Re: [Suit] Fwd: New Version Notification for draft-housley-suit-cose-hash-sig-03.txt

[Russ Housley <[e-mail]>]

Tony:
>  
> Thanks for the new version of the draft.. It addresses all my comments fully except for the question of 'kid'. The point that I was trying to make there was the LMS public key (unlike other public keys) has a well-defined identity, called 'I' in [HASHSIG]. I agree that you can't be prescriptive about this, but I suggest strengthening your new bullet point at the end of section 4 to:
>  "If the 'kid' field is present, then it MAY contain the identity of the public key (called 'I' in [HASHSIG])."
>  
> The situation that I'm thinking of here is one where the IoT device has more than one HSS/LMS public key as trust anchor. Because public key parameters 'I' and 'q' are integral parts of the verification steps, the device may have to attempt verification multiple times and anything which can improve the chances of picking the right public key first time saves computation (energy).

I totally misunderstood your comment, and now that I do, I'm not sure that I agree.  That said, maybe I am still not fully understanding your point.

I think the kid would identify a tree, not a node in the tree.  The signature value itself contains the information necessary to locate the node within the tree.

Russ