5

Where I work, currently for internal sites, self signed certificates are being used. In Firefox, the info regarding the certificate are displayed:

site domain: example.com
Owner: Site doesn't supply owner info
verified by: company name

I am curious as to how vulnerable self signed certificates are to a man in the middle attack, such as SSL strip. The obvious vulnerability, I can think of is that users would be used to trusting the certificate so an SSL strip wouldn't produce any new warnings in Firefox (correct me if I'm wrong). Also in a man in the middle attack, is it possible for the attacker to display the same exact certificate to end-users (also verified by: the same company name)?

1
  • 1
    The biggest problem with using self signed certificates in this way is that you train users to ignore the warnings in the browser about a certificate not being valid. A MITM attack on your internal site might be unliklely, but more likely on another site, and if users blindly ignore the untrusted certificate warning they won't notice it.
    – JonnyWizz
    Nov 27, 2015 at 14:17

3 Answers 3

12

Instead of using self-signed certificates for internal sites, I suggest you set up an internal CA and use it to sign certificate requests. You can import your CA's certificate into each browser used internally (either manually or via automated deployment mechanisms) and this should be as good security-wise as using certificates issued by a major CA -- as long as you take proper measures to secure the system used for signing certificate requests.

This is a much better solution than self-signed certificates and, depending on your environment, your domain controller may already offer CA services (MS Active Directory, FreeIPA, etc).

6

A self-signed certificate is a certificate which is not verifiable by virtue of it being signed by a trusted CA. So you have to be able to check that this is the "right" certificate by some other mean. Otherwise, an attacker could redirect your connection to a fake server (sporting another self-signed certificate), and from there run a man-in-the-middle attack.

If you can make sure once that the self-signed certificate is correct, then you can install it in the "trusted store" of your browser, which will ensure security for all subsequent connections(*). To a large extent, this mimics the usual SSH security model, which Unix sysadmins are quite happy with. One possible way, for the first connection, to check that the certificate sent by the server is the right one, is to ask your browser to display the certificate "thumbprint" (a hash value computed over the whole certificate) and phone the expected server sysadmin for comparison with the reference thumbprint.

(*) Note the tricky point: security is ensured as long as you never accept an unverified self-signed certificate -- pay heed to the browser warnings !

2

As long as you add your self signed certificate to the browsers trusted store, a self signed certificate is for all intents and purposes the exact same as a certificate signed by a certificate authority.

0

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .