Posted by Virus Bulletin on Apr 14, 2014
OpenSSL vulnerability has kept the security community busy.
The 'Heartbleed' vulnerability has kept everyone on their toes over the last week or so - hitting the mainstream media, prompting widespread warnings for users to change their passwords, and causing many admins to review the security of their web servers.
Bruce Schneier, who is not known for over-hyping threats, described the severity of the Heartbleed vulnerability as 11 on a scale of 0 to 10. Whatever you think of his use of language, few experts would disagree that Heartbleed was particularly bad.
Since we blogged about it last week, many other good posts and articles have appeared. We thought it worth collating some of the better ones.
CloudFlare's Nick Sullivan (who will speak at VB2014) was among many experts who were sceptical about the possibility of Heartbleed being used to obtain the holy grail: a server's private SSL signing keys. His company set up a vulnerable server and challenged the community to remotely read its private keys from memory. Much to their surprise, several people found the key. One researcher shared his method.
The only right thing to do if you are running a vulnerable (web) server is thus to revoke its SSL certificates. Unfortunately, some web browsers will happily let a user access the site, even when its certificate has been revoked. For some, however, this might be too late anyway: CBC reports that the Canada Revenue Agency has had social insurance numbers of 900 citizens stolen through a successful Heartbleed exploit.
The fact that anyone can look at OpenSSL's code (because it's open source), yet barely anyone did, has not gone unnoticed. John Levine writes that too many large organisations are using OpenSSL because it's free and aren't paying for code audits, while Dan Kaminsky laments that OpenSSL wasn't treated as the critical infrastructure it had become. Steve Marquess, who calls himself the "money guy" at OpenSSL, also says that the project needs a lot more funding to be able to do its job properly.
Of course, the fact that the vulnerability wasn't publicly disclosed until recently doesn't mean that no one had found it. The vulnerability was introduced in March 2012 and Bloomberg cites two anonymous sources who claim the NSA knew about it all the time and had used it regularly to gather critical intelligence - something the NSA subsequently denied.
A silver lining to Heartbleed might be that it will prompt many organisations to improve the security of their servers. F-Secure's Jarno Niemela suggests you review your config standards, while Sophos's Paul Ducklin looks at whether two-factor authentication would have helped.
Heartbleed isn't just a server-side problem though. Clients that run OpenSSL are also vulnerable if they connect to malicious servers, a point made by Rob VandenBrink of the SANS Internet Storm Center, who also lists some applications that use the OpenSSL library.