Comparative Performance Analysis of Information Dispersal Methods

• Main Authors: Maxim Deryabin, Nikolai Chervyakov, Andrei Tchernykh, Viktor Berezhnoy, Anvar Djurabaev, Anton Nazarov, Mikhail Babenko
• Format: Article
• Language: English
• Published: FRUCT 2019-04-01
• Series: Proceedings of the XXth Conference of Open Innovations Association FRUCT
• Subjects: Information Dispersal; Distributed Systems; Redundant Residue Number System; Performance; Erasure Codes; Reliable Storage
• Online Access: https://fruct.org/publications/fruct24/files/Der.pdf

In this paper, we present an analysis of information dispersal methods for using in distributed storage systems, processing, and transmission of data. We provide a comparative study of the methods most widely used in practice considering performance, reliability and cryptographic security. There are three main approaches to the information dispersal: Information Dispersal Algorithm by Rabin, Residue Number System (RNS) and Polynomial Residue Number System. We propose an efficient data recovery algorithm based on data representation in the RNS. Comprehensive experimental analysis shows that the most productive approach for bit length up to 256 bits is the use of the RNS with our developed algorithm. We show that the use of the RNS for the design of distributed storage systems, data transmission, and data processing, can significantly reduce the time of information processing.