IETF IPCDN Meeting
Vancouver, British Columbia
Thursday November 10, 2005

Reported by Richard Woundy (richard_woundy@cable.comcast.com)

WG Meeting Summary and Administration

The IPCDN WG met to discuss the progress of various MIB documents, especially the outstanding PacketCable/IPCablecom and DOCSIS internet-drafts.

The DOCSIS BPI+ MIB was published as RFC 4131 in September. The DOCSIS QoS MIB and the DOCSIS Event Notification Management MIB are on the RFC Editor Queue.

Most of the meeting discussion centered on resolving open issues with the PacketCable/IPCablecom MTA MIB (syntax of Kerberos realm names and realm organization names, and a textual convention for encryption algorithms), and with the PacketCable/IPCablecom Management Event MIB (whether event throttling applies to local logging on the cable modem). The PacketCable/IPCableCom NCS Signaling MIB, DOCSIS Cable Device MIB, and DOCSIS RFIv2 MIB were also discussed, with a focus on the next steps to get to the 'Publication Requested' state for all five internet-drafts.

There is a strong desire by the WG to conclude all DOCSIS and PacketCable/ IPCablecom MIB drafts by around December 2005 or January 2006.

There were about ten attendees at the WG session in Vancouver. Total meeting time was one hour.

PacketCable/IPCableCom MTA MIB

draft-ietf-ipcdn-pktc-mtamib-07.txt

Eugene Nechamkin and Jean-Francois Mule produced version 07 of the MTA MIB draft, to address comments from the AD review (Bert Wijnen) and from the IPCDN mailing list. The resolution of these comments were posted on the mailing list. Most comments were resolved with simple editorial fixes. One significant addition is specification of the same encryption algorithms as the BPI+ MIB (RFC 4131, ref: DocsBpkmDataEncryptAlg).

The WG discussion was centered on the proper MIB object SYNTAX for Kerberos realm names and Kerberos realm organization names.

With respect to the Kerberos realm organization name (pktcMtaDevRealmOrgName), X.509 and RFC 2459 constrain the length of the organization name to 64 characters. The characters may be encoded as UTF8String, so each character may occupy between one and six octets. Therefore, a single organization name may occupy up to 384 octets. The current MIB object has a SYNTAX of SnmpAdminString(SIZE (1..255)).

The WG recommends to change the pktcMtaDevRealmOrgName object as follows:

1. Remove the sentence referring to the "o=" prefix in the DESCRIPTION clause. Do not include the "o=" prefix in the value of this MIB object.
2. Change the SYNTAX of this object to LongUtf8String, imported from the SYSAPPL-MIB defined in RFC 2287. This enables a maximum size of 1024 octets (for future growth).
3. Add a compliance statement for this object to limit the maximum size of the object to 384 octets, and include justification, e.g. each UTF-8 character can be up to 6 octets long, and the organization name may contain up to 64 UTF-8 characters.

Three distinct MIB objects contain a Kerberos realm name: pktcMtaDevProvKerbRealmName, pktcMtaDevRealmName,and pktcMtaDevCmsKerbRealmName. RFC 4120 specifies a syntax of GeneralString for the Kerberos realm name, although the RFC says the realm name may be non-ASCII (UTF8String) in the future. The WG recommendation is to continue to use SnmpAdminString as the SYNTAX for these three MIB objects.

This draft also has a new textual convention for MTA configuration file encryption algorithms, PktcMtaDevProvEncryptAlg. This textual convention enables future support of triple-DES and AES encryption, similar to the DocsBpkmDataEncryptAlg textual convention in the DOCSIS BPI+ MIB (RFC 4131). It was suggested to compare the PktcMtaDevProvEncryptAlg and DocsBpkmDataEncryptAlg textual conventions, and if they are identical, then have the MTA MIB IMPORT DocsBpkmDataEncryptAlg from the BPI+ MIB.

Editor's note: the DocsBpkmDataEncryptAlg includes 40-bit and 56-bit DES encryption but not 64-bit DES encryption as in PktcMtaDevProvEncryptAlg. These textual conventions are NOT identical nor compatible.

The next steps for this internet-draft are:

1. Bert Wijnen will respond to the author's email and confirm whether or not the draft addresses the AD review comments.
2. The authors will incorporate the recommended changes to the pktcMtaDevRealmOrgName MIB object in version -08.
3. The co-chairs will compose a PROTO write-up and request to put the draft into the 'Publication Requested' state.

PacketCable/IPCableCom NCS Signaling MIB

draft-ietf-ipcdn-pktc-signaling-09.txt

Gordon Beacham, Satish Kumar, and Sumanth Channabasappa submitted version 09 of the NCS Signaling MIB draft, to address comments from the MIB doctor and various WG participants. The resolution of these comments were posted on the mailing list.

Significant updates included changes to a number of MIB objects in the pktcSigDevMultiFreqToneTable, and the removal of the pktcNcsEndPntConfigTxGain and pktcNcsEndPntConfigRxGain objects.

There were no comments concerning this internet-draft in this meeting.

The next steps for this internet-draft are:

1. The WG co-chairs will review this draft.
2. The co-chairs will compose a PROTO write-up and request to put the draft into the 'Publication Requested' state.

PacketCable/IPCableCom Management Event MIB

draft-ietf-ipcdn-pktc-eventmess-05.txt

Wim De Ketelaere, Eugene Nechamkin, and Sumanth Channabasappa submitted version 05 of the PacketCable/IPCableCom Management Event MIB draft, responding to comments from a review by Greg Nakanishi and additional feedback from Randy Presuhn. The resolutions of these comments were posted on the mailing list.

There was some discussion in the meeting about whether the event throttling objects (pktcDevEvThrottleAdminStatus, pktcDevEvThrottleThreshold, and pktcDevEvThrottleInterval) apply to local logging of events, or only to network transmissions of events via SNMP or SYSLOG messages. The current throttling MIB objects only refer to SNMP and SYSLOG, but pktcDevEventDescrReporting and pktcDevEvLogType refer to local logging.

The WG recommends adding the following text to the DESCRIPTION of pktcDevEvThrottleAdminStatus: "The impact of event throttling controls on local logging is an implementation-specific matter."

Editor's note: The following DESCRIPTION of pktcDevEvLogType suggests that the throttling controls do NOT apply to local logging. If throttling of local logging is implementation-specific, then at least this DESCRIPTION should be revised. Bert Wijnen requests that there be explicit text in the relevant DESCRIPTION clause(s) for what folks (e.g. developers) can expect.

pktcDevEvLogType OBJECT-TYPE
       SYNTAX      BITS {
                   local(0), 
                   syslog (1), 
                   trap (2), 
                   inform (3) 
                   } 
       MAX-ACCESS  read-only 
       STATUS      current 
       DESCRIPTION 
               "This MIB Object contains the kind of actions taken by  
                the MTA when the event under consideration occurred. 
    
                A bit with a value of 1 indicates the corresponding 
                action was taken. Setting it to a value of 0 indicates  
                that the corresponding action was not taken. 
     
                An event may trigger one or more actions (e.g.: Syslog  
                and SNMP) or may remain as a local event since  
                transmissions could be disabled or inhibited as defined  
                by the Throttle MIB Objects." 
                 
       ::= { pktcDevEventLogEntry 7 }

The next steps for this internet-draft are:

1. The WG co-chairs will review this draft.
2. The co-chairs will compose a PROTO write-up and request to put the draft into the 'Publication Requested' state.

DOCSIS Cable Device MIB (replacing RFC2669)

draft-ietf-ipcdn-device-mibv2-10.txt

Kevin Marez and Richard Woundy produced version 10 of the DOCSIS Cable Device MIB draft, to address comments from the AD review (Bert Wijnen) and from the MIB doctor review. The resolution of the AD review comments were posted on the mailing list. Most comments were resolved with simple editorial fixes.

Two significant changes in version 10 discussed in the meeting were the new narrative text for MIB module IMPORTs and REFERENCE clauses (section 3.1.1), and the new narrative text discussing the persistance model for cable modem MIB objects (section 3.1.2). There were no significant comments on the IMPORT/REFERENCE narrative text, although this is different from the approach taken by other IPCDN internet-drafts. Bert noted that there were no MUST statements in the persistance model narrative, but that was OK since the normative persistance requirements are in the MIB object definitions.

The next step for this internet-draft is for the IPCDN co-chairs (in this particular case, Jean-Francois Mule) to compose a PROTO write-up and to request to put the draft into the 'Publication Requested' state.

DOCSIS RFIv2 MIB (replacing RFC2670)

draft-ietf-ipcdn-docs-rfmibv2-14.txt

Eduardo Cardona produced version 14 of the DOCSIS RFIv2 MIB, in response to comments from the AD review (Bert Wijnen).

There were no comments concerning this internet-draft in this meeting.

The next steps for this internet-draft are:

1. The WG co-chairs will review this draft.
2. The co-chairs will compose a PROTO write-up and request to put the draft into the 'Publication Requested' state.

IPCDN Next Steps