Jay, your opinion is always valuable to me. If they weren't, I wouldn't bug
you.
>
> - We currently have no interest in those items; we are primarily focused
> on STATUS and far less on CONFIGURATION in our product initiatives at
> the moment. As you have witnessed, dealing with issues on STATUS alone
> has already consumed our available time.
>
> However, now that you have pressed us for a response, let me say this...
>
> After dealing with RFC1759 for close to three years now, we are unsure as
> to whether your rather complex configuration environment (ie, multiple
> printers sharing common devices) can be reasonably handled via RFC1759.
> It just seems too complex to us. (Your--and others'--mileage may vary.)
Unfortunately, two devices are a very real to us at EFI. As you may remember,
we attach copiers to the network via a box call an "XJ". This box has a single
PS rip, supports multiple networks, and feeds two copiers. Some time ago, I
asked how EFI should represent this system via 1759, given that some of the
components were shared, and some were unique. The response that the committee
came back with, was the same that we internally came up with: represent it as
two printers.
Now we had been planning on having one SNMP agent report that there were two
printer devices, and then index the two devices with hrDeviceIndexes of .1 and
.2, hence Eugene's question.
>
> It could very well be that when the dust settles, the PWG may have to
> make some sweeping statements regarding the "reasonable" application of
> RFC1759 (and its successor) to a printer environment.
>
> For example, note the many, many concerns regarding the use of the
> Host Resources MIB. Many folks absolutely DESPISE this mechanism,
> however, how else can one "manage" multiple printers via a single
> SNMP agent implementation?
>
> One solution to this problem is to say, simply, that you CAN'T; that is,
> deprecate the entire HR MIB mechanism and allow only a single instance
> of a printer within an agent. In doing so, Bill Wagner's longstanding
> problems surrounding the Interfaces group could be immediately solved.
> (That is, all objects describing "interfaces" would be implicitly owned
> by the printer, and not the host on which the agent is running, etc.)
I understand why you want to get rid of it, but this poses some complications
for us. The XJ has one IP address, and one, well known, snmp socket. Without
hrDeviceIndex on the printer mib variables, EFI could only report one of the
two attached copiers. This would be a poor solution for us and our OEM's.
>
> I know this is not exactly what you were looking for, but there it is.
> Perhaps others can be of more help to EFI in this area.
>
> ...jay
>
> ----------------------------------------------------------------------
> -- JK Martin | Email: jkm@underscore.com --
> -- Underscore, Inc. | Voice: (603) 889-7000 --
> -- 41C Sagamore Park Road | Fax: (603) 889-2699 --
> -- Hudson, NH 03015-4915 | Web: http://www.underscore.com --
> ----------------------------------------------------------------------
>
--Gail Songer Electronics For Imaging gail.songer@eng.efi.com 2855 Campus Drive (415) 286-7235 San Mateo, CA 94403