Hi John So as I interpret your response, the trade-off is that kid is more flexible as it may be used for key types other than RSA (e.g. symmetric keys) whereas x5t is more efficient if the implementation will be limited to RSA keys. Fair summary?
adam On Mon, Apr 6, 2015 at 10:35 AM, John Bradley <[email protected]> wrote: > kid is a generic name that can be looked up in local keys stores for any > sort of key and value is determined by the signer as part of key exchange. > x5t indicates that it is a thumbprint that can be calculated based on the > public key and doesn't need to be communicated separately. > > Some keystors have specific API to find RSA public keys based on the x5t > value. > > They both assume that the recipient have the key in some sort of local > keystore or some other way to go and retrieve the key remotely by getting a > JWKS. > > One way to look at it is that x5t and x5t#S256 are specific ways to create > a collision resistant kid. Knowing the alg used lets you do some extra > things to optimize looking it up. > > John B. > > On Apr 6, 2015, at 7:44 AM, Adam Lewis <[email protected]> > wrote: > > > > Hi, > > > > The difference of when to use kid vs. x5t to identify the JWT signing > key is not obvious to me. On the surface they seem to do the same thing, > (e.g. identify to the party validating the signature which key to use to > validate it, allow the public key to be retrieved from an endpoint of the > party doesn't already have it). > > > > But JWS defines both, so I'm guessing that there is a subtle difference > that I am missing here? > > > > > > > > adam > > _______________________________________________ > > jose mailing list > > [email protected] > > https://www.ietf.org/mailman/listinfo/jose > >
_______________________________________________ jose mailing list [email protected] https://www.ietf.org/mailman/listinfo/jose
