The Internet of Safe Things

Abstract:

As devices in the world around us are increasingly connected, the possibilities about how to use these devices and the data have exploded. Many of these scenarios such as home automation and office productivity require software tools beyond data visualizations, to express user intents, check for bugs, and ensure interactions do not cause harm to the users and physical world. However, the complexities of the user and environmental contexts these devices operate within present enormous challenges to the software development, testing and verification. I will describe our work on (1) Contextual Fuzzing, a cloud based mobile app testing service that systematically explores environmental conditions (such as networking, sensor input) and user interactions at scale, and (2) SIFT, a safety-centric app synthesis and verification platform for IoT devices that checks whether conflicts or policy violations can arise in the apps. I will demonstrate the power of Contextual Fuzzing and SIFT through user studies and deployments, and highlight the research challenges and opportunities for safe and secure IoT systems.