Vulcan: Lessons on Reliability of Wearables through State-Aware Fuzzing Edgardo Barsallo Yi Heng Zhang Amiya K. Maji
[email protected] Purdue University Purdue University Purdue University West Lafayette, IN West Lafayette, IN West Lafayette, IN
[email protected] [email protected] Kefan Xu Saurabh Bagchi Purdue University Purdue University West Lafayette, IN West Lafayette, IN
[email protected] [email protected] ABSTRACT Wear OS relative to Android. The fundamental driver of the dif- As we look to use Wear OS (formerly known as Android Wear) ferences is the limited display area and the difficulty of executing devices for fitness and health monitoring, it is important to eval- interactive work (such as typing) on a wearable device. As a result, uate the reliability of its ecosystem. The goal of this paper is to wearable apps tend to have more number of Services (which run in understand the reliability weak spots in Wear OS ecosystem. We the background) relative to Activities (which run in the foreground), develop a state-aware fuzzing tool, Vulcan, without any elevated have fewer GUI components, and have tethering to a counterpart privileges, to uncover these weak spots by fuzzing Wear OS apps. app on the mobile device [32]. Moreover, wearable devices are of- We evaluate the outcomes due to these weak spots by fuzzing 100 ten fitted with a variety of sensors (e.g., heart rate monitor, pulse popular apps downloaded from Google Play Store. The outcomes oximeter, and even electrocardiogram or ECG sensor) each with include causing specific apps to crash, causing the running app its own device driver software.