[STA Members] [STA Plugfest] Follow up SAS Plugfest discussion

Elliott, Robert (Server Storage) Elliott at hp.com
Tue Aug 4 09:28:13 PDT 2009


I don't understand that table.  A few points:

1. These are the possible physical link rates that might be in use by a phy:
  SNW-1 based 1.5 Gbps (i.e. legacy 1.5 Gbps)
  SNW-2 based 3 Gbps (i.e. legacy 3 Gbps)
  SNW-3 based 1.5 Gbps no SSC tx/rx
  SNW-3 based 3 Gbps no SSC tx/rx
  SNW-3 based 6 Gbps no SSC tx/rx
  SNW-3 based 1.5 Gbps with SSC tx/rx
  SNW-3 based 3 Gbps with SSC tx/rx
  SNW-3 based 6 Gbps with SSC tx/rx
  SATA 1.5 Gbps no SSC tx
  SATA 3 Gbps no SSC tx
  SATA 6 Gbps no SSC tx
  SATA 1.5 Gbps with SSC tx
  SATA 3 Gbps with SSC tx
  SATA 6 Gbps with SSC tx
  (SATA rates always require SSC rx capability)

Each one is different.  In particular, SNW-1 based 1.5 Gbps is different than SNW-3 based 1.5 Gbps, and SNW-2 based 3 Gbps is different than SNW-3 based 3 Gbps.  Differences include:
    SNW-1/2: max amplitude 1600 mV, DFE probably not used
    SNW-3: max amplitude 1200 mV, DFE may be used

It is difficult to reach SNW-3 based 1.5 and 3 Gbps since they are normally only attempted if 6 Gbps training fails (since a phy supporting SNW-3 probably supports 6 Gbps).

2. Regarding SSC over SAS, end devices always transmit down-spreading SSC, and expander devices always transmit center-spreading SSC.  Both end devices and expander devices must be able to receive both down-spreading and center-spreading.  Expander devices attached to SATA must use down-spreading, though.

3. Multiplexing is never enabled over a 1.5 Gbps physical link.




> -----Original Message-----
> From: scsita-plugfests-bounces at scsita.org [mailto:scsita-plugfests-
> bounces at scsita.org] On Behalf Of David Woolf
> Sent: Tuesday, August 04, 2009 10:55 AM
> To: Michael Fitzpatrick
> Cc: plugfests at scsita.org; members at scsita.org
> Subject: Re: [STA Plugfest] Follow up SAS Plugfest discussion
> 
> To help with this discussion, UNH-IOL has put together a matrix of the
> possible Phy capabilities that a device could potentially support, see
> the attached spreadsheet.
> 
> What the matrix shows is that at a single speed, there are 8 possible
> phy
> configurations, not including different settings for amplitude,
> pre-empahsis, and equalization.
> 
> I'd like to hear what the best way to delineate (i.e. high medium low
> or
> specific values) these other three settings are ( amplitude, pre-
> emphasis,
> equalization ) This way we can determine all the possible phy
> configuration permutations. As you can see this will lead to a big
> number
> of test runs, much more than could be accomplished during 1 week. The
> next
> step will be to whittle down to the most interesting 4-5 combinations.
> 
> thanks,
> David
> 
> 
> On Mon, 3 Aug
> 2009, Michael Fitzpatrick wrote:
> 
> >
> > Everyone,
> >
> > I would like to follow up on our discussion we had last  past Tuesday
> on
> > the plugfest call.
> >
> > I for one would be interested in doing additional phy level testing
> at the
> > up coming plugfest.  I definitely like the idea of having multiple
> devices
> > with different settings already programmed into them or like someone
> stated
> > to have an expander/switch with multiple ports and each port having
> > different settings.
> >
> > In order to flush out this idea on how we do this additional phy
> level
> > testing, we need to have some understanding of just how many
> companies
> > would be interested in doing more phy level testing.
> >
> > If you could let David Woolf and myself know if your company is
> interested
> > in doing additional phy level testing, and there is enough interest,
> we can
> > start putting some ideas together for discussion on the next couple
> of
> > call.
> >
> > Any questions, please let me know.
> >
> > Thanks,
> >
> > Mike Fitzpatrick
> >
> > Email    mfitzpatrick at us.fujitsu.com
> >
> > Cell phone    (405) 820-5809
> >
> >
> > This message contains information which may be confidential and
> privileged.=
> > Unless you are the addressee (or authorized to receive for the
> addressee),=
> > you may not use, copy or disclose to anyone the message or any
> information=
> > contained in the message. If you have received the message in error,
> pleas=
> > e advise the sender by reply [email], and delete the message.
> >
> 
> --
> 
> David Woolf
> ************************************************
> UNH-IOL
> www.iol.unh.edu
> Senior Technical Staff
> 121 Technology Dr. Suite 2
> Durham, NH 03824
> +1-603-862-3749
> ************************************************



More information about the SCSITA-members mailing list