I was reading RFC 4251 (really all of the SSH RFCs) and in section 9 it goes into great length about security considerations listing 3 cases of MITM attacks. In several places in the RFCs the authors mention not having the proper infrastructure to distribute server keys from a central authority.

RFC 6187, if I understand it correctly, changes this does it not? Or am I confusing how the pieces interact with each other?

Granted the probability of a MITM attack seems remote in succeeding anyway based off my reading... or does the use of X.509 close some doors but open others?


X.509 is just a format for public key certificates, the type typically used by web browsers. The critical idea was to have the public key in the certificate signed by someone else who's public key you also trust. In web browsers these are the root certificates that come bundled with the browser. Typically companies like Comodo, DigiCert, GlobalSign, etc. They allow you to trust https://acme.com when acme.com's certificate is signed by a trusted root certificate. The same could be done for each ssh host as it is for web hosts, but it tends to be too much work and cost. So, you typically see a ssh 'fingerprint' when you ssh to a new host and a warning if the host changes its public key.

Notice that the public keys used to sign certificates still need to be distributed to the end user. Computer OSes like Windows and Linux also come bundled with root certificates. So the vulnerability in the X.509 scheme is in someone replacing trusted certificates in the software before it is installed on your system.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.