draft-ops-endpoint-mib-08.txt

[Keith McCloghrie <kzm@cisco.com>]

Hi,

In a MIB being developed for MPLS tunnels, a source IP address is used
as part of the identification for such a tunnel.  The particular IP
address is the one specified in the MPLS signalling message.  The
signalling messages allows for IPv4 or IPv6 addresses, but not DNS
names.  So, the SYNTAX of the object needs to allow for either
IPv4 or IPv6, but not DNS names.  Thus, it needs a TC which allows a
choice of just the two address types.  It appears to me that
draft-ops-endpoint-mib-08.txt does not provide this, since it allows a
choice of three (IPv4, IPv6, and DNS Name) and it recommends against
subtyping:

   The InetAddressType and InetAddress objects SHOULD NOT be subtyped.
   Subtyping binds the MIB module to specific address formats, which
   may cause serious problems if new address formats need to be
   introduced. Note that it is possible to write compliance statements
   in order to express that only a subset of the defined address types
   must be implemented to be compliant.

It is not appropriate in this case to use compliance statements,
because the use of DNS names is not possible due to the need to use the
same format as the signalling messages.

Now, "SHOULD NOT" does not prohibit usage if there is a good reason, as
I believe there is in this case.  That is, it seems to me that this is
a situation in which a variation from the recommendation is warranted.

However, I suspect this kind of situation will occur again.  So,
perhaps the rationale behind the use of "SHOULD NOT" should be
re-examined, and either changed, or at least have additional 
explanatory text.

Keith.