[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: draft-ietf-dnsext-rfc2536bis-dsa-06.txt review



At 12:42 AM 3/27/2006, Eastlake III Donald-LDE008 wrote:
Hi Mike,

See below at @@@


Trimmed all of the text - readers can see previous emails in this thread if they want more detail.


Don -

If I asked you a question of "how do I form a valid DNSKEY record containing a DSA key" I believe you would claim that this ID taken with 4034 provides sufficient detail to do so. I disagree. What I get from 4034 is the general format of a DNSKEY RR. What I get from this ID is the ordering of bits within a key blob of some form. What I am missing is what format identifier (e.g. Algorithm identifier) applies to this ID and where this key blob explicitly fits within the DNSKEY RR. If you're saying this isn't the job of this ID, then my question is "where is the ID that maps this ID into the DNSKEY RR"?

I refer to the Algorithm Identifier field as a format id because that's really what it is. It describes the format of the PublicKey field and needs to reference a specific document to do so rather than the more generic algorithm name. To understand what I mean, let's consider the RSA algorithm. The RSA public key has two common representations - the Public/Exponent representation and the Chinese Remainder Theorem representation. Neither of the two RSA related Algorithm IDs in appendix a.1 of 4034 provide such representation. I would need to define a third (or more) RFC with this format and the assigned algorithm ID would refer specifically to that RFC.

You're proposing obsoleting 2546 and 4034 points at 2546 as being the specific document which defines the format for Algorithm ID 3. You need a replacement document which links that algorithm ID (3) to a specific format. Its just a simple as that. This ID does not do that.


Also, the more I think about it, the less I like the current system of assigning a single number for both the public key representation and an associated signature representation. We're about to re-do RSA with SHA1 into RSA with SHA256 and possibly other SHAs. I would expect similar actions for DSA. I suggest we split the algorithm assignments for key and signature as follows:



Algorithm IDs for Key representations:

0 - Reserved
1 - RSA (modulus/exponent form)
2 - DH
3 - DSA (Fips 182-6)
4 - ECC - not yet defined
5 - RSA (modulus/exponent form - same as 1)

Algorithm IDs for Signature representations
1 - RSA/MD5 - not recommended
2 - none (was DH)
3 - DSA with SHA1
4 - Reserved
5 - RSA/SHA1
6-127 reserved
127-255 - new signature algorithms.


--
to unsubscribe send a message to namedroppers-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>