This chapter provides foundational concepts for Graph Neural Networks, covering static and dynamic graph settings. For static graphs, we define core components such as graph structure, message passing mechanisms, and training processes. We also discuss common graph analytics tasks, datasets, and sparse storage formats. For dynamic graphs, we introduce continuous-time dynamic graphs, temporal GNNs, and their training pipelines, emphasizing chronological constraints and hybrid CPU-GPU data layouts. Finally, we review hardware-aware training scenarios and system optimizations to reduce data transfer and improve bandwidth utilization. This section serves as a comprehensive primer for scalable GNN training across diverse settings.