Research On Lightweight Methods Of Two Kinds Of Encryption Components Based On SPN Structure

With the development and wide application of Internet of Things technology,the security and confidentiality of information systems has become a concern of more and more people.Lightweight encryption technology can provide security solutions for the reliable operation of the information system in resource-constrained environment,but smaller equipment means more strict area limitation.Therefore,how to design a more efficient and compact cryptographic algorithm under the premise of ensuring security has become an urgent problem to be solved.Although the design details of the different kinds of encryption algorithms,encryption parts,such as substitution-permutation network(SPN)structure,and its core component is a linear component with diffusion function and a nonlinear component with confusing function of nonlinear components,so for the two types of password parts design a general optimization method,can be applied to most of the cipher algorithm.This paper is mainly studied from the following two aspects:(1)The optimization of linear components is to reduce the number of XOR gates required to realize the diffusion matrix.The Shortest Linear Programs(SLP)is NP-hard.Since the optimal solution cannot be obtained within the effective time,heuristics are needed to obtain the approximate optimal solution.Based on the classical heuristics BP and Paar,the computational range of4)vector in BP algorithm is optimized,and a heuristic based on frequency-first is proposed.Taking frequency as the criterion of new base selection,the gate with the highest reuse rate is selected as far as possible.In the selection stage,the combination of output targets can be directly matched by using"preemptive gate strategy"to improve the overall efficiency of the algorithm.In the stage of tie,the double standards of(4 and4)were first used to reduce the number of draws,and then the complete randomization was used to increase the possibility of obtaining the optimal solution and minimize the number of XOR gates required to realize the diffusion function.The algorithm is applied to AES Mix Columns optimization,and the circuit depth is 8 and the number of XOR gates is 95.(2)The optimization of nonlinear components mainly lies in the design of S-box,which requires two steps:multiplication inverse operation on(2~8)and affine transformation on(2).Affine transformation is relatively easy to achieve,but the difficulty lies in the inverse process of(2~8).The multiplication complexity of direct inverse operation is too high,so it cannot be applied to resource-constrained environment.In order to reduce the computation cost,this paper transforms the multiplication inverse operation on(2~8)into a series of simple calculations such as multiplication,power and inverse on(2~4)by using the method of composite fields and optimize the implementation structure of various operations.At the same time,affine matrix with fewer logic gates is used to reduce path delay and area loss,so the S-box is designed which is superior to the same design in circuit area and delay.