Precision control and accumulated error eliminating method applied to fully homomorphic encryption

Abstract

The invention discloses a precision control and accumulated error eliminating method applied to fully homomorphic encryption. The method comprises the following steps of splitting a plaintext; encrypting each constituent part and computing to acquire a result ciphertext; performing homomorphic comparison computation on the result ciphertext; performing a reserving operation or a carrying and reserving operation via a comparison computation result; and acquiring a result plaintext via decryption and rounding operations, thus achieving precision control on a polynomial fully-homomorphic encryption method based on coefficient mapping conversion. After such operation, each byte part of the computation result ciphertext can be controlled within a precision range of an original scheme, and afterthe decryption and rounding operations, the result plaintext can be completely accurate. The single ciphertext computation can be completely accurate, and the any times of continuous ciphertext computation also can be completely accurate.

Description

technical field [0001] The invention relates to the technical field of information security, in particular to a precision control and cumulative error elimination method for fully homomorphic encryption. Background technique [0002] The polynomial fully homomorphic encryption method based on coefficient mapping transformation proposed in Chinese patent 201510192143.1 operates on ciphertext data, which is essentially an approximate operation, so how to control the accuracy of the operation is an important issue. Especially after multiple consecutive ciphertext operations, the cumulative error will gradually increase, so that the resulting ciphertext significantly deviates from the correct value after decryption, thus affecting the actual use of the method and system. [0003] Compared with the precise operation characteristics of the fully homomorphic encryption method on the lattice, which has been extensively studied by academics, the polynomial fully homomorphic encryptio...

AI Technical Summary

Problems solved by technology

Especially after multiple consecutive ciphertext operations, the cumulative error will gradually increase, so that the resulting ciphertext significantly deviates from the correct value after decryption, thus affecting the actual use of the method and system

[0003] Compared with the precise operation characteristics of the fully homomorphic encryption method on the lattice, which has been extensively studied by academics, the polynomial fully homomorphic encryption algorithm based on the coefficient mapping transformation adopts the principle of approximate operation in the ciphertext operation, so that it can guarantee the security of the ciphertext. At the same time, the overall efficiency of the ciphertext operation is greatly improved, and the volume of the ciphertext and its expansion speed can be well controlled, but at the same time, the price paid by the approximate operation is the loss of precision.

[0004] The accuracy problem of homomorphic operation of single-pass ciphertext is mainly due to the approximate operation characteristics of the original scheme of the polynomial fully homomorphic encryption method based on coefficient mapping transformation, in which the higher the sampling density of the operation support function is, the higher the operation accuracy is, but At the same time, the volume of computing support functions will expand rapidly, which will have adverse effects on network storage and transmission.

However, if the density is too low, the operation accuracy cannot even meet the single operation requirements of the ciphertext with low effective digits.

Images