Re: [Webpush] Application server authentication new years edition

[Costin Manolache <[e-mail]>]

On Thu, Jan 7, 2016 at 1:57 PM, Martin Thomson <[e-mail]>
wrote:

> On 8 January 2016 at 05:59, Costin Manolache <[e-mail]> wrote:
> > 1. Key identifiers ('kid') - for normal JWT, the key is known from the
> > issuer or
> > 'sub'. Since in our case we'll have self-generated keys, and we want to
> > avoid
> > having each key registered with all providers - all that a push service
> will
> > know
> > when seeing a JWT token the first time will be the content of the token.
> > So we need a way to find the public key needed to verify the content -
> and
> > the only one I know is to use something like the SHA of the public key.
> > On subscribe - the provider can store the public key, with the kid as
> lookup
> > key, than
> > on send it can verify. For example kid == (SHA1 or SHA256 truncated to
> > 64bits) would work.
>
> The push service is going to need the public key, so that means
> including a JWK in the message, I think.  Knowing a hash of the public
> key isn't enough.
>

The push service can store the key when subscribe() is called.

I think we can make it work without a key ID - the send URL will decode
to the registration, which would include the authorized public key. But
I think it would still be a bit cleaner and simpler if the JWT key could be
verified
without having to use the URL.



>
> I think that we can avoid using key identifiers.  We could include one
> and use that to reference another header field, but it is easier to
> say that there is exactly one JWT and exactly one JWK and that one
> depends on the other.
>

That would work too, but the push service still need to verify that the
key in subscribe is the same as the key in JWT. If you load the
subscription -
using either the URL or the key id - there is no need to send the JWK.



> > 2. 'aud' field - not sure what would be the right value, maybe the
> domain of
> > the push provider ?
>
> I specified URL at the suggestion of Chris, but I had the same thought
> you did.  Maybe we can trim it to origin.  That reduces portability
> considerably and also shortens the token somewhat.
>
> > 3. Going back to the 'voluntary'/optional part: I'm not sure if
> encryption
> > is going to be required or
> > optional in the final version ?
>
> The protocol won't absolutely require it, though it will strongly
> recommend it.  (I'd prefer to require it, but there you have it.)  The
> W3C API will require encryption.  Does that make sense?
>

Yes.


> > 4. I think we can have the sender key optional in the subscribe
> operation,
> > in particular for cases
> > like low end IOT - not including it would give permission to anyone with
> the
> > URL and pubkey
> > of the device to send messages. There is another case where it helps -
> > pairing, in cases of D2D -
> > but with restrictions ( either time, or binding the subscription to the
> > sender at first use ).
> > In other words: I'm not opposed to having 'optional' on the subscribe
> (and
> > on contact info).
>
> I agree.  Making the JWT mandatory on the subscription would introduce
> some operational problems for the cases where one endpoint has to be
> used by several different application servers too.
>
>
It shouldn't be mandatory - just recommended that subscribe indicates who
is authorized
to send.

A push service that doesn't require authentication can just ignore it.



> > 5. This may be a bit controversial :-), if we require the developers to
> use
> > a JOSE library to
> > generate JWT tokens, signed with a ES256 - wouldn't make sense to also
> allow
> > them to use
> > the same library to encrypt the payload for the e2e ? There is already a
> > content type defined, so easy to
> > differentiate.
>
> I don't know of any way to do that without also massively inflating
> the size of the payload.  There is a mapping to JWE already in the
> encryption draft for those who would prefer to reuse their library.
> The tricky part is key derivation, but I think that's largely
> unavoidable in this case.
>

Yes, it doesn't seem possible to convert from JWE to the binary format,
so UAs would need to support both, I don't think this would work. And it
may be
more confusing in the end.

Costin