Trellis++ a Practical Privacy-Preserving Food Safety Framework
Primary Investigator:
James Krogmeier
Servio Palacios, Aaron Ault, James Krogmeier, Bharat Bhargava
Abstract
The Internet of Things is ubiquitous. As IoT data volumes increase in a privacy-conscious world, an alternative model where provable computation happens closer to the data is needed. This project proposes to move parts of the computational kernels to the edge of the network to take advantage of the computational capabilities of the edge nodes. Unfortunately, including the edge in the computational resources can lead to a higher risk of data leakage or theft of confidential data. To combat this, and to support privacy-preserving computation and analytics, this project develops a practical computational framework and aggregation techniques to address a critical issue on edge computing: producing auditable computations that also prevent theft of confidential data. In particular, this project aims to prove the safety of food through its lifecycle computing on encrypted data to obtain proof of safety while keeping all these data private to the requirements of the data owners.
In this work, we propose Trellis++: a practical privacy-preserving food safety framework capable of performing auditable computation on private data, reporting real-time aggregate certification status to downstream participants without disclosing underlying private data.