LVD SCSI-April 18th meeting Comments

NHarris at NHarris at
Thu Apr 17 09:41:00 PDT 1997

* From the SCSI Reflector (scsi at, posted by:
* NHarris at
Form: Memo
Text: (97 lines follow)

I respect the right for all who use these reflectors to post their
opinions.  However, I must take serious exception with a couple of 
made in your recent post. My reply is on behalf of Adaptec with whom I
believe you insinuate in your comments although not using our name
directly. I am uncertain as to whether your comments reflect the
position of Unitrode or whether they are just a personal emotional
flame.  However, it is easy for me to gather that if these comments do
represent Unitrode, it is not difficult to see why few of your
terminator parts are in use by Adaptec.

First: I believe any notion that Adaptec was not a continuing
participant in the Ultra 2 effort, lacked any due diligence in effort
or anything else you might try to imply from such a statement is utter
nonsense.  The record is clear on the participation of Adaptec in the
development of the Ultra 2 I/O cell design.  I might add that Wally
Bridgewater from  Adaptec and Ting of Qlogic, both of who are key
silicon people from their respective companies made prominent
contributions to this effort. I believe that technical people from
all companies who participated in this effort would take serious offense
to such nonproductive comments.   If you don't remember, Adaptec hosted
one of the early SPI-2 working groups at Milpitas in 1995 in which much
of this early discussion about asymmetry for driver design took place.
Since many had not completed a basic I/O cell design, this approach
represented the lesser of  the two evils, with the alternative requiring 
additional transistors.
We have now had an opportunity to produce test parts and do a proper
characterization.  We now know that alternative approaches present a
better platform for the future.   In light of these facts, I am a bit
annoyed at your statement alleging nonparticipation on our part.   
Perhaps you've been outside waiting on the spaceship during the entire time.

Adaptec has spent a great deal of time, money and resources to provide a 
complete characterization of
this new technology.  I'm sure other companies have likewise expended 
resources.  Perhaps in your zeal to get to market you've forgotten the 
process of due diligence. Not everything can be fixed with bandaids.
We're familiar with each and every one of the bandaids that you suggest and 
know there are reasons why
these would not work in the current infrastructure.  

Second to state that April 18th is a political event crosses the line. 
Adaptec has presented legitimate concerns about the present design to the
appropriate forum. We did so at the earliest possible time dictated by
the time to develop the test vehicles and hardware which I might add we
developed at great expense.  Although that data may not have been to
your liking, it did shed light on the fact that our choice of design
approach was a poor choice given the problems we know must be conquered
at higher speeds.   Adaptec chose to be responsible to the
marketplace and technology, even in the face of the insinuations from
those such as Unitrode who are anxious to try to establish themselves as
front running players in SCSI.  To the contrary, the public record of
what had been shown to the SCSI committee up until the March San Diego
meeting was noticeably absent of data supporting the functionality of
real Ultra 2.  April 18th exists precisely to bring the answers to
questions against our technical claims.  By the time this is over, there
may very well be more due diligence on the alternative than currently exists
for the status quo.  Of course if such due diligence has been shown
to customers, then all those customers who have adopted LVD can sleep
easy at night knowing that their SCSI investment is safe. I don't
believe that diligence has been performed, either in public or private

Unless you live on a different planet, Adaptec is the market leader by a
wide margin in SCSI revenue. The incremental amount of revenue to be
gained by anyone who takes any early market share of Ultra 2 is
insignificant to us as compared to the amount of damage that could be done 
on bad customer experience with Ultra 2.  We all make a lot of money on
SCSI.  Our customers have a significant investment in this technology. 
We want to continue to provide new generations of SCSI that protect that
investment for as long as is practically possible.  We are anxious to do
anything appropriate to evolve the technology around solutions that have
longevity.  This is our overriding concern.

We know that current state of knowledge in the science of differential
I/O pad design clearly demonstrates that a symmetric design is
superior.  The choice of asymmetry has been around for quite a while. 
It's no accident that its NEVER been employed for high speed signaling
in any other technical standard.

It is indeed a shame that you won't be in attendance at the April 18th 
political convention.

Norman H. Harris
Staff Advisory Engineer
Principal Member X3T11/X3T10
Adaptec Inc.
691 S. Milpitas Blvd
Milpitas, Ca. 95035
email: nharris at,nharris at or
n_harris at
PH. (408) 957-2230 fax (408) 957-7990
Use Proportional Font: true
Attachment Count: 0

X-NVL-Content-Type: UNKNOWN
X-NVL-Content-Typename: UNKNOWN
X-NVL-Content-Charset: X-IBM-437
X-NVL-Content-Filename: ATTRIBS.BND
X-NVL-Content-Transfer-Encoding: X-UUENCODE

begin 777 ATTRIBS.BND
M`0#1"A4+`0`6"UT+`0!>"Y at +`0"9"]X+`0#?"R<,`0`H#&T,`0!N#+0,`0"U


* For SCSI Reflector information, send a message with
* 'info scsi' (no quotes) in the message body to majordomo at

More information about the T10 mailing list