IoT-Enabled Smart Cities Architectures Connectivity Edge Computing Digital Twins and Security

The Internet of Things (IoT) supports urban monitoring and service management by connecting sensors, actuators, communication networks, and computing platforms. This survey examines IoT-enabled smart cities across five themes: layered reference architectures; connectivity, including low-power wide-area networks (LPWAN), fifth-generation (5G) cellular networks, satellite links, and prospective sixth-generation (6G) technologies; edge and fog computing; digital twins; and security and privacy, with emphasis on blockchain based and federated approaches. Application domains in energy, transportation, waste, water, environmental monitoring, and street lighting are reviewed. The central contribution is a conceptual cross-layer decision framework that links service requirements to connectivity selection, computation placement, and digital twin configuration under security and governance constraints. A first-order model of offloading delay and device energy illustrates the computation placement trade-off. The reviewed studies show that no single connectivity technology satisfies all urban requirements; that a field pilot of 5G exhibited significant variation in network performance; that, in simulation, a low Earth orbit (LEO) narrowband IoT constellation of 60 satellites guaranteed a minimum link availability of about 30% of the simulated period; that city-scale digital twins have been applied most often to urban planning; that a blockchain ledger does not guarantee the truthfulness of recorded sensor data; and that learning-based defenses can be evaded. An evidence classification shows that the quantitative studies examined mostly report results from simulations, benchmark datasets, or single pilots, and not from operational city deployments. Persistent challenges include interoperability, energy efficiency, scalability, connectivity reliability, and data governance. Promising directions include hybrid terrestrial and satellite connectivity, adaptive edge and cloud processing, sensor-driven digital twins on standardized data models, privacy-preserving collaborative learning, and selectively applied distributed ledgers. The survey is non-exhaustive and follows no formal systematic protocol.