Re: [IPFIX] obsoleting obsoletion in 5102-bis and IE-DOCTORS

On Aug 22, 2012, at 3:27 PM, Paul Aitken wrote:

> Brian,
> 
>> On Aug 22, 2012, at 2:54 PM, Paul Aitken wrote:
>> 
>>> Brian,
>>> 
>>> No objection.
>>> 
>>> Consider VRRP: although VRRP v2 (RFC 3768) was obsoleted by VRRP v3 (RFC 5798), v2 is still in common use.
>>> 
>>> We should expect the same for IPFIX IEs: obsolete IEs may continue in use for many years because people are unaware, unable, or simply unwilling to update.
>>> 
>>> Can IANA record the obsolescence date individually for each field?
>> Yep. From IE-DOCTORS (current -04 working copy):
>> 
>> The revision number of an Information Element is incremented upon deprecation, and the revision date updated.
> 
> Which revision date is that? The overall revision date of the registry?

The revision date of the IE, will clarify.

>>> Reviewing draft-ietf-ipfix-ie-doctors-03,
>>> 
>>>   The use of deprecated Information Elements SHOULD result
>>>   in a log entry or human-readable warning at the Exporting and
>>>   Collecting Processes.
>>> 
>>> 
>>> That's just not possible. How can an existing EP know that its IEs have been deprecated? - Unless it's able to reference IANA, and method (1) of yesterday's discussion is in use.
>> An EP using a deprecated IE sending to a CP using a deprecated IE has in essence no problem -- sure, they're deprecated, but they're interoperating ("living in the past") one could say.
>> 
>> As soon as _one_ of them is updated, though, then one of them SHOULD log (within reason - once per IE per runtime should suffice, but this is an implementation detail).
>> 
>> How about "the use of IEs known by a CP or an EP to be deprecated"...?
> 
> That would work.
> 
> Note that if the EP is to do this, then it knows that it's using a deprecated IE, so it's part-way to answering Benoit's note/objection.

The only case it doesn't cover is old EP (still using a pre-standardization ESIE, and therefore not exporting equivalence) exporting to a new CP (using IANA, and doesn't know anything about the ESIE).


>>> However, that's not desirable since it allows a third party (whoever causes the IANA deprecation) to cause warnings to appear in your tried and tested and perfectly good EP - CP setup.
>> Which is the last remaining question on IE-DOCTORS, and the short version of the other thread: to remove ESIE Reference from IE-DOCTORS or not?
> 
> We need wider community input on this one.

Indeed.

Cheers,

Brian

_______________________________________________
IPFIX mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ipfix