The discovery spec specifies mechanisms for discovering ACIs and public keys, which allows implementations (rkt and the like) to present users with choices about adding those keys to a trust store of some sort to gate ACI execution.
It could be valuable to specify a mechanism to distribute (discover) key revocation certificates to allow implementations to build features for discovering revocations automatically, updating key stores, and rotating ACI signing keys.
provide ACI signers a way to publish key revocations in a way that is discoverable to tools.
allow users executing ACIs with tools which can check for or discover revocations (due to key loss, theft, or compromise). Perhaps on the next ACI discovery attempt or via polling.
support signing an ACI with a new key and revoking an old key, in order to perform a key rotation
Alternately, maybe tools should piggyback on existing key infrastructure and query key servers. It would be left to implementations to decide how to check the health of a trusted key and how often to do so. Related prior art: certificate revocation lists and OCSP.
The discovery spec specifies mechanisms for discovering ACIs and public keys, which allows implementations (
rkt
and the like) to present users with choices about adding those keys to a trust store of some sort to gate ACI execution.It could be valuable to specify a mechanism to distribute (discover) key revocation certificates to allow implementations to build features for discovering revocations automatically, updating key stores, and rotating ACI signing keys.
Maybe something like:
I can see this being valuable to:
Alternately, maybe tools should piggyback on existing key infrastructure and query key servers. It would be left to implementations to decide how to check the health of a trusted key and how often to do so. Related prior art: certificate revocation lists and OCSP.