MOON: An approach to data management on relational database and blockchain

Abstract

The Relational model has been used widely for decades and was valuable for the popularization of the use of Databases. Nowadays, many applications of several domains continue to be data-oriented, but alternatives to the Relational model have been solidifying. One of them is Blockchain, which is considered a disruptive technology and has relevant properties, such as immutability and no need for centralized third parties. Therefore, applications that use Relational Databases (BDR) can benefit from these properties by migrating part of their data to Blockchains. This research presents the approach to data Management on relatiOnal database and blOckchaiN (MOON), which its client applications use SQL DML to communicate with the MOON. Then, clients send inserts, updates, selects, and deletes queries to the MOON, which execute them regardless of whether the data is in a Blockchain or RDB. Furthermore, the MOON performs mapping between relational and Blockchain model, integrating the two technologies, and Blockchain’s data indexing. There were two experiments to validate this study. The first generated workload from real data from a Portuguese hospital and the second used two validated benchmarks: Voter and Twitter. In all experiments, there was an execution on three scenarios: i) data stored only in the BDR; ii) partitioned data using MOON; and iii) data stored only on the Blockchain. The metrics used in the evaluations were response time and correctness. The conclusion is that the MOON responds to requests correctly and provides RDB and Blockchain features to the data, such as supporting complex queries and data immutability. Moreover, their response time was intermediate between BDR and Blockchains.