<p>The rapid development of Industry 4.0 has accelerated the deployment of Industrial Internet of Things (IIoT) technologies, allowing for extensive communication between industrial infrastructure, sensors, actuators, and cyber-physical systems. However, major architectural challenges arise in terms of scalability, interoperability, latency, and security because of the enormous volume of heterogeneous data produced and the stringent real-time requirements. Fog computing, which connects edge devices and centralized cloud systems, has become a crucial intermediary layer in this environment. With an emphasis on fog computing-based industrial systems, this paper offers an organized and critical literature analysis of IIoT designs. This review is based on a systematic selection of peer-reviewed papers drawn from major scientific databases. It proposes a taxonomy of IIoT designs, categorized into cloud-centric, edge-optimized, fog computing-based, and hybrid models. To assess their performance in terms of energy efficiency, scalability, interoperability, resilience to cyber attacks, and latency management, a comparative study is carried out. The investigation identifies a number of outstanding issues, such as multi-layered security flaws, interaction with current industrial infrastructures, distributed fog computing orchestration, and large-scale deterministic real-time communication. This study offers a systematic framework to direct future improvements of scalable, secure, and energy-efficient IIoT ecosystems by synthesizing previous research and addressing specific research gaps.</p>