Re: [IPsec] AD review of draft-ietf-ipsecme-qr-ikev2-08

[Tero Kivinen <[e-mail]>]

Paul Wouters writes:
> On Wed, 6 Nov 2019, Valery Smyslov wrote:
> 
> > Do you think the current diagrams are confusing?
> 
> Yes. Because often I go back to RFCs and only look at the diagrams
> expecting it to be what I need to implement. So for optional/required
> payloads, I would mostly look at the diagram, and perhaps read a bit
> of text.

That is the reason we added Appendix C in the IKEv2.

So my proposal is to leave the exchanges inside the text as they are,
but add new Appendix that has the different exchanges including the
optional payloads.

> >> That is, the diagrams should represent the state machine, not an
> >> example of the state machine.
> >
> > Hmmmm... It's an open question :-) Aa a counter-example,
> > the EAP and non-EAP case of IKEv2 are not shown
> > on the same diagrams - these are different diagrams,
> > however the state machine for IKE_AUTH is the same.
> 
> Sure.

In RFC7296 Appendix C we do have C.2 IKE_AUTH Exchange without EAP,
and C.3 IKE_AUTH Exchange with EAP. And I would say that the state
machine for IKE_AUTH for them are different, the state machine for
IKE_SA_INIT is same for both and is not included in C.2, or C.3, both
of them use the IKE_SA_INIT from C.1
-- 
[e-mail]