2 ms·
I don't get it. This can't be about decrypting traffic between internal clients and external servers, because the bank doesn't have the RSA key in the first pl
by amluto 10y ago
I don't get it. This can't be about decrypting traffic between internal clients and external servers, because the bank doesn't have the RSA key in the first place. It also seems unlikely that it's about decrypting traffic between external clients and internal servers, because (a) that's not what the bank rep said and (b) the server would just log the traffic.
That leaves two cases: internal client <-> internal server and internal client <-> internal MITM. For both of these cases, I see two easy solutions:
1. Log the session key.
2. Define a custom TLS extension that contains the session key encrypted with against a known bank-controlled public key and have the server/MITM send this extension after the handshake. Problem solved.
In any event, the internal client <-> external server case is highly problematic. What's to stop, say, Snapchat from running an extra unauthenticated DH exchange in JavaScript and encrypting all the application layer traffic against the resulting session key? The MITM proxy won't notice unless it's very clever indeed and then all those nice MITM proxy logs are useless.