SDP Mail Archive: Re: SDP> Suggestions for discussion at SDP session next week

Re: SDP> Suggestions for discussion at SDP session next week

Carl Kugler (kugler@us.ibm.com)
Fri, 15 May 1998 13:09:25 -0400

>Robert Herriot (robert.herriot@Eng.Sun.COM)
>Thu, 14 May 1998 12:37:39 -0700
>
>In my solution, I assume that there is a port dedicated to returning p=
rinter
>status only and it has higher priority for processing than the normal =
IPP
>port or the SNMP port. I want to keep the probability of a reply near =
100%
>even when the printer is very busy with other network requests.
>
What assumptions are you making about the platform to allow this kind o=
f tuning
(some ports processed at higher priority)?

>The concern I have about SNMP is that a client's "get status" query co=
uld be
>behind
>several GetBulk requests and thus not get processed fast enough for th=
e client
>to
>determine that the printer is alive but very busy.
>
This seems like a worst-case scenario. Who would be doing these GetBul=
ks and
why? How often? What are the consequences when the query is not proce=
ssed
fast enough?

>Can an SNMP request have as good a guarantee of a quick response as a
>dedicated port?
>
How good a guarantee is required for this application? No response is
guaranteed over UDP anyway. Is tuning of the snmp server allowed?

>Also in my proposal, I assume that the data returned is in an IPP comp=
atible
>format.
>
>Bob Herriot
>

Carl Kugler
=