Securing cloud data using secret key 4 optimization algorithm

• Published in: PloS one Vol. 19; no. 4; p. e0301760
• Main Authors: Frimpong, Twum, Hayfron Acquah, James Benjamin, Missah, Yaw Marfo, Dawson, John Kwao, Ayawli, Ben Beklisi Kwame, Baah, Philemon, Sam, Samuel Akyeramfo
• Format: Journal Article
• Language: English
• Published: Public Library of Science 16.04.2024
• Subjects: Algorithms; Analysis; Cloud computing; Evaluation; Mathematical optimization; Ghana
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0301760

Cloud computing alludes to the on-demand availability of personal computer framework resources, primarily information storage and processing power, without the customer's direct personal involvement. Cloud computing has developed dramatically among many organizations due to its benefits such as cost savings, resource pooling, broad network access, and ease of management; nonetheless, security has been a major concern. Researchers have proposed several cryptographic methods to offer cloud data security; however, their execution times are linear and longer. A Security Key 4 Optimization Algorithm (SK4OA) with a non-linear run time is proposed in this paper. The secret key of SK4OA determines the run time rather than the size of the data as such is able to transmit large volumes of data with minimal bandwidth and able to resist security attacks like brute force since its execution timings are unpredictable. A data set from Kaggle was used to determine the algorithm's mean and standard deviation after thirty (30) times of execution. Data sizes of 3KB, 5KB, 8KB, 12KB, and 16 KB were used in this study. There was an empirical analysis done against RC4, Salsa20, and Chacha20 based on encryption time, decryption time, throughput and memory utilization. The analysis showed that SK4OA generated lowest mean non-linear run time of 5.545±2.785 when 16KB of data was executed. Additionally, SK4OA's standard deviation was greater, indicating that the observed data varied far from the mean. However, RC4, Salsa20, and Chacha20 showed smaller standard deviations making them more clustered around the mean resulting in predictable run times.