A binary matrix-based data representation for data compression in blockchain

Laouid, Abdelkader, Mostefa, Kara, Al-Khalidi, Mohammed ORCID: https://orcid.org/0000-0002-1655-8514, Chait, Khaled, Hammoudeh, Mohammad and Aziz, Ahmed (2023) A binary matrix-based data representation for data compression in blockchain. In: 2023 Fifth International Conference on Blockchain Computing and Applications (BCCA), 24 October 2023 - 26 October 2023, Kuwait City, Kuwait.

Preview

Accepted Version Available under License In Copyright. Download (339kB) | Preview

Abstract

Blockchain relies on storing and verifying a large volume of data across multiple nodes, making efficient data compression techniques crucial. By reducing the size of data, compression techniques enable more data to be stored within the limited space constraints of the blockchain networks. Furthermore, compressed data consumes less bandwidth for transmission and enhances the overall performance of blockchain networks by reducing the time and resources needed for data storage and retrieval. To overcome this issue, this paper presents a new data representation approach to enable efficient storage and management of diverse data types on the blockchain, ensuring scalability, cost-effectiveness, and improved network efficiency. A binary matrix M of size m x n bits can be converted to two vectors H and V of sizes m’ and n’, respectively. The compression rate expressed by (m‘ + n’ + │ Hash(M) │) x 100/(m × n) increases exponentially, i.e., 2 λ with λ depends on m and n); this makes the proposed technique is very effective in data size reduction. With a matrix, for example, M = 512 x 512 bits, we achieve a rate of reduction equal to 96.42%. The original data can be recovered using H, V, and Hash(M). The conversion from M to (H, V) is simple, which optimizes energy consumption for low-power devices. Meanwhile, the challenge of recovering the original data could be exploited in a blockchain process, where the mining consensus could be identified based on the node that recovered a predefined set of vectors. Furthermore, this technique ensures that data integrity checking is available only at the nodes with a massive computation capacity.