ipsec January 2011 archive
Main Archive Page > Month Archives  > ipsec archives
ipsec: [IPsec] IKEv2 Diffie Hellman retry logic

[IPsec] IKEv2 Diffie Hellman retry logic

From: Gaurav Poothia <gpoothia_at_nospam>
Date: Sun Jan 16 2011 - 21:01:02 GMT
To: "ipsec@ietf.org" <ipsec@ietf.org>

Scenario: When the IKEv2 initiator guesses an incorrect DH group and the responder sends back the DH group hint in INVALID_KE_PAYLOAD notification.

Couple of questions around this:

On what basis does the responder reject the DH group:

1. Because the best match initiator SA payload proposal (against responder policy) has a different DH group from KE payload

2. Because the responder after looking all the SA payload initiator proposals with DH group from KE payload finds none of the initiator proposals acceptable

3. Because the responder altogether ignores the initiator proposals (SA payload) and only checks to see that the DH group in KE payload doesn't figure in its own policy at all

To paraphrase:
Case 1 looks like it will have IKEv1 parity in terms of using the best policy match and restarting negotiation if the initial KE guess doesn't match up to that.
Case 2 will do worse than IKEv1 by not forcing the best policy match but by proceeding with an inferior and acceptable match will save an extra round trip.
Case 3 is actually non deterministic because the hint is not guaranteed to work (since other transforms have not been evaluated while choosing hint)

Once rejected on what basis does the responder choose the DH group to put in the INVALID_KE_PAYLOAD hint (corresponding to above rejection criteria):

* For cases 1 & 2: It is the DH group in the initiator SA proposal that facilitates the best policy match (against responder policy).

* For case 3 it the DH group in responder's most preferred proposal.

Thanks

_______________________________________________
IPsec mailing list
IPsec@ietf.org
https://www.ietf.org/mailman/listinfo/ipsec