Re: [Acme] Clarification about Identifier Validation Challenges

[Romain Fliedel <[e-mail]>]

Any feedback ?

2015-07-27 12:40 GMT+02:00 Romain Fliedel <[e-mail]>:

> Hi,
>
> I've been reading the spec, and I don't really understand why the reply to
> validation challenges is not designed the same way as other api. If I
> understand the spec correctly, instead of sending a JWS reply, the jws is
> embedded in a json object containing 'type' and 'validation'.
>
> For example the dvsni challenge response is :
> {
>   "type": "dvsni',
>   "validation": {
>     "header": { "alg": "HS256" },
>     "payload": "qzu9...6bjn",
>     "signature": "xxxxxxxxxxxxxxxxxx"
>   }
> }
>
> Why don't use a regular JWS as body for this challenge reply ?
> In that case the reply would be:
> {
>     "signature": "xxxxxxxxxxxxxxxxxxx",
>     "protected": "eyJub25jZ...In0",
>     "header": {
>         "alg": "RS256",
>         "jwk": {
>             "kty": "RSA",
>             "n": "ox33_lEk....Eg9zM",
>             "e": "AQAB"
>         }
>     },
>     "payload": /* in cleartext for readability */ {
>        "type": "dvsni",
>        "token": "fgf...gfdg"
>     }
> }
>
> Also for the dns challenge I don't understand why there is
> "clientPublicKey" attribute in the reply.
> Still regarding dns challenge I am concerned about the length of the
> "signature" generated when RS256 is used to sign the JWS object with a 4096
> bits key. It will then exceed the maximum txt record length. maybe using a
> hash of this signature would solve this ?
>
>