Parameter Optimization for Robust Cryptosystem Performance in Encryption
by MultiThreading.Tech: #1 Publication on Concurrent ProgrammingMarch 15th, 2024
Too Long; Didn't Read
The article analyzes the significance of parameter setup in optimizing encryption performance. It investigates the impact of assistant threads and rounds of confusion and diffusion on cryptosystem speed and statistical performance, highlighting strategies for efficient and robust encryption processes.
Authors:
(1) Dong Jiang, School of Internet, Anhui University, National Engineering Research Center of Agro-Ecological Big Data Analysis and Application, Anhui University & [email protected];
(2) Zhen Yuan, School of Internet, Anhui University;
(3) Wen-xin Li, School of Internet, Anhui University;
(4) Liang-liang Lu, Key Laboratory of Optoelectronic Technology of Jiangsu Province, Nanjing Normal University, National Laboratory of Solid State Microstructures, Nanjing University, Nanjing & [email protected].
Table of Links
6. Parameter Setup
According to the encryption process, clearly, the number of assistant threads n and the rounds of confusion and diffusion r significantly affect the the encryption speed and the statistical performance of the deployed cryptosystems. In all experiments carried out in this paper, n is set to 8, 12, and 32 for laptop, personal computer, and workstation, respectively, and r is set to 5. The reasons for such settings are analyzed in this section.
We use different number of assistant threads to encrypt a plain video of size 576 × 576 using three hardware platforms. The relationship between the average encryption time and the number of assistant threads are drawn in Fig. 8. Set n to the maximum number of threads supported by CPU is unnecessary, however, can guarantee the deployed cryptosystems reach their fastest encryption speed.
The purposes of diffusion and confusion operations are to improve the plaintext sensitivity of the cryptosystems and decrease the correlation of the plain images. We, thus, generate byte sequences, perform 10 rounds of diffusion operations on a plain image, randomly select a pixel, change its value, perform 10 rounds of diffusion operations on the modified image with the same byte sequences, and calculate NPRC and UACI between the generated cipher images. The results are plotted in Fig. 9. Similarly, we perform 10 rounds of confusion operations on a plain image. The plain image and the scrambled images after 1-5 rounds of confusion operations are shown in Fig. 10. The correlation coefficients between the plain image and the scrambled images after different rounds of confusion operations are drawn in Fig. 11. When r is equal to 5, clearly, NPCR, UACI, and the correlation coefficient reach the upper and lower bounds, respectively.
This paper is available on arxiv under CC 4.0 license.
L O A D I N G
. . . comments & more!
. . . comments & more!
