RE: CONSENSUS TEST: Fragmentation handling

[Stephen Kent <kent@bbn.com>]

At 2:34 PM +0300 4/8/04, Tero Kivinen wrote:
>Stephen Kent writes:
>>  if neither #2 or #3 is a SHOULD, then I would like to add text that
>>  every implementation MUST implement at least one of these, to give us
>>  a decent chance of having a way to accommodate fragments for
>>  port-specific SAs. in fact, maybe that is the best way to state this,
>>  given the current set of comments on this topic.
>
>Then we can have two implementations both implementing different parts
>of that MUST and they do not interoperate. Also the case #3 can also
>be used with IPv6, and the case #2 cannot securely be used with it, so
>I think it is better to have case #3 as SHOULD and case #2 MAY. I do
>not think we need requirement for "MUST" for at least one, as there
>will be implementations which do not care about port selectors and/or
>fragments.
>--
>kivinen@safenet-inc.com

I too would prefer one MUST, for interoperability, be we can't get 
consensus on that, hence my suggestion.

Your final comment above is problematic:
"as there will be implementations which do not care about port 
selectors and/or fragments."

Implementations MUST support port selectors; it was a 2401 
requirement and it is a 2401bis requirement. The question is how to 
handle the port selector specific SAs in contexts where fragmentation 
occurs.

Steve