Researchers Target Shared Memory to Hack Android Apps

Researchers at the University of Michigan and the University of California, Riverside have developed a new attack method that can be used to obtain sensitive information from applications running on Android and likely other operating systems.

Dubbed “UI state interference attack,” the method relies on exploiting what experts call a “newly- discovered public side channel,” namely the shared memory, which enables various processes running on the operating system to share data between them. In this case, the problem is a design flaw in graphical user interface (GUI) frameworks that can be used to determine every UI state change through this shared memory, which can be accessed without special permissions.

“The assumption has always been that [the apps installed on a device] can’t interfere with each other easily,” explained Zhiyun Qian, an assistant professor at UC Riverside and one of the researchers involved in the project. “We show that assumption is not correct and one app can in fact significantly impact another and result in harmful consequences for the user.”

The researchers have showed that an attacker can use a malicious Android app, which only requests minimal permissions, to harvest data entered by users into other applications. The attacker’s program runs in the background and monitors changes in shared memory. These changes are then correlated to various activities performed by the victim, such as logging in to an account or taking a picture.

The findings have been tested on Android, but researchers believe that the same principle can also be applied on OS X, iOS and Windows.

Experts conducted tests on the applications for Gmail, H&R Block, WebMD, Newegg, Chase Bank, Hotels.com and Amazon. With a success rate of up to 92%, they managed to use their malicious app to hijack data from these programs as it was inserted by the user. They claimed a success rate is as follows: Gmail (92%), H&R Block (92%), Newegg (86%), WebMD (85%), Chase Bank (83%), Hotels.com (83%) and Amazon (48%).

Through this method, an attacker can obtain login credentials and any other sensitive data entered into the applications, and even pictures — for example, the scanned copy of a check made with the Chase app.

The researchers are presenting their paper (PDF), titled “Peeking into Your App without Actually Seeing It: UI State Inference and Novel Android Attacks,” at the USENIX Security Symposium in San Francisco on Friday. The team has also published some videos to demonstrate how the attack works.