take the MAC frame (Preamble + Data + IPG) and encapsulate
it
		in STS-192c frame without any involvement, visibility into
the MAC framing
		process. So, I am not getting where and why the WIS should
open the MAC
		frames and try to change it. Could you please explain it a
bit more?

		Thanks,
		Sanjeev 


		>
		>Cheers,
		>Brad
		>