The IP Over NBMA (ION) Archive

Cell Relay Retreat>ION Archive>month:1996-Jun> msg00144



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

comments on draft-ietf-iplpdn-frmib-dte-07.txt

  • From: alan@irvine.xylan.com (Alan Bartky)
  • Date: Mon, 17 Jun 96 14:51:33 PDT

Comments edited in below:

- Alan K. Bartky
  Xylan Corporation
  6/17/96 14:45

> From owner-ion@nexen.com Mon Jun 17 12:49:48 1996
> X-Sender: fred@stilton.cisco.com
> Mime-Version: 1.0
> Content-Type> : > text/plain> ; > charset="us-ascii"> 
> Date: Mon, 17 Jun 1996 12:22:53 -0700
> To: ion@nexen.com
> From: roy.spitzer@adn.alcatel.com (Roy Spitzer) (by way of fred@cisco.com (Fred
>  Baker))
> Subject: comments on draft-ietf-iplpdn-frmib-dte-07.txt
> Sender: owner-ion@nexen.com
> X-Info: [Un]Subscribe to ion-request@nexen.com, submissions to ion@nexen.com
> X-Info: Hypermail archive at http://netlab.ucs.indiana.edu/hypermail/ion/
> X-Info: FTP archive at ftp://ftp.nexen.com/pub/ion/
> Content-Length: 7114
> 
> Carolyn, Fred, and Charles,
> 
> I am sorry that these comments are so late.  I just found out about this
> new iteration of RFC1315 at the last Frame Relay Forum meeting from Andy
> Malis.  At the same time our offices was moving and I just was able to retrieve
> a copy of the proposed MIB for the Montreal IETF June 24, 1996.
> 
> How do I get on the proper distribution list for these discussions.  Andy
> gave me ion and it did not seem right.  Please forward this mail to the
> proper distribution list.
>

This MIB is for all intents and purposes done for now.  We at Xylan
have a working version of the draft MIB and I see no need for any
of the changes you have requested.  There will be a lot of changes
needed for SVCs and possibly for RFC1490 multiprotocol and other
issues in the upcoming MIBs for Frame Relay DTE (USER) and network
devices.


As far as learning about this "so late" the MIB working group
that worked on this is the same one that drafted RFC1315 
(iplpdn) and if anyone is doing any MIB work it is very
easy to find these working groups.  Drafts are kept
in the public FTP servers and anyone doing work on any
RFC MIB they should go to the FTP servers and do searches
for drafts on the RFCs and/or subjects they are working on.
The Frame Relay Forum is not where this work is going
to draft the MIBs, the IETF is the appropriate body
for this work.

> Comments:
> 
> 1.      We believe that a definition of frTrapState at the interface
>         basis is more useful for some vendors than for all interfaces.
>         Some interfaces may be managed by different agents than others.
>         I would suggest that frDlcmiTrapState be added to the frDlcmiTable.
>

 
> 2.      The frDlcmiTable does not support bidirectional signalling.  Will
>         this be a problem for the future?  Currently, I know of no DTE
>         vendor that supports bidirectional signalling.
>

I don't know of any either.  That is the point of working in
the MIB for consensus.  If someone has a dire need for it
they can contribute work on the next version.  Preferably
as an optional table.
 
> 3.      frClcmiAddressLen is not adequate.  Since the MIB is going to be
>         changed, I suggest the same definition as frLportAddrDLCILen from
>         RFC 1604 be used.
> 

The existing draft object is also in RFC1315 and is defined as:

         frDlcmiAddressLen OBJECT-TYPE
             SYNTAX  INTEGER {
                     two-octets (2),
                     three-octets (3),
                     four-octets (4)
             }
             ACCESS  read-write
             STATUS  mandatory
             DESCRIPTION
                "This variable states which address  length  in
                octets.  In the case of Q922 format, the length
                indicates the entire length of the address  in-
                cluding the control portion."


The RFC1604 version is:

 frLportAddrDLCILen OBJECT-TYPE
     SYNTAX  INTEGER {
      twoOctets10Bits(1),
      threeOctets10Bits(2),



Since this object is in RFC1315 and should not be deprecated
(since there are lots of RFC1315 implementations for agents
and managers) this object must remain as it
is currently defined.




> 4.      The MIB does not support different quality of service parameter
>         (CIR, Bc, Be) on a per direction basis.  Is this a problem for the
>         future.
> 

I don't know of any DTE vendor (or switches) that support
it in different directions.  Again if someone has a need
they can express it and contribute to the next version.

> 5.      The following statistics would be helpful in the frDlcmiTable:
> 
>   frDlcmiLinkRelErrors OBJECT-TYPE
>      SYNTAX  Counter32
>      MAX-ACCESS  read-only
>      STATUS  current
>      DESCRIPTION
>              "The number of user-side local in-channel
>              signaling link reliability errors (i.e., non-
>              receipt of Status/Status Enquiry messages or
>              invalid sequence numbers in a Link Integrity
>              Verification Information Element) for this UNI/NNI
>              logical port.  If the logical port is not
>              performing user-side procedures, then this value
>              is equal to noSuchName."
>      REFERENCE "ITU Q.933 Annex A and ANSI T1.617 Annex D"
>      ::= { frDlcmiEntry 13 }
> 
>   frDlcmiProtErrors OBJECT-TYPE
>      SYNTAX  Counter32
>      MAX-ACCESS  read-only
>      STATUS  current
>      DESCRIPTION
>              "The number of user-side local in-channel
>              signaling protocol errors (i.e., protocol
>              discriminator, message type, call reference, and
>              mandatory information element errors) for this
>              UNI/NNI logical port.  If the logical port is not
>              performing user-side procedures, then this value
>              is equal to noSuchName."
>      REFERENCE "ITU Q.933 Annex A and ANSI T1.617 Annex D"
>      ::= {frDlcmiEntry 14 }
> 
>   frDlcmiChanInactive OBJECT-TYPE
>      SYNTAX  Counter32
>      MAX-ACCESS  read-only
>      STATUS  current
>      DESCRIPTION
>              "The number of times the user-side channel was
>              declared inactive (i.e., N392 errors in N393
>              events) for this UNI/NNI logical port. If the
>              logical port is not performing user-side
>              procedures, then this value is equal to
>              noSuchName."
>      REFERENCE "ITU Q.933 Annex A and ANSI T1.617 Annex D"
>      ::= { frDlcmiEntry 15 }
> 

In my opinion all of the above statistics will put an 
additional burden on the DTE to monitor what the
DCE is doing wrong.  Maybe usefull for debugging Certification
tests, but not very usefull for real world cases.  The
existing Fault handling in the draft MIB is my opinion
adequate and easy to implement.

> 6.      What is the interrelationship between frCircuitMulticast and
>         frDlcmiMulticast?  Can frCircuitMulticast be set to anything
>         other than unicast if frDlcmiMulticast is nonBroadcast?  Can the
>         same interface have both unicast and multicast DLCI?  If so,
>         what is the setting of frDlcmiMulticst set?
> 

This one goes back a while.  If any DLCI runs multicast, then
frDlcmiMulticast is set to true.  Yes an interface can have
both unicast and multicast.  The frDlcmiMulticast is more of
an indicator of whether the implementation can support multicast 
or not, not necessarily whether they have any DLCIs that 
are multicast or not.

> 7.      What provisions are being made for SVCs or is it out of the
>         scope of this MIB modification?
>

As far as I'm concerned it is beyond the scope as this MIB
is going out for last calls.  We never saw any contributions
on SVCs during the drafting of this MIB that I can recall.

I fully support moving forward on SVCs in the next version
but not in this version.
 
> 8.      The following statistics would be useful either as part of frDlcmiTable
>         or part of a separate table:
> 
>    frErrorUserLinkRelErrors OBJECT-TYPE
>        SYNTAX Counter32
>        MAX-ACCESS read-only
>        STATUS current
>        DESCRIPTION -- (from RFC1604 frMgtVCSigUserLinkRelErrors)
>              "The number of user-side local in-channel
>              signaling link reliability errors (i.e., non-
>              receipt of Status/Status Enquiry messages or
>              invalid sequence numbers in a Link Integrity
>              Verification Information Element) for this UNI/NNI
>              logical port.  If the logical port is not
>              performing user-side procedures, then this value
>              is equal to noSuchName."
>        REFERENCE "ITU Q.933 Annex A and ANSI T1.617 Annex D"
>        ::= { frErrorEntry 1 }
> 
>   frErrorUserProtErrors   OBJECT-TYPE
>        SYNTAX Counter32
>        MAX-ACCESS read-only
>        STATUS current
>        DESCRIPTION  -- (from RFC1604 frMgtVCSigUserProtErrors   )
>              "The number of user-side local in-channel
>              signaling protocol errors (i.e., protocol
>              discriminator, message type, call reference, and
>              mandatory information element errors) for this
>              UNI/NNI logical port.  If the logical port is not
>              performing user-side procedures, then this value
>              is equal to noSuchName."
>        REFERENCE "ITU Q.933 Annex A and ANSI T1.617 Annex D"
>        ::= { frErrorEntry 2 }
> 
>   frErrorChanInactive   OBJECT-TYPE
>        SYNTAX Counter32
>        MAX-ACCESS read-only
>        STATUS current
>        DESCRIPTION  -- (from RFC1604  frMgtVCSigUserChanInactive )
>              "The number of times the user-side channel was
>              declared inactive (i.e., N392 errors in N393
>              events) for this UNI/NNI logical port. If the
>              logical port is not performing user-side
>              procedures, then this value is equal to
>              noSuchName."
>        REFERENCE "ITU Q.933 Annex A and ANSI T1.617 Annex D"
>        ::= { frErrorEntry 3 }
>

RFC 1604 supports the above statistics.  I think that requiring
those types of statistics for a DTE does not give enough
benefits for the space it requires for code and data on the
Frame Relay DTE agent code, let alone what benefit it gives
for the Manager as far as debugging problems.  The existing
ERR table is adequate in my opinion.
 
>   frErrorRcvdShort   OBJECT-TYPE
>        SYNTAX Counter32
>        MAX-ACCESS read-only
>        STATUS current
>        DESCRIPTION
>            "The number of frames that were to short to be legal
>             frame relay frames"
>        ::= { frErrorEntry 4 }
> 
>   frErrorRcvdLong   OBJECT-TYPE
>        SYNTAX Counter32
>        MAX-ACCESS read-only
>        STATUS current
>        DESCRIPTION
>             "The number of frames that were longer than the length
>              specified in the ifTable by ifMtu for this interface."
>        ::= { frErrorEntry 5 }
> 
> 

The above two statistics look like better candidates to lobby
them to be added to the next version of the Interface MIB.  Although
personally I don't think they are important enough to add.

>    frErrorRcvdIllegalDLCI   OBJECT-TYPE
>        SYNTAX Counter32
>        MAX-ACCESS read-only
>        STATUS current
>        DESCRIPTION
>            "The number of frames received with an illegal Data Link
>             Connection Identifier (DLCI)."
>        ::= { frErrorEntry 6 }
> 
>   frErrorRcvdUnknownDLCI   OBJECT-TYPE
>        SYNTAX Counter32
>        MAX-ACCESS read-only
>        STATUS current
>        DESCRIPTION
>            "The number of frames received with an unknown Data Link
>             Connection Identifier (DLCI)."
>        ::= { 6fdnVsFrErrorEntry 7 }
> 

I can't again see the benefit of these ones either.  It seems
that we are trying to police the DCE errors.

> Regards,
> Roy Spitzer
> 
> ==========================================================================
> 
> Roy Spitzer                             Roy.Spitzer@adn.alcatel.com
> Alcatel Data Networks                   (703) 724-2412
> 44983 Knoll Square                      FAX: (703) 724-2005
> Ashburn, VA 20147
> 
> 
>

I can see no reason for not pushing forward the
current draft "as is" for publishing as an RFC.


Regards,

Alan

**********************************************************
Alan K. Bartky                       Email:alan@xylan.com
Principal Software Engineer
XYLAN Corporation                    Voice:(714) 597-8042
15707 ROCKFIELD BLVD STE 155F        FAX:  (714) 597-8342
IRVINE CA 92718-2830
**********************************************************