Distributed Storage with Strong Data Integrity based on Blockchain Mechanisms

Abstract

A blockchain is a datastructure that is an append-only chain of blocks. Each block contains a set of transaction and has a cryptographic link back to its predecessor. The cryptographic link serves to protect the integrity of the blockchain. A key property of blockchain systems is that it allows mu- tually distrusting entities to reach consensus over a unique order in which transactions are appended. The most common usage of blockchains is in cryptocurrencies such as Bitcoin. In this thesis we use blockchain technology to design a scalable architec- ture for a storage system that can provide strong data integrity and ensure the permanent availability of the data. We study recent literature in blockchain and cryptography to identify the desired characteristics of such a system. In comparison to similar systems, we are able to gain increased performance by designing ours around a permissioned blockchain, allowing only a predefined set of nodes to write to the ledger. A prototype of the system is built on top of existing open-source software. An experimental evaluation using different quorum sizes of the prototype is also presented.