[STA Members] SAS 2.0 market requirements -- minutes from Aug. 15 call

Alice Tate alice at scsita.org
Thu Aug 24 10:13:20 PDT 2006

STA members,

Below please find the notes from the August 15 telecon that discussed 6 
G considerations / SAS 2.0 market requirements. Attached is the 
presentation that we reviewed during the call at that time.

Marty Czekalski (Seagate) is revising the presentation, to review with 
the membership on Tuesday, Sept. 12 in Nashua immediately following the 
presentation by analyst Steve Denegri (who is scheduled for 6:30 PM 
that evening).

If you have further feedback after seeing these minutes, please make 
sure to get it to Marty.


STA 6 G considerations
August 15, 2006
10 AM pacific
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 656 bytes
Desc: not available
Url : http://lists.scsita.org/pipermail/scsita-members/attachments/20060824/fc6c02d2/attachment.bin 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: SAS 2.0 Requirements#9ED6EE.ppt
Type: application/vnd.ms-powerpoint
Size: 293888 bytes
Desc: not available
Url : http://lists.scsita.org/pipermail/scsita-members/attachments/20060824/fc6c02d2/attachment.ppt 
-------------- next part --------------

Bill Boyne Amphenol
Viad Pamtini, Agere
Yamini Shastry, Finisar
Richard Tomaszewski, HP
Jim Wong, Hitachi GST
Dan Reno, Hitachi GST
Steve Fairchild, HP
Rob Elliott, HP
Tonya Comer, HP
Mark Adams, LeCroy
Harry Mason, LSI Logic
Paul Wassenberg, Marvell
Rachelle Trent, PMC-Sierra
Marty Czekalski, Seagate
Martin Or, Sierra Logic
Alice Tate, STA

Presentation was circulated to the members reflector with the call 

We started by reviewing the amended Slide 3:

How much do we want to burden the standard in terms of the drives, for 
driving the longer cables? The information we have from the cable 
manufacturers is that 95% of cables are a certain length, and there are 
solutions that can get you to the longer distances.

We don?t want to burn up extra power in 90% of the drives that don?t 
need the longer distances.

Marty heard from the PHY group: If you make the half meter work in the 
backplane, 10m is not going to be a problem.

There is a current solution with the length.

Cables that don?t need the loss budget with 3G, should be labeled 
accordingly with a disclaimer.

Comments on Slide 4:

Rob noted that there is no additional cost burden as commented on the 
slide. It was noted that perhaps it should be added to say ?greater 
than 10m?

We need a rolling two generation window for multiplexing.

Comments on Slide 5:

The ?desired? statement was discussed. 4-to-1 ratio is not needed, it 
was described as a ?non-goal.?

Comments on Slide 6:

Depending on the size you want the channel to be, you can dial in your 
error rate.

Statistical simulation of the channel.

Hitachi GST noted that their design folks said that SSC wouldn?t 
necessarily be a requirement for 6G. They asked if anyone had data that 
showed the necessity of SSC? HP said from their perspective it is 
necessary. Protocol is being designed in T10 so that system vendors can 
choose to use it or not. Benefit vs. cost, complexity & compatibility. 
It was noted that SATA has had a lot of problems with this. Do we want 
to spend the cost on this for limited benefit?

How successful are we going to be using 3G with both transceivers and 
SSC? The real hold back would be VMEi (???) in existing systems. HP 
noted that we have voltage levels that would help this in 6G.

All devices shall be capable transceivers and SSC but the system 
vendors can choose to ?turn it on? or not.

This is still a question as to whether it should be a ?should? ?shall? 
or ?nice to have?

Zoning: questioned if this should be emphasized more. Commented that 
SAS 1 also has this in some devices, so it is not limited to only SAS 

Slide 7

No comments, discussed already on slide 1

Slide 8

Supporting 3 G SATA over Gen2i is already possible in certain scenarios.

Drive band has to be simpler due to SATA.

If it works fine, but lets not burden the cost of the receiver. Another 
?non-goal.? If it happens on its own, fine.

HP would like to see backplanes partitioned into two camps.

Slide 9

3G multiplexing?? do we need it or not. Long discussion ensued. 
Probably will be early 2009 before there are systems with multiplexing. 
Is the complexity worth it?

There need to be solutions that will accommodate both 6G and 3G.


Some concerns about the comment ?Bandwidth aggregation is optional?

It is in the spec as a feature.

Slide 10

First generation products are not required to be SAS 2 compatible.

Hitachi GST noted that they want there to be general guidelines that 
stay simple ?you have to have it all? rather than bits & pieces. Just 
four things: fixed gig, use of MiniSAS (n/a for drives), support SAS 
2.0, SSC, Multiplexing. If you implement it, it has to be done this 
way, but it is optional.

Comment re: full-duplex, active disk drive ? we should place 
requirements on spec with variances, because we will want to hit 
different price points and those are often covered in OEM vendor specs.

Next steps:

We are going to firm up the presentation and then re-present it to the 
group. We will do it in the evening before the Steve Denegri 

-=-=-=-=-=-=- SCSI Trade Association -=-=-=-=-=-=-
Alice Tate                                  Tel. +1.781.407.0288
Marketing Manager                 Fax +1.781.407.0199
10 Marlboro Street              E-mail: Alice at scsita.org
Dedham, MA 02026                   http://www.scsita.org

The SCSI Trade Association is managed by LoBue & Majdalany Management 
Group, an IAAMC Charter Accredited Firm.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 4571 bytes
Desc: not available
Url : http://lists.scsita.org/pipermail/scsita-members/attachments/20060824/fc6c02d2/attachment-0001.bin 

More information about the SCSITA-members mailing list