CURRENT_MEETING_REPORT_
Reported by Bob Stewart/Xyplex
CHARMIB Minutes
The Character MIB Working Group held its third meeting at the IETF
meeting in Boulder. The meeting was well attended, with representatives
of many terminal server companies and other interested parties.
Overall, the meeting showed good consensus and resulted in the
completion of the business at hand. As a result, following edits based
on the meeting and summarized below, we have three MIB drafts that are
ready for initial implementation.
Agenda
o Discuss the Following Three Internet Drafts.
- draft-ietf-charmib-charmib-00.txt
- draft-ietf-charmib-rs232like-00.txt
- draft-ietf-charmib-parallelprinter-00.txt
o Discuss Implementation Plans.
We discussed the issue of permanent versus operational database. We
decided to leave such a distinction as implementation specific and to
say so in the specification. This general issue was to be mentioned in
the SNMP meeting.
We discussed the approach to action variables (such as charPortReset).
The specific questions were whether the apparent autoreset of the value
is acceptable, and why not use a write-only value. The editor explained
that the approach in the specification was taken after recommendations
from some of the SNMP gurus on the mailing list.
The Chair is to obtain official experimental MIB object identifier
numbers for the three MIBs.
The group would like a better understanding of the life cycle of an
experimental MIB. So would the Chair.
We discussed the life cycle of a charPortIndex as compared to an
ifIndex. They're basically the same, with the same restrictions, which
1
have been discussed quite thoroughly in the greater SNMP context.
We decided to count the BREAK condition as a character on the basis that
it is a significant event on the data lines. Furthermore, the character
counters are to include all characters, such as XON and XOFF.
We decided to add charLastChange, to operate similarly to ifLastChange.
We decided to add charPortSessionMaximum to indicate the maximum number
of sessions allowed on the port. A value of -1 indicates no maximum.
Setting the maximum to less than the current number of sessions has
unspecified results.
In page-by-page review, we agreed on the following specific edits to the
Character MIB:
o Page 7, remove the double negative in the second full sentence. No
change of meaning.
o Page 11, charPortIndex description ``network management system''
becomes ``network management agent''. The same change applies to
the other two MIBs.
o Page 12 and page 17, charPortReset and charSessKill, add ``in
response to a get-request or get-next-request'' to statement on
returning ``ready''.
o Page 12, charPortAdminStatus and charPortOperStatus, change
``test'' to ``maintenance'' and use test as an example.
o Page 17, remove charSessName.
o Page 18, remove ``active'' from charSessState.
o Page 18, change type of charSessProtocol to OBJECT IDENTIFIER and
define values for existing list plus local.
o Page 19, remove charSessRemoteName.
o Page 19, charSessConnectionId, indicate that the object identifier
should point to the highest available related MIB, such as Telnet
if available.
Page-by-page reviews of the RS-232 and Parallel MIBs resulted in no
edits.
2
We discussed the parity, framing, and overrun counters in the RS-232
MIB, and decided to keep them as they are.
In a quick, informal poll of who had implementation intentions, positive
responses were obtained from Hughes LAN Systems, Digital Equipment
Corporation, Datability, Xylogics, and Xyplex.
3
Attendees
Anne Ambler anne@spider.co.uk
Anthony Chung anthony@hls.com
Jim Reinstedler jimr@ub.com
Bob Stewart rlstewart@eng.xyplex.com
Ron Strich ssds!rons@uunet.uu.net
Emil Sturniolo
Laszlo Szerenyi
William Townsend townsend@xylogics.com
David Waitzman djw@bbn.com
Jesse Walker walker@eider.enet@decpa.dec.com
4