Paper 2020/095

SCloud: Public Key Encryption and Key Encapsulation Mechanism Based on Learning with Errors

Zhongxiang Zheng and Anyu Wang and Haining Fan and Chunhuan Zhao and Chao Liu and Xue Zhang

Abstract

We propose a new family of public key encryption (PKE) and key encapsulation mechanism (KEM) schemes based on the plain learning with errors (LWE) problem. Two new design techniques are adopted in the proposed scheme named SCloud: the sampling method and the error-reconciliation mechanism. The new sampling method is obtained by studying the property of the convolution of central binomial distribution and bounded uniform distribution which can achieve higher efficiency and more flexibility w.r.t the parameter choice. Besides, it is shown to be more secure against the dual attack due to its advantage in distinguish property. The new error-reconciliation mechanism is constructed by combining the binary linear codes and Gray codes. It can reduce the size of parameters, and then improve the encryption/decryption efficiency as well as communication efficiency, by making full use of the encryption space. Based on these two techniques, SCloud can provide various sets of parameters for refined security level.