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

Re: [10GBT] FW: Contribution for May: Clause 55: Link Segment Specification



Chris-

You cite 1000BASE-T as your style guide
Unfortunately, clause 40 is probably our 2nd worst clause from an editorial standpoint.

Geoff

At 02:29 PM 5/5/2004 -0400, CDimi80749@aol.com wrote:
Geoff,
 
Thanks for the review.....
 
Geoff Thompson wrote>>
>>P1L5
The text:
"Each of the four pairs supports an effective data rate of (2500) Mbps in each direction simultaneously."
...could easily be misunderstood to imply that there is a separate data path over each of the 4 pairs and that there is PHY level pair bonding as in EFM Cu. As I understand it this is not the case, rather that group coding across the pairs is being done.
I would suggest just dropping the sentence unless better text can be developed.
Chris DiMinico Reply>>
 
I've utilized the language of 1000BASE-T whenever possible as the path of least resistance.
 
See 40.7
Link segment characteristics
 
"Each of the four pairs
supports an effective data rate of 250 Mbps in each direction simultaneously."

 
 
Geoff Thompson wrote>>
P1L7
The sentence:
"The term 'duplex channel' will be used to refer to a single channel with full duplex capability."
should not be necessary as this is already a defined term in 802.3, see 1.4.106 in 2002
Further, if you look at the rest of 802.3 and ISO/IEC 11801 the use of the term "channel" refers to the full channel, not the per pair channel. Not a good idea to invent new usage on a per clause basis.
 
Chris DiMinico Reply>>
 
I've utilized the language of 1000BASE-T whenever possible. I've included
the usage of "duplex channel" for consistency with 1000BASE-T.
 
See 40.7
Link segment characteristics
 
"The term “duplex channel” will be used to refer to a single
channel with full duplex capability."

 
Geoff Thompson Wrote>>
P1L14
The text:
"The cabling system used to support 10GBASE-T requires 4 pairs of ISO/IEC 11801 Class E or Class F balanced cabling with a nominal impedance of 100 ohms."

(By the way, it is 802.3 convention to NOT use Greek symbols in drafts so that meaning is not lost when pasting text from drafts across to other applications in ASCII. The term "ohms" should be used. The conversion to Greek symbols is a task for the publications editor.)

Should be changed to:
"The cabling system used to support 10GBASE-T requires ISO/IEC 11801 4pair Class E or Class F balanced cabling with a nominal impedance of 100 ohms."
Rationale:
4 pair cable is required. 2 instances of 2 pair cable (which is "4 pairs") is not assured of doing the trick.
 
Chris DiMinico Reply>>
 
I've utilized the language of 1000BASE-T whenever possible.
 
see
40.7.1 Cabling system characteristics
The cabling system used to support 1000BASE-T requires 4 pairs of Category 5 balanced cabling with a
nominal impedance of 100 ohms.

 
I'll accept your change to "4 pair" in order to eliminate the possibility of 2 instances of 2 pair....
 
Geoff Thompson Wrote>>
P1L19
The text:
"10GBASE-T uses a star topology with Class E or Class F balanced cabling used to connect PHY entities."
is not correct. It should be:
"10GBASE-T uses point-to-point with Class E or Class F balanced cabling used to connect PHY entities."
Since this is not a repeatered network the topology of interconnecting the bridges is way out of scope for this project. The scope of the project is a point-to-point link.
Chris DiMinico reply>
 
I see your point but...network topologies comprised of point-to-point links
are often referred to as star topologies (distinguished from bus). Star topology is used here in
the context of deployment over Class E or Class F balanced cabling.
 
see 40.7.1 Cabling system characteristics
a) 1000BASE-T uses a star topology with Category 5 balanced cabling used to connect PHY entities.
 
 
Geoff Thompson wrote:>>

P1L32
The text:
"segment ... and ... segment ... will provide a reliable medium."
should be:
"segment ... and ... segment ... will provide reliable media."
-OR-
"segment ... or ... segment ... will provide a reliable medium."        (PREFERRED)
Chris DiMinico Reply>>>
 
I'll change to "segment ... or ... segment ... will provide a reliable medium." 
 
Thanks.....
 
Chris
 
 

At 07:16 PM 5/4/2004 -0700, Booth, Bradley wrote:
Dear Task Force members,
 
This will be posted on website in a couple of days, but here's a copy now for you to review.  Thanks to Chris for putting this together.
 
Regards,
Brad
 
-----Original Message-----
From: CDimi80749@aol.com [mailto:CDimi80749@aol.com]
Sent: Tuesday, May 04, 2004 8:50 PM
To: Booth, Bradley
Cc: SKasturia@teranetics.com
Subject: Contribution for May: Clause 55: Link Segment Specification

Brad,
 
Please find attached contribution D1.0 Clause 55 Link Segment specification for posting.
 
Regards,
 
Chris DiMinico
editor: Clause 55.X Link Segment
cdiminico@ieee.org
MC Communications