Architecting blockchain network simulators: a model-driven perspective

Abstract

Blockchain networks have been suggested to have the potential to support some of the most critical functions of modern societies. When used in such capacities, failures of blockchain networks imply catastrophes that extend beyond individuals, organizations and countries. As such, before considered for wide adoption, blockchain network protocols and technologies must undergo the highest standards of analytical and empirical validation subject to key security, reliability and performance qualities. When performing empirical evaluation, however, the sheer size of open-access blockchain networks in their envisioned scale rules out the possibility of exact reproduction and validation in a lab environment. Rather, abstract working models – simulators – of proposed technologies need to be considered. To have value as research instruments, such simulators need to be widely validated for their accuracy by the research community, and also be highly transparent and reusable for allowing quick implementation and comparison of design ideas. We claim that established software engineering paradigms, namely model-driven development and software product lines can help address this need. We outline our own effort to develop a domain meta-model and object-oriented framework for efficient and reliable derivation of specialized blockchain network simulators.