Paper 2018/1249

Two round multiparty computation via Multi-key fully homomorphic encryption with faster homomorphic evaluations

NingBo Li, TanPing Zhou, XiaoYuan Yang, YiLiang Han, Longfei Liu, and WenChao Liu

Abstract

Multi-key fully homomorphic encryption (MKFHE) allows computations on ciphertexts encrypted by different users (public keys), and the results can be jointly decrypted using the secret keys of all the users involved. The NTRU-based scheme is an important alternative to post-quantum cryptography, but the NTRU-based MKFHE has the following drawbacks, which cause it inefficient in scenarios such as secure multi-party computing (MPC). One is the relinearization technique used for key switching takes up most of the time of the scheme’s homomorphic evaluation, the other is that each user needs to decrypt in sequence, which makes the decryption process complicated. We propose an efficient leveled MKFHE scheme, which improves the efficiency of homomorphic evaluations, and constructs a two-round (MPC) protocol based on this. Firstly, we construct an efficient single key FHE with less relinearization operations. We greatly reduces the number of relinearization operations in homomorphic evaluations process by separating the homomorphic multiplication and relinearization techniques. Furthermore, the batching technique and a specialization of modulus can be applied to our scheme to improve the efficiency. Secondly, the efficient single-key homomorphic encryption scheme proposed in this paper is transformed into a multi-key vision according to the method in LTV12 scheme. Finally, we construct a distributed decryption process which can be implemented independently for all participating users, and reduce the number of interactions between users in the decryption process. Based on this, a two-round MPC protocol is proposed. Experimental analysis shows that the homomorphic evaluation of the single-key FHE scheme constructed in this paper is 2.4 times faster than DHS16, and the MKFHE scheme constructed in this paper can be used to implement a two-round MPC protocol effectively, which can be applied to secure MPC between multiple users under the cloud computing environment.