SSL Users Receive A Wakeup Call

Posted by on Jan 30, 2009 | 3 Comments

This bothers me while still encouraging my overall view behind SSL in the long term. As you can see from the link above, SSL has essentially for lack of a better description…been “hacked”. By coming up with a means of creating a fake Certification Authority, a group of hackers have discovered that SSL can be overtaken with greater ease than we might have ever thought previously.

Now for the good news. Duplicating this is not too likely, at least according to the hackers who are refusing to release the exact key and code used to make the magic happen. Still, the idea of exploiting  a weak MD5 cryptographic algorithm in digital signatures and certificates is going to keep me up at night for quite sometime.

It is one thing to simply claim that you will not share what I deem to be very dangerous information with those who might opt to use it for nefarious purposes, but it is also quite another to believe this claim of “doing the right thing”. In the end, I am glad the weakness is public and sincerely hope that those who are in charge of our data security using SSL are working with these individual hackers to find a means of dealing with the MD5 issue itself. Clearly, this does not make me want to put a lot of faith into SSL until this problem has been resolved. Well, at least until the next exploit is figured out that is.

  • http://blog.ivanristic.com Ivan Ristić

    You’ll be happy to learn that Verisign fixed the problem some time ago:

    https://blogs.verisign.com/ssl-blog/2008/12/on_md5_vulnerabilities_and_mit.php

    Ivan

  • http://www.matthartley.com Matt Hartley

    Ivan: Awesome to hear, thanks for the update…wonder why more news agencies did not cover this fix?

  • Mark Hill

    While what you say is true, to someone who doesn’t read the article, it sounds as if SSL depends on MD5 exclusively. MD5 has been hacked but SHA-1 has not, so the only possible forged certificate would be signed only with MD5. What needs to happen is to remove the use and trust of MD5. While all MD5-only certificates are being replaced, all browsers need to be updated to flag the risk when they depend on such a cert. Then at a chosen date in the future, all the MD5-only certs need to be revoked.

    In the future, another digest algorithm needs to be developed and included in addition to SHA-1 so that when SHA-1 is hacked, we have another fallback to bridge us over again.