Stemming the heartbleed

The Internet community got a wake-up call last week when news emerged that a bug in a piece of encryption software had been lurking in cyberspace for two years.

Announced last Tuesday, the bug, known as Heartbleed, lets hackers obtain user names, passwords and even encryption keys from websites that use OpenSSL, an open source software used to secure a majority of online communications.

As Heartbleed has stayed undetected until now, it is possible that the bug could have already been exploited to fish for data from websites that use OpenSSL.

Bloomberg reported last Friday that the U.S. National Security Agency (NSA) knew about the Heartbleed bug and regularly used it to gather critical intelligence, an allegation that the NSA has denied.

In any case, there is no way of knowing if the bug has been exploited, as Heartbleed leaves no traces.

From a technical perspective, the Heartbleed vulnerability is a missing “bounds check” error, a type of mistake that is common and easy to make with the C programming language used to develop OpenSSL.

So how is this a problem? When you are at a secure server of, say, an online banking website, you will see a green padlock in your browser’s address bar, indicating that your connection is secure.

To check that the server end of the connection is still there, your browser would normally state the amount of “heartbeat” data that it would send to the server, and send that amount of data over.

The server would then return that same data, keeping the connection “alive”.

With the Heartbleed bug, the server would not perform a “bounce check” to verify that the actual amount of data it received from the browser matches the stated amount.

So, a hacker behind a Web browser could potentially send less data than what was stated, forcing the server to retrieve additional data from the server’s working memory – which could contain passwords and encryption keys – to make up for the difference.

With OpenSSL being used widely to secure much of Internet communications, some people have questioned if it is wise to leave its development to a core team of open source developers. OpenSSL is developed by four people, and only one of them counts it as his full-time job.

However, the beauty of open source comes from the fact that anyone, including tech companies and governments, can take part in the development or fund the efforts.

“Some argue that open source is safer, because the more eyes review the code, the less vulnerabilities there will be”, says Sami Petäjäsoja, Asia-Pacific vice president at Codenomicon, the security firm that discovered the bug, along with engineers at Google.

“Open source makes it possible for any user to do their due diligence before deployment, including security testing and code review. If something is discovered and the software is fixed, it will help the whole community in the long run, as is the case with Heartbleed. In a closed source, similar vulnerabilities can stay hidden and only be known to malicious actors,” he adds.

The bigger question is this: when should severe vulnerabilities like Heartbleed be made public? Should big Internet companies get the heads-up before others?

While content distribution networks such as Akamai, which secures and hosts a large portion of the Internet’s content, were notified before everyone else, others like Yahoo, Amazon and some governments were caught off-guard.

The Canadian government, for example, has shut down public access to some of its websites which may be affected by Heartbleed.

To be sure, there are no roadmaps to follow in the disclosure of software flaws. It is often a judgment call for developers to make.

Announcing a loophole like Heartbleed publicly before it is plugged on a majority of websites will be catastrophic. Disclosing it privately to some organisations and not others will leave the uninformed out in the cold.

With much of the global economy dependent on the Internet, a more egalitarian way of managing and disclosing security threats that plague critical parts of the Internet’s infrastructure is crucial, before malicious hackers have time to act.

To stem any bleeding, the Internet’s blood needs to clot a lot faster.