20.7 Insecure renegotiation

In 2009, Marsh Ray and Steve Dispensa, two employees of a company providing a multi-factor authentication solution that was eventually acquired by Microsoft and integrated into Azure, discovered a renegotiation-related vulnerability in then-current TLS versions that allowed Mallory to inject an arbitrary amount of chosen plaintext into the beginning of the application protocol stream [111].

Conceptually, then-current TLS versions were vulnerable to insecure renegotiation because server Alice did not verify whether the source – that is, her communication peer – of the old and the new data in a TLS session was the same.

Using the insecure renegotiation attack, Mallory can inject data that Alice will process as if it came from Bob. For instance, in a web application, Mallory can inject an unauthenticated HTTP request and trick Alice into processing that request in the context of the authenticated user Bob.

Technically, the attack is carried...