A HTTPS site uses a self-signed certificate. A developer argues this is fine since 'the traffic is still encrypted'. What critical protection does a self-signed certificate LACK compared to a CA-signed certificate?
- A.Self-signed certificates lack a path to revocation through the standard CRL or OCSP infrastructure and browsers accept whatever chain the server presents: an operator who loses control of the private key has no way to alert clients unlike a CA-issued certificate that can be revoked and blacklisted globally
- B.A self-signed certificate provides no third-party verification of the server's identity; an attacker performing a MITM attack can present their own self-signed certificate and users have no way to distinguish the real server from the attacker's interception proxy
- C.Self-signed certificates rely on the server's private key for both signing and verification and do not touch an independent root store: this lets an attacker who compromises the web server also reissue trusted-looking certificates for other internal hostnames
- D.Self-signed certificates use the same cipher negotiation and provide identical encryption strength to a CA-issued certificate: the fingerprint is simply not published to any public transparency log and no client ever checks for one