SCSI, Networks & Sense Data

Charles Monia monia at am.shrmsg.shr.MTS.dec.com
Tue Apr 4 16:24:48 PDT 1995


Ken Hallam wrote in part:

>
>Your bland assurance that the ULP only needs to know if the link works or
>is broken, does not cover all of the cases that I can think of. Most larger
>systems want to know something about the progress of time dependent
>operations, (like those on tape drives for instance) and about types and
>classes of errors in the LLP. This is not to say such applications create a
>mandate for their inclusion into SAM, only that the possibility exists and
>should be allowed for.
>

Hi Ken:

Rather than reply at length, I'd like to blandly point out that SAM reflects
the way layered, distributed systems are commonly designed and specified. Such
a reply would, therefore, only repeat ideas that can be found in any text on
the subject.

Regarding the problems you say are in SAM, I think you should prepare a
detailed proposal for consideration by the committee. The proposal should
describe the problems in detail and the changes you think are needed..

Thanks,
Charles Monia





More information about the T10 mailing list