Vulnerabilities / Threats
8/1/2013
10:19 AM
Connect Directly
Twitter
Twitter
RSS
E-Mail
50%
50%
Repost This

Creating Browser-Based Botnets Through Online Ad Networks

Researchers demonstrate how ads invoking JavaScript on viewers' browsers en masse could create untraceable networks to wreak DDoS damage

LAS VEGAS -- BLACK HAT USA -- For several years security researchers and black hat hackers have fine-tuned methods of manipulating the eccentricities and vulnerabilities of the way browsers work to make user machines visit certain sites, download illegal content, and even carry out attacks like SQL injection without the user knowing it. However, these attacks have always been thought of as invoking one-off behavior that wouldn't scale well enough to leverage for something like a distributed denial-of-service attack (DDoS). But yesterday at Black Hat USA, a pair of researchers showed it is possible to maneuver browsers on a massive scale through online advertising.

Click here for more of Dark Reading's Black Hat articles.

In a demo at their session, WhiteHat Security researchers Jeremiah Grossman, CTO and founder, and Matt Johansen, manager of the firm's threat research center, showed it's possible to essentially create a hard-to-trace browser botnet that can easily trigger DDoS with a minimal investment in a fake online ad. As they explained, networks that serve up advertisements on ad-supported sites across the Internet frequently allow their advertisers to run arbitrary JavaScript on browsers displaying their ads. Using JavaScript to make hundreds of thousands or millions of advertising viewers connect at once to a particular target site could quickly create enough connections to take down most sites on the Web.

"The Web runs on advertising -- that's how all these websites are paid for," Grossman said. "So the reach of these advertising networks is phenomenal."

In their demonstration, Grossman and Johansen purchased $20 worth of advertising impressions through an unnamed advertising network and placed an innocuous-looking ad with a call to automatically load an external site where they could change their JavaScript payload on the fly.

"The hardest part of all this research was the approval process. Not for the reasons you'd think -- they happened to not be very good at reading JavaScript, or even caring about JavaScript. What they actually cared about was that the ad looked pretty and worked like an ad," Johansen said. "But anytime we wanted to tweak something, like change a URL, it had to go through reapproval. So instead of putting the code directly in the ad, we just put script source and sourced it out to a file on our side."

The researchers stood up an Apache server on AWS to crash it in front of the audience within a few seconds of targeting their script-running browsers toward it.

"This whole time we did not hack anybody. We just used the way the Web works and took down our own service," Johansen said. "We stayed completely on the legal side here. But you can kind of get an idea of how this could get fun if you didn't."

Not only could malicious hackers do much more damage with more malicious code, but "there's no particular reason why the bad guys couldn't use a stolen credit card" to carry out this kind of attack, Grossman said.

According to Johansen, the advantage to an attacker of using this method is its disappearing footprint.

"So why not just do a traditional denial-of-service attack? It's not persistent. It goes away," he says. "There's no trace of this -- we put the money in the machine, the JavaScript gets served up, and then it goes away. And it's very, very easy." Have a comment on this story? Please click "Add Your Comment" below. If you'd like to contact Dark Reading's editors directly, send us a message. Ericka Chickowski specializes in coverage of information technology and business innovation. She has focused on information security for the better part of a decade and regularly writes about the security industry as a contributor to Dark Reading.  View Full Bio

Comment  | 
Print  | 
More Insights
Register for Dark Reading Newsletters
White Papers
Cartoon
Current Issue
Video
Slideshows
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2013-1421
Published: 2014-04-22
Cross-site scripting (XSS) vulnerability in Craig Knudsen WebCalendar before 1.2.5, 1.2.6, and other versions before 1.2.7 allows remote attackers to inject arbitrary web script or HTML via the Category Name field to category.php.

CVE-2013-2105
Published: 2014-04-22
The Show In Browser (show_in_browser) gem 0.0.3 for Ruby allows local users to inject arbitrary web script or HTML via a symlink attack on /tmp/browser.html.

CVE-2013-2187
Published: 2014-04-22
Cross-site scripting (XSS) vulnerability in Apache Archiva 1.2 through 1.2.2 and 1.3 before 1.3.8 allows remote attackers to inject arbitrary web script or HTML via unspecified parameters, related to the home page.

CVE-2013-4116
Published: 2014-04-22
lib/npm.js in Node Packaged Modules (npm) before 1.3.3 allows local users to overwrite arbitrary files via a symlink attack on temporary files with predictable names that are created when unpacking archives.

CVE-2013-4472
Published: 2014-04-22
The openTempFile function in goo/gfile.cc in Xpdf and Poppler 0.24.3 and earlier, when running on a system other than Unix, allows local users to overwrite arbitrary files via a symlink attack on temporary files with predictable names.

Best of the Web