<p>Edge IoT Systems (EIoT) play a pivotal role in the advancement of smart technologies by seamlessly integrating computational models with physical systems. Edge computing serves as a key enabler for smart IoT systems, addressing challenges commonly associated with cloud-based or server-based approaches, including latency, energy consumption, bandwidth limitations, and accuracy. Research on Edge-AI in IoT is scattered, as most studies focus on specific cases or designs. A detailed review is needed to bring these works together, compare their performance, and highlight challenges and future opportunities. To address this gap, this comprehensive survey focuses on the key performance parameters of Edge-assisted Edge IoT Systems (EaEIoT). It analyzes metrics such as latency, energy efficiency, accuracy, scalability, security, and reliability, emphasizing their role in enhancing operational efficiency. The study reviews traditional methods and models in EIoT, tracing their evolution and the improvements enabled by edge computing. It also evaluates the integration of Artificial Intelligence (AI) within EIoT and its impact on overall system performance. Furthermore, the review explores current research trends, challenges, and future directions, providing a holistic perspective on optimization strategies to improve the efficiency and reliability of EaEIoT.</p>