On 7/5/2016 10:33 AM, Black, David wrote:
> Joe,
>
>>>    Implementations MUST NOT use option kind 69 unless and until it is
>> replace 69 with TBD
>>
>>>    assigned to TCP-ENO by IANA.  In the meantime, implementations MUST
>>>    use experimental option 253 [RFC6994], to which IANA has assigned
>>>    ExID 0x454E (encoded by decimal bytes 69, 78 in Figure 1).
>>>    Conversely, after IANA assigns a dedicated option kind to TCP-ENO,
>>>    the use of option 253 is deprecated.
>> The above is OK.
>>
> Is it also worthwhile to have this draft record the unauthorized use of
> option kind 69 for a prior protocol that is superseded by TCP-ENO?    This
> would be only as a description of what's happened.
If and when 69 is assigned - yes.

But not before and not if a different codepoint is assigned.  It's
sufficient to say that "all previous variants of this protocol are
superceded by this", which applies both to content and codepoints.


> I can argue this both ways - not blessing prior bad behavior vs. recording 
> what
> happened.   The latter provides IANA with a reference for the unauthorized
> use and would enable this draft to at least discourage such use.   OTOH,  
> having been in an analogous situation in the past, I'm not optimistic about
> ever reclaiming that option kind value.
>
> Right now, it looks like the intent is not to ask for option kind 69 - if 
> that were
> to be requested, this draft ought to have additional text on how a receiver
> distinguishes TCP-ENO from past use of that option kind (and what happens
> if the use of option kind 69 that shows up doesn't match what the receiver
> expects).
That seems like to would be a very appropriate punishment for the behavior.

Or worse, being assigned a codepoint others are squatting on.

Sauce for the goose, as it were.

IMO, if you pee in the well, you'd better be ready to drink from it.

Joe


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

Reply via email to