Mike Jones writes:
For your point "4) Thumbprint formats" if you or someone else wants to
define an additional thumbprint format for use in IoT contexts (or any
other contexts), I encourage you to write an Internet Draft that does
so, registering the new header parameter defined in the JSON Web
Signature and Encryption Header Parameters registry.
That can of course be done, but I would have hoped the initial
version of the specification would also be usable in the IoT
context, where the use of raw public keys will most likely arise.
If what you want is a thumbprint over a raw key, see the individual
submission draft
https://tools.ietf.org/html/draft-jones-jose-jwk-thumbprint-01,
which defines a method for doing this. The -01 version incorporates
working group feedback from Toronto. In Toronto, I'd asked whether
the working group wanted to adopt it as a working group draft and a
decision hasn't been made on that yet. If this would be useful for
IoT applications, that would be good to know.
That looks ok for the jwk use, but for the hash over the SPKI parts of
the X.509 is better because that is already used in other places. I.e.
if you want to create fingerprint that can be used to match the key
used in other protocols, they are not using that format defined in
your draft, thus you need to regenerate the JWK format from their
internal public/private key formats and generate new hash.
For example DANE x 1 x format (i.e. 3 1 1 for SHA-256, or 3 1 2 for
SHA-512) defined RFC6698 section 2.1.3 are calculated over the exact
same binary object which is transmitted in the raw keys used in the
TLS (RFC7250 section 3), which is again same binary object used in the
in the IKEv2 (draft-kivinen-ipsecme-oob-pubkey).
In the IoT context it will most likely be quite common to define the
configuration of who can connect to you by using list of hashes of
raw public keys. I.e. the device has list of hashes, and when
connection comes (either over TLS or IKEv2 or whatever), then that raw
public key sent inside the connection protocol is hashed and it is
matched against that list of hashes. If match is found, the connection
is allowed, if not the connection is dropped. Now json might be one
way of this configuration could be transmitted to the IoT device, thus
ability to be able to represent hashes in the format that makes it
possible to match the binary blobs used on the wire, would be useful.
One of the reasons the SPKI is used, that it can also be extracted
from the self-signed certificate, i.e. early implementations might use
self-signed certificates in the TLS (for example) before the RFC7250
implementations come out.
The draft-jones-jose-jwk-thumbprint format is in such format that it
is quite hard to match that against the binary blob we get from the
wire, as to do so would require to format the public key received to
JWK and then calculating the fingerprint of that newly created object.
Parsing SPKI format (and parsing JSON also if we use that for
configurations) is required in the implementations anyways, but in
normal case the IoT devices do not need code for generating JSON
objects.
So I do not think the format you are specifying there is suitable for
IoT uses, but I assume it will still be useful in the JWK in general.
--
kivinen(_at_)iki(_dot_)fi