SAS 2.1 connector considerations

Paul Von Stamwitz pvstamwitz at amcc.com
Wed Dec 3 13:50:42 PST 2008


* From the T10 Reflector (t10 at t10.org), posted by:
* "Paul Von Stamwitz" <pvstamwitz at amcc.com>
*
Alvin, Here are my 2 cents.
First, my assumptions:
1. We should provide an active copper solution for 2.1 to support 20-25m
cables.
2. I don't see a need for optical for 6Gb SAS since the vast majority
(if not all) of the applications will be handled by the passive/active
copper cables.
3. I do see the need for optical at 12Gb, since active copper may not be
able to achieve the 25m distance.
4. Lane density (# of lanes per sq. inch) is more important for internal
than external.
With these in mind, I am wary of defining a new connector for optical in
2.1. We have a limited timeframe which will probably require us to use
existing solutions such as the current 8088 or QSFP which have their
drawbacks (lack of sidebands, size of footprint, etc.) If we do choose a
new connector for 2.1, I believe we will be shackled with it for 3.0,
because of backward compatibility and to avoid confusing the customer
base by changing cables too often.
Therefore, I recommend the following:
1. We table optical considerations until 3.0 on the assumption that
there will be a new connector for copper/optical.
2. For 2.1 we adopt 08-358 to add active power to the existing Mini SAS
4x and live with the keying confusion until 3.0.
3. I don't think we need an external connector with greater than 4
lanes, but if we do, we should use the 68-circuit 8088 (Mini SAS 8x)with
appropriate sidebands instead of keys. The primary usage model should be
8x-to-8x (i.e. let's try to avoid the Y-cable to two Mini SAS 4x.) I
also don't think we need the modular approach for external (e.g.
SFF-8644 as proposed in 08-455). Many applications would use a hybrid
modular plug to Mini SAS 4x, so the sidebands on the modular side would
have to match the keying on the Mini SAS side. And, I don't think we
have the lane density issue for external that we have for internal. 
4. So, for internal, we should consider the modular approach (e.g.
SFF-8643 in 08-455). The currently defined sidebands should be used (but
the pinouts can be made less confusing) and they should be duplicated on
every 4x connection.
In summary, I think optical should wait for 3.0. We should also avoid
major changes to the external connectors for 2.1 and make sure the
solutions we make for 3.0 address all the issues. Any changes to the
external connections should be "variations on the them", such as active
mini SAS 4x, mini SAS 8x, etc. For internal, a narrower version of the
mini SAS 4i by stacking 2 lanes over 2, while maintaining the same
sidebands per 4 lanes, makes good sense.
-Paul 
________________________________
From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of
Alvin.Cox at seagate.com
Sent: Monday, November 24, 2008 2:25 PM
To: t10 at t10.org
Subject: SAS 2.1 connector considerations
Feedback from users would help to narrow the selection field and number
of proposals to better fit the application model. In addition to the
posting by Alvin Cox regarding the input from the November t10 meeting,
a few opinions would be extremely helpful. Please consider both internal
and external applications. 
08-453 <http://www.t10.org/cgi-bin/ac.pl?t=d&f=08-453r0.pdf&gt;;  
Is a modular design favored over a set number of connections? 
	This would likely consist of 4x blocks and could allow for
multiple cables or different single cables that have the 4x, 8x, 12x,
16x configurations. 
	Trade-off of modular may be encountered due to area. 
	Suggestions for sideband support? 
		Separate module? 
		Include in one module? 
		Duplicate on all modules? 
			Partial or all lines? 
Sidebands? 
	How many required for internal and external? 
	Does external provide more information than just cable power and
type? 
Power requirements? 
	How much power is required for a 16x configuration? 
		Copper 
		Optical 
	What voltage and current? 
		Cable perspective 
		User perspective 
Alvin Cox
Seagate Technology, LLC
Office 405-392-3738
Cell 405-206-4809
E-Mail	alvin.cox at seagate.com
--------------------------------------------------------
CONFIDENTIALITY NOTICE: This e-mail message, including any attachments, is
for the sole use of the intended recipient(s) and contains information that
is confidential and proprietary to Applied Micro Circuits Corporation or its
subsidiaries. It is to be used solely for the purpose of furthering the
parties' business relationship. All unauthorized review, use, disclosure or
distribution is prohibited. If you are not the intended recipient, please
contact the sender by reply e-mail and destroy all copies of the original
message.
*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org



More information about the T10 mailing list