Wide Area Network
A wide area network (WAN) is a computer network that spans a large geographic distance, such as multiple cities, countries or continents, linking smaller networks, typically local area networks, into one connected system. Because no single organization can run cables across public land and oceans on its own, a WAN is usually built from circuits and services leased from telecommunications carriers, and its defining difficulty is not the equipment but the distance.
The WAN's central constraint is latency governed by physics: signals in optical fiber travel at roughly two thirds of the speed of light in a vacuum, about 200,000 kilometers per second, so distance sets a floor under delay that no amount of bandwidth can remove. A round trip between New York and London over fiber cannot take less than several tens of milliseconds, and across the Pacific it takes longer, whereas a LAN round trip is a fraction of a millisecond. For that reason WAN design emphasizes caching content near users, replicating data between regions, reducing the number of back-and-forth exchanges, and protocols that tolerate delay. Applications that feel instant in an office can seem slow across a WAN purely because they were written to make many small round trips.
Historically the WAN came first: the ARPANET of 1969 linked computers at distant universities over leased telephone lines, and it established the practices that WANs still follow. Later technologies include X.25 packet networks, frame relay and ATM, which carriers sold as private circuits to businesses, followed by MPLS, a label-based forwarding technique widely used from the late 1990s to give enterprises predictable private paths between sites. Today a large share of intercontinental traffic runs through submarine fiber-optic cables, with repeaters or amplifiers along the route, and carriers and large cloud companies own or lease capacity on them. The link between a customer site and its carrier may be a leased line, a fiber connection or ordinary broadband, depending on cost and reliability needs.
Organizations choose among private and public WAN options by trading cost, control and security. A leased line or MPLS service offers predictable bandwidth and delay under a contract but is expensive and slow to provision; a virtual private network (VPN) builds an encrypted tunnel over the ordinary internet at far lower cost, though performance varies with the public path. Software-defined WAN (SD-WAN), widely adopted since the 2010s, adds a software layer that watches multiple links and steers each application's traffic over the best one. Whatever the transport, the sites are joined by routers, and a WAN requires addressing, routing protocols and security policies to be agreed across organizational boundaries.
The internet is often described as the largest WAN, but the two terms are not synonyms, and mixing them up is the commonest misconception. A WAN is any wide-area network, public or private; the internet is one specific global interconnection of thousands of independently operated networks, and a company can run a private WAN that never touches it. Likewise a wide area network does not mean a fast one, since bandwidth and latency are separate qualities, and a high-capacity link across an ocean still has the delay of the distance. WANs matter more than ever for cloud computing, content delivery networks, remote work, and distributed machine learning, where moving datasets and model checkpoints between regions is limited by wide-area bandwidth, and where moving the computation closer to the data is often cheaper than moving the data.
Read a wide area network as the long-haul stretch that stitches local networks into one system: built from carrier circuits, tunnels and undersea fiber, limited more by distance and delay than by raw capacity, and designed around the fact that the speed of light cannot be negotiated.