Malisa Vucinic writes:
> Ok, I see your point. If you don’t think we should design a new one,
> is there something already specified that does not require 10+
> exchanges that we could use?

Depends on your requirements and what kind of authentication and
authorization you are doing.

The example I gave with IKEv2 + EAP-AKA is 3 exchanges between JN and
JA (exchange = message pair, i.e. one request and one reply) and 2
exchanges between JA and JCE.

The exchanges between JN and JA would most likely be fragmented as
they will not fit 100 bytes, so actual 802.15.4 frames count is even
higher. The exchanges between JA and JCE can most likely be fit in
802.15.4 frames, so they are single frames.

I am not expert in EAP methods, so I have not really checked what can
be done there, but they usually do require two exchanges between JA
and JCE, as there needs to be some kind of challenge from JCE to be
sent to the JN through JA, and JN needs to reply to that, and only
after that JCE can signal JA and JN that authentication and
authorization is successful.

There might be some method where the challenge might be generated in
such way that JA could generate it and forward both the challenge it
generated and the reply from JN to the JCE, but in that case I think
that puts more trust on the JA by the JCE, and I do not know if there
us such EAP method, and how secure it is.
-- 
[email protected]

_______________________________________________
6tisch mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6tisch

Reply via email to