Network Types
Part of Networking
Classification of networks by scale, ownership, and purpose.
Why This Matters
Networks are classified by scale, ownership, and purpose because different use cases have different technical and economic requirements. A network connecting three computers in one office needs very different technology than a network connecting thousands of locations across a country. Understanding network types helps you match the right technology to the right problem and understand why certain design choices exist.
Classification also provides vocabulary for describing network architecture clearly. When someone says “our WAN connects multiple LANs,” both parties understand the scale and structure being described without spelling out every detail. This shared vocabulary enables efficient communication among people building and maintaining networks.
In a reconstruction context, understanding network types helps prioritize which networks to build first, what technology each requires, and how they might interconnect as the communications infrastructure grows.
Local Area Network (LAN)
A Local Area Network covers a limited geographic area — typically a single building or campus — and is owned and operated by a single organization. LANs are characterized by high bandwidth (typically 100 Mbps to 10 Gbps), low latency (submillisecond), and relatively low cost per connection.
Modern LANs use Ethernet (IEEE 802.3) for wired connections and 802.11 (Wi-Fi) for wireless connections. A typical LAN consists of network interface cards in computers, patch cables connecting computers to wall plates, structured cabling (Cat5e/Cat6) running from wall plates to wiring closets, patch panels in wiring closets, and Ethernet switches connecting everything together.
The economics of LANs favor centralized infrastructure. The cost of running cable through walls and ceilings is fixed regardless of future bandwidth upgrades — upgrading from 100 Mbps to 1 Gbps requires only replacing the switches and NICs, not the cables. This amortization of physical infrastructure cost makes planned, structured cabling a sound long-term investment.
A Home Area Network (HAN) is a LAN for residential use. The same technology applies; the scale and management complexity are smaller.
Metropolitan Area Network (MAN)
A Metropolitan Area Network covers a city or metropolitan region, spanning distances too large for standard LAN technology (which is limited to 100 meters for twisted pair) but not large enough to require WAN economics. MANs are typically operated by telecommunications carriers, ISPs, or large organizations with facilities spread across a city.
MAN technologies include fiber optic runs between buildings (using dark fiber, passive wavelength division multiplexing, or active fiber optic equipment), wireless point-to-point links for locations where fiber is impractical, and carrier Ethernet services (where a telecom provider offers Ethernet connectivity between sites).
For a reconstruction context, the relevant MAN technology is point-to-point wireless links. Line-of-sight wireless links at 2.4 GHz, 5 GHz, or licensed microwave frequencies can span 10-50 km with modest equipment. A pair of directional antennas on towers or buildings, connected to wireless bridge hardware, creates a high-bandwidth link without laying cable. This is the fastest way to extend connectivity across distances that exceed Ethernet cable reach.
Wide Area Network (WAN)
A Wide Area Network spans large geographic distances, potentially continental or global. The internet is the ultimate WAN — a network of networks connecting billions of devices worldwide. WANs are characterized by lower bandwidth, higher latency, and much higher cost per bit than LANs.
WAN links use: leased lines (dedicated circuits between two points, provided by telecommunications carriers), DSL and cable modem (shared broadband connections over existing copper and coaxial infrastructure), fiber optic internet connectivity, and satellite links (available almost anywhere on Earth, but with high latency due to the distance to geostationary satellites).
WAN economics fundamentally differ from LAN economics. LAN bandwidth is cheap and shared within an organization. WAN bandwidth is expensive and typically metered or capacity-limited by the connection speed. This difference drives compression, caching, and traffic prioritization at WAN boundaries — techniques that are unnecessary inside a LAN.
Personal Area Network (PAN)
A Personal Area Network connects devices within a very short range — typically within a few meters of a person. Bluetooth and USB are the primary PAN technologies. Bluetooth connects phones to headsets, computers to keyboards and mice, and sensors to smartphones. NFC (Near Field Communication) enables contact-less payments and device pairing.
PANs are relevant for integration of wearable and portable devices with local network infrastructure. A health monitoring device that communicates via Bluetooth to a smartphone, which relays to a LAN, is a PAN bridging to a LAN.
Virtual Private Network (VPN)
A VPN extends a private network across a public network (typically the internet). Remote users connect to a VPN server, and their traffic is encrypted and tunneled through the internet to emerge on the private network. To the user, remote access via VPN feels the same as being on the local LAN.
VPNs are essential for connecting remote sites and remote users to organizational networks. Instead of paying for dedicated WAN links between sites, a site-to-site VPN uses existing internet connectivity with encrypted tunnels to provide secure connectivity. This is dramatically cheaper than leased lines for lower-bandwidth requirements.
OpenVPN and WireGuard are widely used open-source VPN implementations. WireGuard is newer, simpler, and faster; OpenVPN is more mature and works in more environments (including strict firewalls that block WireGuard’s UDP-based protocol). Either is appropriate for small-to-medium deployments.
Storage Area Network (SAN)
A Storage Area Network is a dedicated high-speed network connecting servers to storage systems. Instead of storage being directly attached to each server, storage is on a shared network accessible to any authorized server. This enables flexible storage allocation, high availability through redundant paths, and centralized management.
SANs use Fibre Channel (a specialized high-speed protocol designed for storage) or iSCSI (SCSI commands over TCP/IP, which allows use of standard Ethernet equipment). Fibre Channel SANs require specialized switches, host bus adapters, and cables, making them expensive. iSCSI SANs use standard Ethernet infrastructure at the cost of some performance overhead.
For a reconstruction context, SANs are probably overkill. The complexity and cost are justified only when many servers need shared access to the same storage, which requires significant organizational infrastructure to justify. NFS (Network File System) or SMB over a standard LAN provides shared storage for most small-to-medium needs without the complexity of a dedicated SAN.
Network Tiers and the Internet
The internet is organized in tiers. Tier 1 networks are the backbone — large telecommunications companies with global networks that interconnect with each other for free (settlement-free peering). Tier 2 networks have regional scope and pay Tier 1 networks for transit to reach destinations outside their own networks. Tier 3 networks (ISPs) serve end users and pay Tier 2 or Tier 1 networks for transit.
This hierarchy means that a packet sent from one continent to another traverses multiple networks: the sender’s LAN, the sender’s ISP (Tier 3), a regional network (Tier 2), a global backbone (Tier 1), then back down through Tier 2 and Tier 3 to the destination’s LAN. Each boundary between networks involves routing policy decisions about preferred paths and backup paths.
For reconstruction purposes, the ability to connect any two points in the world via the internet is a powerful capability. Understanding that the internet is a network of networks, each operated independently but connected through standard protocols, explains why the internet is resilient — no single point controls all traffic, and failures in one part route around by the distributed routing protocols.