Abstract The Industrial Internet of Things is transforming industrial processes through the integration of edge, fog, and cloud computing to offer predictive maintenance to reduce unexpected downtime and increase operational effectiveness. Limited use of predictive maintenance is due to antiquated legacy systems, the availability and security of dependable communication networks, and the provision of big datasets in real time. This study proposes a novel multilayer framework that uses MQTT protocol for transferring data across the layers along with systematically integrating pre-processing and machine learning algorithms to ensure robust and accurate predictions. It uses real-time data collection, low-latency processing, and advanced analytics to predict and prevent equipment failures. Bibliometric analysis of 1,281 publications between 2015 and 2024 shows the rapid growth in IIoT-based predictive maintenance research, where India and China are emerging leaders in contributions and citations. This paper discusses the key enabling technologies, including IoT sensors, edge computing, fog computing, cloud computing, machine learning, and blockchain, and also discusses the challenges such as energy efficiency and system scalability along with the solutions to overcome those challenges. The applications of predictive maintenance in the manufacturing, energy, and automotive industries point to extensive opportunities for reducing downtime, enhancing the life of equipment, and reducing operational costs. The bibliometric insights, coupled with the proposed framework highlight the transformative role of predictive maintenance in modernizing industrial ecosystems. This study paves the way for scalable, energy-efficient, and sustainable predictive maintenance systems to advance the next generation of industrial processes.
Kapoor et al. (2025) studied this question.