RE: Command Queue Depth (was asymmetric/Symmetric)

[Peter Johansson <PJohansson@ACM.org>]

At 08:38 PM 9/10/00, Jim McGrath wrote:

>So yes, I think the SBP-2 style of target initiated actions might make 
>some sense in heavily loaded systems, which a more traditional approach is 
>appropriate for lightly loaded situations.  The trick is having the 
>distributed targets and initiators figuring out when to use each one 
>(difficult) or focusing on one (as we have traditionally) and basically 
>optimizing latency for light or heavy loads, but not both.

Jim, I agree with what you've said (except to correct a small nit, 
responded to out of order below so as not to distract from the main topic) 
and think it's especially interesting in the light of the SBP-2 study group 
convening at T10 this week in Huntington Beach.

When the working group first developed SBP-2 we were pleased with our 
efforts in one particular aspect: the protocol became more efficient the 
more heavily it was loaded with commands.

Now that we have two years experience, there is a better understanding that 
it is worthwhile to have a special optimization to reduce startup latency 
when a target is idle, with no commands to process. In the case of SBP-2 
this seems to be simple to do; I look forward to your comments if your able 
to be there on Friday.

>In the extreme the target would only has to accept out of the blue an 
>alert message (just a bit of information) from each initiator, telling it 
>that the initiator has something for it ...

This is exactly what SBP-2 does with the DOORBELL register (that signals 
the target that new work awaits in the initiator).








Regards,

Peter Johansson

Congruent Software, Inc.
98 Colorado Avenue
Berkeley, CA  94707

(510) 527-3926
(510) 527-3856 FAX

PJohansson@ACM.org