2015 Data Science Workshop

Wed, April 29, 2015
Location: Fisher Conference Center, Arrillaga Alumni Center

"Secure Analytics on the Internet of Things"

Abstract:

The high-level, long-term project goal is to: Research and define new cryptographic computational models for end-to-end secure analytics and actuation on enormous streams of real-time data from the Internet of Things. Achieving this goal requires expertise and research in a wide range of disciplines, including cryptography, analytics, security, software, networking and hardware design. The research, if successful, will enable a new class of IoT applications that can analyze huge streams of sensor data without knowing either what the data is or even what the result is. When data first leaves a sensor, it is encrypted. Novel algorithms will take encrypted data as input and produce encrypted data as output and be unable to decrypt either. Only an end user, who has the proper key, can decrypt the data and see the actual result of the computations.

The initial application is an embedded network of smart shower heads, taps, and other water outflows in a Stanford dormitory.These connected devices will provide precise, fine-grained information on when, how much, and what temperature water is used.The devices will have an integrated Bluetooth Low Energy to communicate with student phones.This phase of the work will focus on four components:I. The sending device and data collection network.II. The analytics engine for the streams of data.III. Algorithms that use the analytics engine to learn important insights about water use.IV. Scientifically exploring interventions that lead people to conserve water.

Security and privacy is a cross-cutting concern in components I – III (sensors to analytics).For the application to preserve privacy and security end-to-end, we will need to define precise privacy models and ensure that each component maintains them.