Blockchain-AI Integrated Framework for Efficient and Secure Big Data Processing in IIoT Using Enhanced DPoS Consensus

Abstract

The blistering development of the Industrial Internet of Things (IIoT) has brought serious issues to the maintenance of large-scale sensor data security and processing with low latency and scalability. Conventional central and edge-only solutions are either limited in the number of trust bottlenecks or restricted in the detection accuracy, thus a hybrid solution is required. This study establishes a Blockchain-AI composite model, where federated anomaly detection and a superior Delegated Proof-of-Stake (eDPoS) consensus mechanism system are used to efficiently and safely process big data on IIoT scenarios. This methodology gives the analytical models that are vital in throughput, latency and the likelihood of hostile takeover. Researcher experimented the Indian IIoT and Blockchain Synthetic Dataset which includes DPoS information under a wide range of conditions, including safe and malicious adversarial stake attacks. It was found to significantly (up to 20 percent) improve throughput over vanilla DPoS, but latency is minimized under medium-delay networks and can anomaly detect (AUC [?] 0.93) with errors nearly equal to centralized (under 5 percent) baselines. Security analysis provides resistance to stake-boost attacks and optimization of storage using lightweight anchoring. This paper makes the framework a scalable and secure IIoT deployment solution, between blockchain consensus and AI-driven anomaly detection.