Adobe apps have been in the bull's eye as one of the most popular attack vectors, especially in the form of malware-laden PDF files. Adobe has been beefing up its security posture during the past year, starting with its automatic updater feature for Reader and Acrobat and a regular patching schedule. Another big piece of Adobe's security strategy is hardening its code against attacks, says Brad Arkin, senior director of product security and privacy. "Sandboxing was one of the early technology candidates we looked at" for code-hardening, he says.
The new Protected Mode feature is based on the sandboxing technique used by Microsoft in Office 2010 and by Google in its Chrome browser. The first release of the sandboxing feature will keep all "write" calls inside the protected space so that if a PDF is infected with malware, the malicious code can't spread outside that file to the system itself or to other files. The second phase of the Protected Mode release for Reader will prevent attackers from reading information on the victim machine's system via an infected PDF. "It will be harder for the bad guys to install malicious software or anything persistent on a computer that could survive from one reboot to [the next]," Arkin says.
Sandboxing basically quarantines any malware that's embedded in a PDF. "Real-world attacks today are almost always some variant of a memory trespass vulnerability," Arkin says. Sandboxing, which restricts what an attacker is able to do with a PDF file, should eliminate many of those types of threats, he says.
But the cat-and-mouse game between security experts and attackers will likely just move to another vector, he says. "Once we put this out, attackers will either move to another application or they might look for creative ways to work around it," he says.
Security experts welcomed Adobe's adoption of sandboxing, but they don't expect it to dissuade attackers from continuing to hammer away at Adobe apps. "Reader will continue to be a really big target for a while, even with the sandbox environment," says Andrew Storms, director of security operations at nCircle. "But we need to congratulate Adobe for another tool to help protect ourselves and our users."
Storms says the true weakest link is still the user. "The idea is to use the tools the vendors are giving you, but you still have to apply your own risk assessment," he says.
Adobe's Arkin says Adobe worked with Microsoft Office's security team, the Google Chrome team, as well as other outside experts in implementing the sandboxing technology in Reader. It's based on Microsoft's practical Windows sandboxing technique, and works like this in Reader: when Reader needs to execute any action not allowed in the sandbox such as launching an attachment inside a PDF file from another app, for example, that action gets vetted by a broker process with specific policies that prevent access to any functions that could open the system up to attack.
The goal is to stop attackers from writing files, changing registry keys, or installing malicious code via PDFs -- but without the user noticing any difference in the Reader app, Arkin says. "There's no performance impact, no change to the interface, no dialog box – it's just happening in the background. And users don't change the way they interact with PDFs," he says.
Sandboxing does not, however, prevent phishing or clickjacking attacks, he notes. "There's nothing we can do to protect against users who follow" links in infected PDFs purportedly from their banks, for example, he says.
Just how long sandboxing techniques will stand up to attacks is unclear. nCircle's Storms says attackers probably will come up with ways to subvert sandboxing much the way researchers have done with Microsoft's Address Space Layout Randomization (ASLR) and Data Execution Protection (DEP) security features. "The question is how quickly Adobe, Microsoft, and Google react" to any challenges to the technique, he says.
Have a comment on this story? Please click "Discuss" below. If you'd like to contact Dark Reading's editors directly, send us a message.