Securing Homogeneous Big Data Using Augmented and Light-Weighted Homomorphic Encryption Scheme

Authors

  • D. Anuradha, S. Bhuvaneswari

Abstract

As the business and scientific data size grows faster in tremendous rate, big data would be the choice for the storage and management of the data. Big data analytics improves the corporate revenues and customer satisfaction. Also, big data analytics helps governments to plan people welfare schemes effectively. But sharing sensitive data among data analysts may lead to data privacy leakage in some cases. Homomorphic class of encryption schemes allow us to do mathematical calculations on encrypted data and the results can be used for analysis purposes meaningfully. But the existing homomorphic algorithms are relatively expensive in terms of complexity and runtime if used for big data. A new augmented homomorphic encryption algorithm, which is light-weighted when compared with other traditional algorithms is proposed in this work. Even though this design is light-weighted, it is robust against cryptanalysis, as this new algorithm produce different cipher for same message at different occurrences in a data block. A new method to transform different ciphers of the same message into a cipher, which is very useful for map-reduce operation.is also proposed here.Java programs are tested in HDFC-Hive environment and the results obtained shows that my algorithms are faster than other existing homomorphic algorithms.

Downloads

Published

2020-05-12

Issue

Section

Articles