Trusted relational databases with blockchain: design and optimization

Abstract

With the emergence of large scale data collection from Internet of Things and mobile devices, the notion of trust is now an increasingly important aspect of the next generation of data processing systems. We propose a blockchain enabled relational storage system that supports immutable transactions and temporal snapshots. By embedding blockchains in relational tables, the database stores trust related information in a tamper proof fashion, making the data provenance provably verifiable. To support large query workloads, we further propose an optimization algorithm that determines the best temporal snapshots to materialize in order to minimize the total time cost of answering a given query workload. Experimental evaluation shows that our materialized snapshot approach improves the performance of large analytical query workloads by as much as 50 times.