Adaptive Encryption Algorithms for Big Data: Balancing Security and Performance

Abstract

The availability of huge amounts of data in big data science and analyzing them requires computational efficiency as well as possessing privacy and security at the same time is apparently a challenge. Convention encryption techniques are also problematical to achieve the volume, velocity and variety of the big data. Hence, this paper aims to examine the use of adaptive encryption algorithms pointing out the possibility of achieving an optimal ratio of security and performance in big data. There are types of encryption techniques that allow for the control of the level of protection that is being offered with reference to the amount of security needed for the particular data and the resources available for the process. First, we discuss various adaptive encryption methods as part of prior work, including variable key size, adaptive encryption types, and context sensitivity. We also talk about how such algorithms can be incorporated with big data processing platforms such as Hadoop and Spark. In addition, we also assess the effectiveness of the developed adaptive encryption algorithms with different big data benchmarks. From our early finding, it can be seen that adaptive encryption challenged and outperforms static encryption techniques in terms of processing time with no neglect to optimum security. Also, we present the advantages, disadvantages, issues involved and threats concerning the adaptive encryption such as the keys management issue, schema evolution and possible insecurity. At last, we propose a few avenues for further research in the area, such as proposing new types of adaptive encryption schemes based on machine learning and data classification approaches to shed more light on the optimization of the security-performance paradigm.