Re: IKE V2 Open Issues

[Uri Blumenthal <uri@bell-labs.com>]

Derek Atkins wrote:
>>Yes, your way it's clearer. However Charlie's way is simpler,
>>and one less payload type to worry about...
> 
> IMHO payload types are cheap.

Hm, OK.

> I can go either way.  I don't feel extremely vetted to one way or the
> other.
> 
> However, another benefit for using two payload types: it makes it
> easier for protocol analyzers like tcpdump or ethereal.  They can
> differentiate the cookie request N(COOKIE_REQUIRED{cookie}) from a
> cookie response N(COOKIE{cookie}) to aid in analysis and debugging...
> A small benefit indeed, but a tangible one for, IMHO, little
> additional coding.  You have to have the code to parse the packet
> either way -- whether you look for IKEV2_NOTIFY_COOKIE or
> ..._COOKIE_REQUIRED is a one-line change.

OK, sold. I'm convinced in the value of COOKIE_REQUIRED and
support it.