PMP Mail Archive: RE: PMP> [delete ppmPrinterEnabled] Restru

PMP Mail Archive: RE: PMP> [delete ppmPrinterEnabled] Restru

RE: PMP> [delete ppmPrinterEnabled] Restructured Port MIB (18 Jul y 2005)

From: McDonald, Ira (imcdonald@sharplabs.com)
Date: Thu Jul 28 2005 - 11:49:38 EDT

  • Next message: Stuart Rowley: "RE: PMP> [delete ppmPrinterEnabled] Restructured Port MIB (18 Jul y 2005)"

    Hi Bert,

    It does help, yes. We have been prototyping against the equivalent
    of Internet-Drafts (PWG Working Drafts). After a very successful
    initial interoperability test two weeks ago, we realized that
    restructuring (from two object groups to three object groups)
    was necessary. I posted an ASN.1-only sketch of the restructuring.

    Of necessity, the OIDs changed (because the columnar objects moved
    from one table to the other).

    Jerry Thrasher's team at Lexmark was so quick that they've already
    prototyped the new sketch.

    Probably we should close the hole. That's what I just asked for
    feedback about on this mailing list. Both Client (SNMP Command
    Generator) and Agent (SNMP Command Receiver) applications have been
    prototyped by OS and printer vendors.

    Cheers,
    - Ira

    Ira McDonald (Musician / Software Architect)
    Blue Roof Music / High North Inc
    PO Box 221 Grand Marais, MI 49839
    phone: +1-906-494-2434
    email: imcdonald@sharplabs.com
    -----Original Message-----
    From: Wijnen, Bert (Bert) [mailto:bwijnen@lucent.com]
    Sent: Thursday, July 28, 2005 11:28 AM
    To: McDonald, Ira; 'thrasher@lexmark.com'
    Cc: 'pmp@pwg.org'
    Subject: RE: PMP> [delete ppmPrinterEnabled] Restructured Port MIB (18 Jul y
    2005)

    I have not followed detailed discussion on this... but

    - SMI rules state that one a document with a MIB is published, you can NEVER
      re-use an OID for some other purpose. All you can do is obsolete the
      old object and add a new one.
    - In IETF, when we just have internet drafts, we allow people to re-use an
    OID
      (supposedly no-one has used the OID while at internet draft state,
    besides.
      such MIB modules normally have a

         ::= { mib-2 xxxx } -- xxxx to be assigned by IANA

      so there is no definite complete OID, and people who want to do early
    implementations
      fill in a number under their enterprise prerelease branch or so.

      the final xxxx gets assigned by IANA upon document approval (right before
    RFC
      publication) and so no OID changes are allowed anymore.
      See the SMI documents for the rationale.

    - But once published as RFC, we never re-use an OID, not even when a new rev
    is
      being published as a intrenet-draft.

    Hope this helps,

    Bert
    -----Original Message-----
    From: pmp-owner@pwg.org [mailto:pmp-owner@pwg.org]On Behalf Of McDonald, Ira
    Sent: Thursday, July 28, 2005 16:59
    To: 'thrasher@lexmark.com'; McDonald, Ira
    Cc: 'pmp@pwg.org'
    Subject: RE: PMP> [delete ppmPrinterEnabled] Restructured Port MIB (18 Jul y
    2005)

    Hi Jerry,

    There's no such thing as a reserved no-access object, but I'd be happy to
    leave
    the hole (so that the OIDs don't change for any other columnar objects).
    Would
    you prefer that?

    Note that ALL of the columnar objects were cleaned up and reordered (and
    renumbered) in this latest revision, so it would be hard for there to be too
    many
    implementations of the new OIDs already.

    Others - shall I leave the hole and leave the other new OID assignments
    stable?

    Cheers,
    - Ira

    Ira McDonald (Musician / Software Architect)
    Blue Roof Music / High North Inc
    PO Box 221 Grand Marais, MI 49839
    phone: +1-906-494-2434
    email: imcdonald@sharplabs.com
    -----Original Message-----
    From: thrasher@lexmark.com [mailto:thrasher@lexmark.com]
    Sent: Thursday, July 28, 2005 10:33 AM
    To: McDonald, Ira
    Subject: RE: PMP> [delete ppmPrinterEnabled] Restructured Port MIB (18 Jul y
    2005)

    Ira,

    Are you just planning on creating a reserved no-access object where
    ppmPortEnabled "used" to be, or are
    you going to delete the entry and shift everything else in the table up.??

    Jerry

    "McDonald, Ira" <imcdonald@sharplabs.com>
    07/23/2005 01:56 PM
            
            To: "McDonald, Ira" <imcdonald@sharplabs.com>,
    "'thrasher@lexmark.com'" <thrasher@lexmark.com>
            cc: "'pmp@pwg.org'" <pmp@pwg.org>,
    "'Ron.Bergman@rpsa.ricoh.com'" <Ron.Bergman@rpsa.ricoh.com>
            Subject: RE: PMP> [delete ppmPrinterEnabled] Restructured
    Port MIB (18 Jul y 2005)

    Hi,

    Let's delete ppmPrinterEnabled, because the 'false' is redundant
    with ppmPortEnabled of 'false' for all supported ports. And
    because otherwise, we must say "ignore ppmPortEnabled of 'true'
    for installation if ppmPrinterEnabled is 'false'". Which leads
    me to agree that we should delete ppmPrinterEnabled.

    When a Printer is permanently removed from an ENA interface,
    the whole ppmPrinterTable row and all subordinate rows in
    ppmPortTable can just be deleted.

    For 'ppmPortProtocolType' using 'PrtChannelTypeTC', we're working
    to get 'unknown(2)' registered quickly with IANA, so that we can
    do the 'right thing' that Bert Wijnen pointed out.

    Cheers,
    - Ira

    Ira McDonald (Musician / Software Architect)
    Blue Roof Music / High North Inc
    PO Box 221 Grand Marais, MI 49839
    phone: +1-906-494-2434
    email: imcdonald@sharplabs.com
    -----Original Message-----
    From: McDonald, Ira
    Sent: Friday, July 22, 2005 1:37 PM
    To: 'thrasher@lexmark.com'; McDonald, Ira
    Cc: pmp@pwg.org; Ron.Bergman@rpsa.ricoh.com
    Subject: RE: PMP> Restructured Port MIB (18 July 2005)

    Hi Jerry,

    I agree that feedback from the clients (Microsoft and Apple) on how they'd
    like this to work would be helpful.

    Remember, that a value of 'ppmPrinterIndex' must NEVER be reassigned
    to a different instance of a Printer at a later date. While the MIB may
    grow and shrink, the base 'ppmPrinterIndex' should be immutably
    associated with exactly one specific instance of a Printer. This is both
    correct MIB practice and required by the object definition in the current
    MIB draft.

    Cheers,
    - Ira

    Ira McDonald (Musician / Software Architect)
    Blue Roof Music / High North Inc
    PO Box 221 Grand Marais, MI 49839
    phone: +1-906-494-2434
    email: imcdonald@sharplabs.com
    -----Original Message-----
    From: thrasher@lexmark.com [mailto:thrasher@lexmark.com]
    Sent: Friday, July 22, 2005 1:28 PM
    To: McDonald, Ira
    Cc: pmp@pwg.org; Ron.Bergman@rpsa.ricoh.com
    Subject: RE: PMP> Restructured Port MIB (18 July 2005)

    I would think that the usefullness of the ppmPrinterEnabled for an
    ENA would only be for the "transcient" case of the plugging/unplugging
    of the related printer. However at some point (in the case that the printer

    stays unplugged "permanently") the printer entry and associated protocol
    tables would be removed such that the the MIB could grow and shrink
    over time.....

    For example, an ENA with a USB host interface that supports up to
    128 attached printers, In my opinion, shouldn't need to have 128 printer
    entries
    in the MIB tables from first power on......only the entries that have
    detected (valid 1284ID)
    printers etc. attached.....otherwise you'd end up in most cases with 127
    default printer entries
    with associated port tables that don't actually go anywhere.

    Of course this behaviour is different from the current TCPMON.ini file
    which has static entries.......

    I think maybe the clients should give guidance on how they want it to work.

    Jerry Thrasher

    "McDonald, Ira" <imcdonald@sharplabs.com>
    Sent by: pmp-owner@pwg.org
    07/22/2005 12:13 PM

    To: "'Bergman, Ron'" <Ron.Bergman@rpsa.ricoh.com>, "McDonald,
    Ira" <imcdonald@sharplabs.com>, "Wijnen, Bert (Bert)" <bwijnen@lucent.com>,
    pmp@pwg.org
    cc:
    Subject: RE: PMP> Restructured Port MIB (18 July 2005)

    Hi Ron,

    If the Printer entry is deleted when an ENA interface
    is disconnected, then all the subordinate Port entries
    MUST be deleted too (because they are indexed by the
    object ppmPrinterIndex). This is ugly if the local
    printer is promptly plugged _back_ into the interface.

    If the Printer entry is left in place but _not_ clearly
    marked 'disabled', then ppmPrinterIEEE1284DeviceId,
    ppmPrinterHrDeviceIndex and all the other Printer
    columnar objects must be reset (to default values).

    That's why the ppmPrinterEnabled object should be kept.

    The WG concensus was strong that ppmPortEnabled was
    required to keep the port list static (fixed number
    of ports for an interface). Therefore, I added the
    ppmPrinterEnabled object.

    If others want ppmPrinterEnabled removed, would they
    please speak up soon?

    Cheers,
    - Ira

    PS - Remember that this MIB is supposed to work for
    Network Spoolers too, where the concept of 'the printer
    is removed' is fuzzy. The 'printer' is just some
    configured downstream network printer that may well
    be administratively disabled _without_ removing the
    configuration at the Network Spooler.

    Ira McDonald (Musician / Software Architect)
    Blue Roof Music / High North Inc
    PO Box 221 Grand Marais, MI 49839
    phone: +1-906-494-2434
    email: imcdonald@sharplabs.com

    > -----Original Message-----
    > From: Bergman, Ron [mailto:Ron.Bergman@rpsa.ricoh.com]
    > Sent: Thursday, July 21, 2005 8:24 PM
    > To: McDonald, Ira; Wijnen, Bert (Bert); pmp@pwg.org
    > Subject: RE: PMP> Restructured Port MIB (18 July 2005)
    >
    >
    > Ira,
    >
    > Base on my experience with ENAs, they do not provide a feature to
    > disable an output port unless the printer is removed. Normally,
    > this is to replace a worn-out unit or upgrade a printer.
    > In this case the old printer is gone forever. So how does your
    > "STATIC entries" handle this situation?
    >
    > Ron
    >
    > -----Original Message-----
    > From: McDonald, Ira [mailto:imcdonald@sharplabs.com]
    > Sent: Wednesday, July 20, 2005 8:38 AM
    > To: Bergman, Ron; McDonald, Ira; Wijnen, Bert (Bert); pmp@pwg.org
    > Subject: RE: PMP> Restructured Port MIB (18 July 2005)
    >
    >
    > Hi Ron,
    >
    > Based on previous IPP experience, it will take MONTHS to add one
    > new enum to the PrtChannelTypeTC with IANA - that would stop the
    > Port Mon MIB dead in its tracks until it was accepted by IANA.
    >
    > About ppmPrinterEnabled - same rationale as ppmPortEnabled - keeps
    > the number of Printer entries STATIC in an implementation - lets
    > the user see that the one Printer (i.e., hardward output interface)
    > on an External Network Adapter should presently be ignored.
    >
    > Remember that the Port Mon MIB MUST NOT depend on either Host
    > Resources or Printer MIB, by common concensus - it may only
    > AUGMENT them, if they are present.
    >
    > Cheers,
    > - Ira
    >
    > Ira McDonald (Musician / Software Architect)
    > Blue Roof Music / High North Inc
    > PO Box 221 Grand Marais, MI 49839
    > phone: +1-906-494-2434
    > email: imcdonald@sharplabs.com
    >
    > > -----Original Message-----
    > > From: Bergman, Ron [mailto:Ron.Bergman@rpsa.ricoh.com]
    > > Sent: Tuesday, July 19, 2005 7:40 PM
    > > To: McDonald, Ira; Wijnen, Bert (Bert); pmp@pwg.org
    > > Subject: RE: PMP> Restructured Port MIB (18 July 2005)
    > >
    > >
    > > Ira,
    > >
    > > I am not sure what value ppmPrinterEnabled adds to the MIB.
    > > This appears to be analogous to
    > > On Line/Off Line. If I want to create a driver for the
    > > printer I don't care what the current
    > > state is. That information is only necessary when I am ready
    > > to print and then this MIB is
    > > not used.
    > >
    > > I believe that Bert has a valid point in using
    > > ppmPortProtocolType. It is not a major effort
    > > to add unknown(2) to the IANA registrations.
    > >
    > > Otherwise, the changes are inline with our discussions
    > > following the test.
    > >
    > > Ron
    > >
    > > -----Original Message-----
    > > From: pmp-owner@pwg.org [mailto:pmp-owner@pwg.org]On Behalf
    > > Of McDonald,
    > > Ira
    > > Sent: Tuesday, July 19, 2005 9:46 AM
    > > To: 'Wijnen, Bert (Bert)'; McDonald, Ira; 'pmp@pwg.org'
    > > Subject: RE: PMP> Restructured Port MIB (18 July 2005)
    > >
    > >
    > > Hi Bert,
    > >
    > > Thanks for your quick feedback. My replies inline below.
    > >
    > > Cheers,
    > > - Ira
    > >
    > >
    > > Ira McDonald (Musician / Software Architect)
    > > Blue Roof Music / High North Inc
    > > PO Box 221 Grand Marais, MI 49839
    > > phone: +1-906-494-2434
    > > email: imcdonald@sharplabs.com
    > >
    > > > -----Original Message-----
    > > > From: Wijnen, Bert (Bert) [mailto:bwijnen@lucent.com]
    > > > Sent: Tuesday, July 19, 2005 9:08 AM
    > > > To: McDonald, Ira; 'pmp@pwg.org'
    > > > Subject: RE: PMP> Restructured Port MIB (18 July 2005)
    > > >
    > > >
    > > > Only did a very very quick scan.
    > > >
    > > > Comments.
    > > >
    > > > - ppmPortProtocolTargetPort OBJECT-TYPE
    > > > SYNTAX Integer32 (0..65535)
    > > > I propose that you use InetPortNumber TC from RFC4001
    > > >
    > >
    > > Won't work, because this port is not limited to Internet Suite
    > > protocols. The 'service:' URI in ppmPortServiceNameOrURI may
    > > also be for non-Internet suites (AppleTalk, NetWare, etc.).
    > >
    > > I'll correct the DESCRIPTION in the MIB and make clear that
    > > (as with the Printer MIB) ports/channels may be from multiple
    > > protocol suites.
    > >
    > >
    > > > - ppmPortProtocolType OBJECT-TYPE
    > > > SYNTAX Integer32 (0..2147483647)
    > > >
    > > > WHy not use TC PrtChannelTypeTC as the SYNTAX?
    > > > I do see that you want to use zero (meaning not supported).
    > > > But maybe better is to use none(1) in that case, or maybe
    > > > adding an enumeration to the TC of notSupported(xx) ??
    > > > It is now an IANA-maintained TC, so it should not be that
    > > > difficult to get a label added.
    > > >
    > >
    > > Won't work. PrtChannelTypeTC currently only defines 'other(1)'
    > > and (foolishly) does NOT define 'unknown(2)' (unlike every other
    > > textual convention in the Printer MIB). Because the Printer MIB
    > > v2 still doesn't define DEFVAL clauses for most objects, this
    > > oversight has not surfaced before. We could register 'unknown(2)'
    > > with IANA, but _not_ fast enough (because this MIB's going into OS
    > > and printer vendor products right now).
    > >
    > >
    > > > - ppmPortPrtChannelIndex has a reference to RFC1213, while I
    > > > think I would reather reference RFC2863 (the current IF-MIB)
    > > >
    > > > Bert
    > > >
    > >
    > > Agreed. My mistake from the old Printer MIB (RFC 1759).
    > >
    > > I'll correct the references in the MIB.
    > > - Ira
    > >
    >



    This archive was generated by hypermail 2b29 : Thu Jul 28 2005 - 11:50:18 EDT