These notes outline the evolution and design of Local Area Networks (LANs), the roles of Layer 1/2/3 devices, and the transition to full-duplex, switched Ethernet. They explain transparent bridging, smart forwarding based on a filtering database with backward learning, and the problems caused by loops—together with the Spanning Tree Protocol (STP) as the remedy. The second half motivates Virtual LANs (VLANs) for segmentation on shared physical infrastructure, details IEEE 802.1Q tagging (including header fields and limits), and distinguishes access vs trunk ports and VLAN assignment methods. The deck closes with limits of isolation, interactions with STP, VLAN-aware vs unaware switches, and high-level LAN design conclusions.
Wide Area Networks (WANs) predate LANs: WANs arose in the 1960s to share a few expensive mainframes; LANs appeared in the late 1970s/early 1980s as minicomputers and then PCs enabled local resource sharing while mainframes persisted for other tasks. Initially WANs and LANs evolved independently with vendor-specific protocols (e.g., Decnet, SNA, IP; Novell, Banyan VINES, NetBEUI). Interconnection efforts ultimately converged on IP as the winner.
Modern LANs rely on standards across layers: IEEE 802 at data link (e.g., 802.3 Ethernet, 802.11 Wi-Fi), EIA/TIA 568, and ISO/IEC 11801 for structured cabling. Device roles align with layers: L1 repeaters/hubs separate physical domains but share a collision domain; L2 bridges/switches separate collision domains but share a broadcast domain; L3 routers/L3 switches separate broadcast domains (not specific to LANs).
A repeater interconnects at the physical layer, re-propagating bit streams. It can join networks with the same Medium Access Control (MAC) across different physical layers and counter signal degradation over long cables.
Multiport repeaters (hubs) are required for twisted-pair/fiber hub-and-spoke topologies; they extend the collision domain across all attached links.
