Embedded Software Implementation of Privacy-Preserving Computation

Number of Students : 2
Guide : Utsav Banerjee

The growth of IoT has led to concerns about the lack of privacy of data collected, processed, and communicated by embedded systems. Homomorphic Encryption (HE) is the most versatile cryptographic solution for privacy-preserving computation. However, its high memory demands and computational bottlenecks make it expensive to implement on embedded devices. We will implement efficient software for HE-based encryption, decryption and computation on various embedded micro-controller platforms, e.g., ARM Cortex-M4, ARM Cortex-M7, etc using the open-source SEAL-Embedded library. Platform-specific algorithmic optimizations will be explored. Various standard computations will be benchmarked in the encrypted domain, while maintaining side-channel security requirements.