SNMPv2 Classic+ overview

[Jeff Case <[e-mail]>]

SNMPv2 Classic+

One proposal under consideration by the working group consists of a series
of 17 documents (labeled SNMPv2 Classic+ for convenience):

	draft-ietf-snmpv2-adminv2-ds-02.txt, draft-ietf-snmpv2-bcm-ds-01.txt
	draft-ietf-snmpv2-coex-ds-02.txt,    draft-ietf-snmpv2-conf-03.txt
	draft-ietf-snmpv2-conf-ds-02.txt,    draft-ietf-snmpv2-intro-ds-02.txt
	draft-ietf-snmpv2-ip-ds-02.txt,      draft-ietf-snmpv2-mib-ds-02.txt
	draft-ietf-snmpv2-party-ds-02.txt,   draft-ietf-snmpv2-proto-ds-02.txt
	draft-ietf-snmpv2-scm-ds-01.txt,     draft-ietf-snmpv2-sec-ds-02.txt
	draft-ietf-snmpv2-smi-ds-02.txt,     draft-ietf-snmpv2-tc-ds-03.txt
	draft-ietf-snmpv2-tcp-ds-02.txt,     draft-ietf-snmpv2-tm-ds-02.txt
	draft-ietf-snmpv2-udp-ds-02.txt

These documents have been through the refining fire of implementation and 
trial deployment and constitute the only proposal which has benefitted from
these experiences.  Recent implementation and deployment experience has shown
them to be sound, stable, implementable, and deployable.  The addition of
BCM and some other changes to simplify the configuration task, in conjunction
with GUI-based configuration tools, have addressed to a large extent many of
the concerns about difficulty of configuration. 

The documents are quite solid, but based on the implementation and trial
deployment, a few changes are recommended.  The complete editorial directions
suggested changes are described more fully in a separate note to follow this 
message, but the brief descriptions are as follows:

	1.  a small modification to constants used by maintenance functions

	    this change allows SNMPv2 Classic+ implementations to interoperate
	    with existing SNMPv2 Classic (RFC 1441-1452) implementations

	2.  automatic clock acceleration mechansisms are invoked if and only
	    if the discrepancy between the received timestamp value and the
            value in the LPD is greater than or equal to one second

	    this change obviates observed leapfrog behaviors which result from
            sub-second accelerations coupled with rounding effects.

	3.  the SCM document is deleted from the series and the intro document
	    is edited accordingly to remove references to it

	    this change is based on deployment feedback which has shown that
	    BCM is more simple to implement, deploy, and use than is SCM

	4.  minor editorial corrections, for example, to correct multiple
	    incorrect section cross-references between and within documents

	    this change makes reading the documents much easier for the 
	    uninitiated -- new readers are unnecessarily confused by the
	    documents, in part, because of numerous references to a section
	    3.1 or 3.2 which are now 5.1 and 5.2, etc

In spite of these minor needed changes, the Working Group owes the
editor a large debt of thanks because the documents are remarkably solid.

The point of enumerating these suggested changes is twofold.

First, the small number of changes clearly illustrates the quality of the
work and shows that responding to destablilizing pressures was unnecessary.
The work was done and the quality of the work compares well with many of the
emerging proposals.

Second, since these documents are used as the basis of many of the current
emerging proposals, it is important to avoid duplication of these errors.

regards,
jdc