Closed marsella closed 2 years ago
The server has selected the key_id
and should be storing the association between the key_id
and the ciphertext received in Step 3.ii in their database already. This is why the client doesn't send the key_id
to the server in Step 3.ii.
That is, there is nothing for the server to check here. The server does not know the key used to encrypt the ciphertext send in Step 3.ii, so would be unable to check any associated data the client were to send, anyway.
The bug that I see is that we should make storage of the association between the received ciphertext and key_id
explicit in Step 4.ii.
Does this make sense to you, @marsella?
Ah, yes it does make sense. Thanks for clarifying.
Right now the client goes to the key server to get a key id, then generates a key and sends it to the key server to store. There's not currently a step where the key server makes sure that the ciphertext they send is associated with the same key id it produced.
I think this should be explicit in 4.i.