Blockchain-based anonymous data exchange with machine learning for cybersecurity in industrial control systems for the Internet of Things.

Abstract

Security is an important and critical issue of industrial systems with infrastructures on IoT that benefit users only if they maintain security. An Industrial Control System is an essential infrastructure in business environments and in the development of ICT it is difficult to prevent cyber attacks with traditional methods. This paper aims to design an architecture for Blockchain-based anonymous data exchange systems with Machine Learning techniques. The methodology uses the bibliographic review, the empirical-analytical research, the graphing of the logical infrastructure, the qualitative approach. As results are cited: a framework divided into 4 levels called IoT, network, Blockchain and Machine Learning was designed, to detect anomalies with a prediction algorithm and visual analysis to inspect the existence of attacks against the system and take appropriate measures; the literature review resulted in 61 articles and provides information on architectures in the three technologies; It is emphasized that only 13% were implemented, 66% are software simulations and 21% are theoretical; simulation evaluations provide a maximum throughput of 99.70% and a minimum throughput of 93.86%; the three simulations of ten scenarios each maintain acceptable yields, and it is concluded that the performance is very acceptable at a general level. It was concluded that the framework resulting from the combination of these technologies is a viable solution to improve security in industrial systems and approve distributed access to data.