A Small-Size FHE Scheme for Better Privacy Protection of IoT

Fully homomorphic encryption (FHE) fundamentally solves the problems of confidentiality when data and operations are entrusted to third parties. It can play an important role in secure data computation for Internet of Things (IoT). However, the storage complexities of existing FHE schemes are still not friendly. Our work focuses on optimizing the storage complexity of FHE, in order to further promote the practical process of FHE in IoT. The FHE schemes based on learning with errors (LWE) have become the mainstream of FHE schemes due to its simplicity, security, and ease of efficient implementation. The key switching technology is the ingenious and crucial technique that led to LWE-based FHE. By proposing a low-noise key switching technology, and using Bin-LWE assumption twice (first on the public key, and then again on the ciphertext), we design a leveled FHE scheme with small parameters (without expensive bootstrapping technology). Meanwhile, a detailed analysis of noise growth for homomorphic evaluation is made, and the specific security parameters are given.