A blockchain-based machine learning intrusion detection system for internet of things

Abstract

The Internet of Things (IoT) is the major evolution of Internet also known as Internet of Everything which made a network with smart sensors heterogeneous devices. Nowadays, the usability of IoT networks is increasing very rapidly from smart home, smart industry to smart everything. But, these smart devices like as traditional Internet are vulnerable to various attacks such as denial of service (DoS), spoofing attacks, ransomware attacks, and many more. There are also various protocols such as DTLS, IPv6, and many other lightweight protocols used for IoT data security. But despite these, these attacks are also occurred via sniffing or manipulating of header information to both encrypted and non-encrypted protocols. Attacks generated via header information can be mitigated by various methods as ML-based intrusion detection systems (IDSs) is one of them. These IDSs security depends on the accuracy/integrity of training data (IoT data) and trust on the ML/DL algorithms. Recently, blockchain, a new advanced technology, is emerged, which has several use cases in the IoT domain for providing security. Due to the various advantages of blockchain and ML/DL methods in IoT data security, we combine these technologies and provide a secure blockchain-ML-based framework for heterogeneous IoT data security environment.