This article is more than 1 year old
Google (finally) adds protection for common Web 2.0 attack
Better late than never
Google has beefed up the security of Gmail and its other services by adding a feature to login pages that blocks one of the more common forms of web attacks.
The upgrade is designed to protect against CSRF, or cross-site request forgery, attacks. The technique subverts basic website defenses by exploiting the often-misplaced trust one site has in another. Over the past three years, vulnerabilities in Gmail, YouTube, and other Google properties have put user accounts at risk of being accessed by miscreants who use the method.
Sometime in the last three days, Google's login pages began setting a cookie with a unique token on each user's browser, according to Mike Bailey, a senior researcher for Foreground Security. That same value is also embedded into the login form. If the two don't match, the user will be unable to log in.
"It's one of those things that people have been telling them to fix for a long time and for whatever reason, they haven't done so until just now," Bailey told El Reg. "They finally implemented the protection that pretty much everybody in the industry recommends they use."
A Google spokesman confirmed that the company added CSRF protection to login pages.
"The security of our users' accounts is very important to us, and we're always looking for ways to make improvements," he said.
While Google is good at quickly squashing reported bugs that put accounts in jeopardy, Bailey and other security researchers have often criticized the web giant for not being more proactive. A case in point: In September 2008, Jeremiah Grossman, CTO of WhiteHat Security, demonstrated a way to use CSRF to learn which YouTube videos a user had viewed.
Grossman's proof-of-concept worked by uploading an Adobe Flash video into Gmail and then duping a victim into accessing it. Because such files are automatically granted access to an entire domain, Grossman was then able to access users' login names, email addresses, and viewing history.
While Google security pros fixed that vulnerability and other CSRF holes, the criticism has been that they were essentially treating the symptoms rather than curing the sickness. What's more, Bailey said, the token-cookie protection has been available on other Google forms but was conspicuously absent from its login page. That's similar to installing an alarm sensor on all the windows of a mansion, but not bothering to put one on the front door.
To be sure, Google, with its millions of users, probably wanted to proceed gingerly with the new protection lest it break the the normal sign-in flow for legitimate users. And it's also fair to point out that Bailey said plenty of other large websites, which he declined to name, have yet to adopt the measure.
The important thing is that Google has finally joined the rest of its security-conscious peers by adding an industry standard for preventing one of the most common forms of attacks plaguing Web 2.0 properties.
"It's our best solution so far," he said. "It's not entirely ideal, but it works well." ®