P1394 Mail Archive: Re: P1394> units and logical units definition

P1394 Mail Archive: Re: P1394> units and logical units definition

Re: P1394> units and logical units definition

Greg Shue (gregs@sdd.hp.com)
Mon, 16 Mar 1998 09:57:26 -0800 (PST)

Ats asks:
> In the SBP-2 spec, it says ;
>
> "A logical unit is part of the unit architecture and is
> an instance of a device model e.g., mass storage,
> CD-ROM or printer"
>
> In section 17:Model Configuration of document 0.1d -PWG
> printer profile, it says ;
>
> "Each unique combination of the above information represents
> a unique service provided by the device and is identified as a
> separate Unit within the device. Zero or more instances of
> a unit may exist within the device at any given time.
> All instances of a unit have identical functionality."
>
> It may be my lack of mis-understanding, but doesn't the
> 2 statements conflict ?

The two statements do not conflict. The PWG profile is more
restrictive than the SBP-2 proposal.

Note that SBP-2's concept of a printer certainly does not match
well with our devices. Disks are easily modeled. Scanners are
fairly easily modeled. Printers definitely are not. (They would
have fits trying to deal with the degree of flexability in the
Printer MIB! :-)

> SBP-2 considers logical units as instances of "device" models
> whereas the printer profile considers logical units as instances
> of a "logical channels".

Almost. The PWG profile considers a logical unit as an instance
of a service. The login establishes an instance of a
communications channel to the associated service. If you think
of the service definition as being it's own device model, then
everything should line up.

> If there is 1 instance of a printer device with multiple
> SBP-2 logical channels, how do I address this device as a
> unit architecture following both rules ?
>
> 1 "printer" unit with 2 logical unit directories?...
> but then,a logical unit has to be "an instance" of a
> device model,so there should be only 1 logical unit ????

The confusion comes from trying to model a printer as "1" device.
There are really many independent services, which is why
everybody wants Multiple Independent Channels.

In my mind, a "printer" is composed of (at least):

1+ Print Job interpretors (PRINTING service) running concurrently
1+ Printer Status monitors (DEVICE STATUS service) running concurrently
1+ Printer Control services (DEVICE CONTROL service) running concurrently

There are a number of ways in which this could be arranged according
to SBP-2, and fewer according to the PWG proposal. Let's take the
case of a "printer" which can simultaneously manage accepting and
parsing 2 print files, servicing N different remote queries, and
allows 1 client control access at a time:

The SBP-2 proposal allows this to be modeled as:
1 unit with:
2 logins to the Print Job logical unit
N logins to the Printer Status logical unit
1 login to the Printer control logical unit

Similarly, since each of these services operate independently
of each other (and may be used by other protocol stacks), SBP-2
allows for them to be modeled as:
2 + N + 3 units with 1 logical unit each
or:
1 Print Job Unit Directory with 2 logical units (instances)
1 Device Status Unit Directory with N logical units (instances)
1 Device Control Unit Directory with 1 exclusive logical unit

The current PWG proposal restricts the options to the last one.
For the "Disk & Media Changer" example, we would have to ask the
question "Do the 2 functions operate independently of each other?"
I don't know. If they do, then they would be modeled as 2 units
with one logical unit per supported connection to each.

Does this help?

Just to check my assumptions:
Am I incorrectly modeling a "printer"?

> Ats

-- 
Greg Shue
Hewlett-Packard Company
Office Products Division			gregs@sdd.hp.com
----------------------------------------------------------------