Security and privacy concerns, which are made worse by the growing
number of Internet-connected devices, are the primary obstacles to the Internet of
Things (IoT) widespread implementation. Everyone is now very concerned about
Internet of Things security, including businesses, governments, and consumers. Even
though no system can ever be completely protected from attacks, effective system
defense depends on real-time threat detection. There is a dearth of studies on intrusion
detection systems that work well in IoT-based blockchain systems. In this study, we
present a novel approach to intrusion detection in the Internet of Things-based
blockchain systems by employing the DL (Deep Learning) subfield of machine
learning to identify security abnormalities. This detection platform provides security as
a service and allows interoperability with numerous network communication protocols
utilized by the Internet of Things. We go into great details about the suggested system's
architecture and intrusion detection method. Real network traces are evaluated along
with simulated data to illustrate the scalability of the proposed intrusion detection
system to prove its viability. Our results confirm that the proposed intrusion detection
system can correctly detect real intrusions.
Keywords: Smart contracts, Blockchain, Deep learning, IoT, Intrusion detection.
