Dew Computing (DC) has recently emerged as a complementary computing paradigm that extends cloud, fog, and edge computing by enabling autonomous, local-first computation directly at end-user devices. Unlike traditional distributed models that rely on centralized or near-edge infrastructures, Dew Computing emphasizes offline-capable, low-latency, and resilient processing at the extreme edge, while maintaining synchronization with fog and cloud layers when connectivity is available. This paper presents a systematic and comprehensive review of Dew Computing based on a structured literature analysis, covering its conceptual foundations, architectural models, and operational mechanisms. The study analyzes key Dew-based architectures, including cloud–dew and hybrid edge frameworks, highlighting their role in reducing latency, improving fault tolerance, enhancing energy efficiency, and supporting privacy-preserving local processing. In contrast to existing surveys, this work provides a critical synthesis of current approaches by identifying their strengths, limitations, and deployment trade-offs across different application scenarios. Furthermore, the paper examines major application domains such as Internet of Things (IoT), smart healthcare, smart agriculture, and cyber-physical systems, where Dew Computing demonstrates advantages in real-time responsiveness and operational resilience. Security and privacy challenges are also analyzed, focusing on recent solutions such as blockchain-based trust management, federated learning, lightweight cryptographic protocols, and AI-driven intrusion detection, while highlighting unresolved issues related to scalability and resource constraints. Unlike prior works, non-computing interpretations such as meteorological dew-point modeling are excluded or clearly distinguished to avoid conceptual ambiguity. Finally, the survey identifies open research challenges, adoption barriers, and future research directions, positioning Dew Computing as a key enabler for decentralized, user-centric, and resilient next-generation computing systems.