What Is a WAN (Wide Area Network)?

A Wide Area Network (WAN) is what happens when you need networks to talk across distances that physics makes expensive.

Light takes 67 milliseconds to cross the Atlantic Ocean. No amount of engineering will ever change this. When your New York office sends data to London, those bits spend most of their journey simply traveling—not being processed, not being routed, just moving through fiber at two-thirds the speed of light.

Every WAN technology is a strategy for living within this constraint.

The Distance Tax

A LAN exists where distance is free. Your laptop talks to the printer across the room in microseconds. Bandwidth is abundant—10 Gigabits per second is routine. You own the cables. If something breaks, you can walk over and fix it.

A WAN exists where distance costs. Between cities, you don't own the infrastructure—telecommunications companies do. Bandwidth becomes expensive. Latency becomes unavoidable. And the bits must cross territory you don't control, raising questions about security and reliability that never arise when the cable runs through your own ceiling.

This isn't just a matter of scale. Distance changes what's possible. An application that works beautifully on a LAN—snappy, responsive, reliable—might become unusable over a WAN. Not because the WAN is broken, but because the application assumed distance was free.

Buying Your Way Across Distance

Organizations have several ways to span geographic distance, each representing a different bargain:

Leased lines buy you a dedicated path. A T1 line gives you 1.544 Mbps that's yours alone—no sharing, predictable performance, substantial cost. Modern dedicated circuits can reach 10 Gbps or higher. You're paying for guaranteed isolation.

MPLS networks buy you priority on shared infrastructure. The telecommunications carrier labels your traffic and gives it preferential treatment across their network. You get quality-of-service guarantees without paying for a fully dedicated path. Voice and video work reliably. The bill is still significant.

VPN over Internet buys you security on commodity infrastructure. You encrypt your traffic and send it across the public Internet. Cheap, because you're using infrastructure that already exists. Unpredictable, because you're sharing with everyone else streaming video and downloading files.

SD-WAN buys you intelligence. Instead of choosing one path, you use multiple—perhaps MPLS for voice calls, cheap Internet for file transfers, LTE as backup. Software decides moment by moment which traffic takes which path. You're trading complexity for optimization.

Satellite buys you reach at the cost of latency. When terrestrial connections don't exist—offshore platforms, remote research stations, ships at sea—satellites provide connectivity. But signals travel 35,000 kilometers up and 35,000 kilometers back down. That's 500+ milliseconds of latency before the conversation even starts.

5G cellular buys you mobility and quick deployment. Connect a temporary site in hours, not weeks. Provide backup when wired connections fail. Serve mobile workers wherever cell coverage exists. Performance varies with coverage, but improving rapidly.

The Geometry of Connection

How sites connect matters as much as how they connect.

Hub-and-spoke puts one location at the center. Every remote office connects to headquarters. Simple to manage—you secure one central point, and all traffic flows through it. But if the New York office wants to send a file to the Chicago office, the bits might fly to London first—not because anyone designed it that way, but because the hub was placed where executive leadership sits. Geography isn't the only distance that matters.

Full mesh connects every site directly to every other site. No central bottleneck. If New York and Chicago communicate frequently, they have a direct path. But connection count explodes: 10 sites need 45 connections. 100 sites need 4,950. Complexity and cost grow faster than the network.

Partial mesh finds middle ground. Connect the sites that actually talk to each other. Let the rarely-communicating sites route through a hub. This requires understanding your traffic patterns—which means someone has to think about them.

What Distance Does to Performance

WAN performance characteristics shape what's possible:

Latency is the time tax. A packet crossing the continental United States experiences 30-50 milliseconds of delay. Crossing the Pacific, 100-150 milliseconds. These add up. An application that requires 10 round trips to complete an operation will take at least a full second on an intercontinental link—even with infinite bandwidth.

Bandwidth is the capacity tax. Most business WAN links range from a few Mbps to a few hundred Mbps. Your LAN probably runs at 1,000 Mbps or more. That order-of-magnitude difference means the file that transfers in seconds on the LAN takes minutes over the WAN.

Jitter is the consistency tax. Latency that varies makes real-time applications miserable. A voice call tolerates consistent 100ms latency. It cannot tolerate latency that jumps between 50ms and 200ms—the audio breaks up, words overlap, conversation becomes impossible.

Packet loss is the reliability tax. When routers get overwhelmed, they discard packets. LANs rarely lose packets. WANs might lose 0.1% to 1% during congestion. Every lost packet must be retransmitted, compounding the latency tax.

Fighting the Distance Tax

Organizations don't accept these constraints passively:

Compression shrinks data before transmission. Why send the full-resolution image when the display is only 800 pixels wide?

Deduplication notices that the same data crosses the WAN repeatedly. If you sent that file yesterday, today you only send the delta.

Caching keeps frequently-accessed data close to users. The remote office stores local copies of common files rather than fetching them across the WAN every time.

Protocol optimization fixes protocols that assumed distance was free. Some applications exchange dozens of small messages to accomplish simple tasks—fine on a LAN, agonizing over a WAN. Optimization appliances combine those messages or answer them locally.

Quality of Service ensures that the CEO's video call doesn't stutter because someone started a large file download. Traffic gets prioritized by importance, not arrival order.

The Internet Is a WAN

The Internet is the planet's largest WAN—millions of networks agreeing to forward each other's traffic.

But it's a WAN with no guarantees. When you send traffic across the Internet, you're asking dozens of organizations you've never met to carry your data. They'll try. They make no promises about how fast, how reliably, or what priority your traffic gets.

For many purposes, this is fine. Best-effort delivery works for web browsing, email, and most business applications. Encrypt the traffic, and security is adequate.

For other purposes—consistent voice quality, guaranteed bandwidth for critical applications, regulatory compliance—the Internet's lack of guarantees becomes a problem. Private WAN technologies still exist because some guarantees are worth paying for.

Securing the Crossing

WAN traffic crosses territory you don't control. Security becomes essential:

Encryption makes intercepted traffic meaningless. IPsec, TLS, and similar protocols ensure that even if someone captures your bits in transit, they can't read them.

Authentication ensures you're actually talking to your London office, not an imposter. Certificates, keys, and protocols establish identity before sharing sensitive data.

Segmentation contains damage. If guest WiFi gets compromised, it shouldn't provide access to your WAN. Different types of traffic travel different paths with different trust levels.

The Cloud Changes Everything

Traditional WANs connected branch offices to corporate headquarters, where the data center lived.

Today, there often is no central data center. Applications run in AWS. Email lives in Microsoft 365. The CRM is Salesforce. "Headquarters" might just be another office—not the center of the network universe.

This inverts the logic. Why route traffic from a branch office through headquarters to reach AWS, when AWS is closer to the branch than headquarters is? SD-WAN recognizes this and sends cloud-bound traffic directly to the cloud. Direct cloud interconnection services bypass the public Internet entirely, providing dedicated paths into AWS, Azure, and Google Cloud.

The WAN is no longer a hub-and-spoke network centered on corporate headquarters. It's becoming a mesh of connections between everywhere and everywhere—branch offices, cloud providers, SaaS applications, mobile workers, home offices.

The Cost Equation

WAN costs include:

Circuits—monthly charges for bandwidth between locations. Often the largest expense.

Equipment—routers, firewalls, SD-WAN appliances at each site. One-time costs with ongoing maintenance.

Management—people to configure, monitor, and troubleshoot. Tools to provide visibility.

Opportunity cost—the business impact when inadequate WAN capacity means applications run slowly, video calls drop, or remote offices feel like second-class citizens.

SD-WAN and Internet-based approaches often cost less than traditional MPLS. But "cheaper" isn't always "better." The right choice depends on what your applications need, what failures you can tolerate, and what your users will accept.