Secure and scalable IoT device management with blockchain, K-anonymity, and deep reinforcement learning

Abstract

The vast expansion of Internet of Things (IoT) ecosystems has posed significant challenges regarding security, scalability, latency, and user privacy, which are not properly handled by the conventional architectures. To solve these issues, the authors of this paper suggest an integrated IoT management framework that employs blockchain technology, K-anonymity, and deep reinforcement learning (DRL). Blockchain guarantees security and trust in a decentralized manner, Delegated Proof of Stake (DPoS) enhances scalability and energy efficiency, and IPFS offers decentralized data storage that is efficient. K-anonymity is used to keep user and device privacy secure by stopping the identification of sensitive data. Moreover, DRL is applied to dynamically optimize resource allocation, routing, and latency in changing IoT settings. The experimental results show that the proposed framework can process 15,000 transactions every second with an average latency of 100 ms, which is better than the CPC model. Such results reveal that the method proposed is a versatile, secure, and privacy-respecting answer for smart city, industrial automation, and healthcare IoT applications.