Types of Networks: LAN, WAN, MAN, PAN

Right now, within arm's reach of you, there is very likely a network so small and so intimate that you have never once thought of it as a network at all. If you are listening to anything through wireless earbuds, a burst of radio energy is hopping from your phone's antenna to a chip smaller than a grain of rice, dozens of times a second, keeping the sound in your left ear in lockstep with the sound in your right. Zoom out a little and there is a second network: the Wi-Fi humming quietly through your walls, herding your laptop's traffic, your smart speaker's traffic, and your roommate's video call toward the same plastic box blinking in a corner of the room. Zoom out further still and there is a third: somewhere under an ocean, at a depth that would crush a submarine, a cable barely as thick as a garden hose is carrying a signal that started as somebody's video call on one continent and will land on a screen in another one before this sentence finishes being read out loud.

None of those three networks look alike. None of them are built the same way, none of them run the same protocols at their edges, and none of them would work at all if you swapped their jobs — you cannot run a transatlantic cable's protocol between two earbuds, and you certainly cannot run Bluetooth's protocol across an ocean. That contrast is the entire subject of this chapter. Networking is not one thing that happens at one scale; it is the same handful of ideas — nodes, links, protocols, the three ingredients from the very first chapter — recruited again and again at wildly different distances. And every time the distance changes by an order of magnitude, the rules that work best change right along with it.

Engineers gave these scales names decades ago, and the names have stuck well enough that you will hear them in datasheets, job interviews, and casual hallway conversation among network engineers for the rest of your career: PAN, LAN, MAN, WAN. Think of them as concentric circles drawn around you — the tightest one barely larger than your own body, the widest one wrapping the entire planet — and know that the size of the circle is never just a matter of distance. It is a decision about cost, speed, trust, and how much you are willing to lose if a single link fails.

Scale changes everything in networking: the same goal — getting data from A to B — demands entirely different tools depending on whether A and B are ten centimeters or ten thousand kilometers apart.

PAN: Personal Area Network

Start with the smallest circle, the one drawn tightest around you. A Personal Area Network connects the handful of devices that belong to a single person and rarely wander more than a few meters from their body: a phone, a pair of earbuds, a smartwatch, maybe a fitness ring or a pair of wireless headphones. It is the network equivalent of your own pockets and wrists — nothing in it is meant to be shared with a stranger, and almost nothing in it needs to survive a wall, let alone a city block.

The interesting story here is not the range; it is the tradeoff. A PAN is built entirely around one brutal constraint: battery size. Your earbuds are powered by a battery not much bigger than a lentil, and it has to last all day on a single charge, then recharge overnight in a case not much bigger than your thumb. Any radio technology that tried to reach across a house, the way Wi-Fi does, would drain that battery in minutes. So PAN technologies deliberately give up range and raw speed in exchange for sipping power instead of gulping it.

Before Bluetooth, the answer to "how do two nearby gadgets talk without a cable" was infrared — the same invisible light your television remote still uses. It was cheap and required almost no engineering cleverness, but it demanded a direct line of sight, which made it perfect for "point your laptop at the printer and press a button" and useless for anything you wanted to happen in your pocket. Bluetooth, standardized in the late 1990s, solved that by switching from light to radio, so devices no longer needed to physically face each other to talk.

A PAN, in other words, is not a lesser network because it is small. It is a network engineered, on purpose, to be exactly as small and as power-frugal as the problem demands — nothing more.

flowchart TB
    master["Phone / Laptop<br/>(Bluetooth Master)"]
    left["Left Earbud<br/>(BLE Slave)"]
    right["Right Earbud<br/>(BLE Slave)"]
    watch["Smartwatch<br/>(BLE Slave)"]
    tracker["Fitness Tracker<br/>(BLE Slave)"]
    master --- left
    master --- right
    master --- watch
    master --- tracker

In a Bluetooth piconet the master coordinates all communication — slaves never talk directly to each other. The star topology and deliberately short radio range keep the power draw low enough for coin-sized batteries.

LAN: Local Area Network

Widen the circle past your own body and you reach the network most people actually picture when they hear the word "network" at all: the Local Area Network, confined to a single building, floor, or campus, and almost always owned by one person or one organization. Your home Wi-Fi is a LAN. So is the row of desks in a small office all plugged into the same switch, and so is a school computer lab where thirty machines share one printer down the hall.

A LAN is the network equivalent of a household — everyone under the same roof, sharing the fridge, the Wi-Fi password, and, if you go back far enough, sharing an actual physical wire. The earliest local networks in the 1970s and 1980s ran over thick, stiff coaxial cable, and connecting a new machine to that cable meant literally piercing it with something engineers nicknamed a "vampire tap" — a clamp with a spike that bit through the cable's insulation to make electrical contact with the copper core inside, without cutting the cable in two. It was exactly as unglamorous and slightly alarming as it sounds, and it is a large part of why, once cheap switches and twisted-pair cabling arrived in the 1990s, nobody looked back.

What makes a LAN feel so different from everything bigger than it is trust and speed happening almost for free. Devices on the same LAN are usually a few dozen meters apart at most, so signals arrive in a fraction of a millisecond, and the whole thing is typically administered by one person or one small team who can see, physically, every cable and every device on it. That closeness is exactly why a LAN is where you first plug in a printer, share a file, or stream a movie to the television without a second thought about who else might be listening in.

flowchart TB
    router["Router / Switch<br/>Gateway 192.168.1.1<br/>(DHCP, NAT, Firewall)"]
    laptop["Laptop<br/>192.168.1.10"]
    printer["Network Printer<br/>192.168.1.20"]
    tv["Smart TV<br/>192.168.1.30"]
    ap["Wi-Fi Access Point"]
    router --- laptop
    router --- printer
    router --- tv
    router --- ap
    ap -.- phone["Phone<br/>192.168.1.40"]
    ap -.- tablet["Tablet<br/>192.168.1.50"]

Every device shares the same private subnet and the same gateway — the router — which is the only device that sees the public internet. Communication between any two devices on the LAN stays entirely inside the building and adds sub-millisecond latency.

MAN: Metropolitan Area Network

Keep widening the circle and you leave the building behind entirely, moving up to something that spans a whole city or a large multi-site campus: the Metropolitan Area Network. Picture a university with a library downtown and a dormitory across town, both wired into the same high-speed fiber backbone, so a student's laptop reaches the exact same file server whether they are sitting in a lecture hall or lying in bed. Picture a hospital system linking several clinics scattered across a city to one shared patient-records system, so a doctor at any location sees the same chart. Picture a city government running its own fiber ring beneath the streets to connect traffic-light controllers, public Wi-Fi kiosks, and municipal offices without leasing capacity from a private telecom company for every single link.

A MAN is best understood as a network of LANs — it takes several buildings, each with its own tidy local network, and stitches them together with links fast and reliable enough that the seams mostly disappear. Telecom companies and universities were the earliest builders of these city-scale backbones, often laying fiber rings specifically because a ring, unlike a single straight line, can survive one break by simply sending traffic the other way around — an idea that echoes all the way back to the self-healing dual-ring designs you will meet again when topologies come up in the next chapter.

The word "metropolitan" is doing real work in that name. A MAN is deliberately sized to a city's geography — usually somewhere between one and a hundred kilometers — which is exactly the range where a private company or a single university system can justify running and owning its own dedicated fiber, but where reaching an entire country or continent would require a completely different scale of investment. That in-between quality is precisely why MANs, of the four categories in this chapter, are the one most people have never consciously heard of, even while they benefit from one every day they use a big city's public Wi-Fi.

flowchart TB
    ring["City Fiber Ring<br/>self-healing: traffic reroutes<br/>around any single break"]
    lib["Library (LAN A)"]
    dorm["Dormitory (LAN B)"]
    hall["Lecture Hall (LAN C)"]
    hospital["Hospital (LAN D)"]
    lib --- ring
    dorm --- ring
    hall --- ring
    hospital --- ring

A MAN stitches together several independent LANs across a city. The fiber ring is the critical architectural choice: if any single segment is cut — by construction, weather, or equipment failure — traffic simply flows the other direction around the ring. Each building still runs its own internal LAN; the ring makes them behave as one.

WAN: Wide Area Network

And now widen the circle all the way out, past any single city, past any single country, until it wraps the entire globe: the Wide Area Network, the category the internet itself belongs to. A WAN is built by leasing or physically laying long-distance links between many independently owned smaller networks — a bank connecting every branch and ATM back to central servers, a multinational company tying its New York, London, and Tokyo offices together over private circuits or encrypted tunnels, or, at the largest possible scale, the public internet knitting together essentially every network on Earth that has agreed to exchange traffic with its neighbors.

The physical story of the WAN is, quite literally, a story about the bottom of the ocean. The first transatlantic telephone cable, TAT-1, went into service in 1956, carrying a mere handful of simultaneous phone calls between Scotland and Newfoundland along a copper line thick enough to survive years underwater. Fiber optics changed the math entirely: TAT-8, laid in 1988, was the first transatlantic cable to carry light instead of electricity, and it could carry roughly a thousand times more simultaneous calls than TAT-1 — from a few dozen to tens of thousands — through a strand of glass thinner than a human hair. Today there are hundreds of these cables criss-crossing every ocean on the planet, and if you strung them all together they would wrap the Earth more than thirty times over. They have names, almost like ships — MAREA, Southern Cross, Dunant — and they occasionally suffer a threat that sounds invented but genuinely is not: sharks have been documented biting undersea cables, apparently drawn to the faint electromagnetic field the cables give off, forcing some operators to wrap sections in shark-resistant sheathing.

WANs are where the polite fiction of "the network never fails" finally gives way to the truth every network engineer eventually learns to live with: links this long will be cut by a fishing trawler's anchor, a backhoe on a construction site, or a curious shark, and the entire discipline of WAN engineering is built around routing around that reality rather than pretending it will not happen.

flowchart TB
    internet["The Public Internet<br/>(WAN backbone)"]
    ny["New York Office<br/>Private LAN"]
    london["London Office<br/>Private LAN"]
    tokyo["Tokyo Office<br/>Private LAN"]
    ny ---|"~70 ms RTT"| internet
    london ---|"~35 ms RTT"| internet
    tokyo ---|"~150 ms RTT"| internet

A WAN connects independently owned networks across continents. The latency labels — New York to London at roughly 70 milliseconds, Tokyo to either at over 150 milliseconds — are physical limits set by the speed of light in glass, not by the capacity of the line. No amount of money buys those milliseconds back.

Seeing the Same Idea at Every Scale

Lay all four side by side and a pattern falls out that is easy to miss when you meet them one at a time: every step outward trades intimacy for reach. A PAN knows you personally and barely reaches past your skin. A LAN knows your household or your office and barely reaches past your walls. A MAN knows your city and reaches across it on infrastructure someone deliberately built and owns. A WAN knows nothing about you at all — it simply agrees, link by link, continent by continent, to carry whatever arrives and hand it to whoever the address says should receive it.

None of this is purely academic trivia, and it is worth resisting the temptation to file it away as a table to memorize for a quiz. A network engineer troubleshooting a slow connection asks, almost instinctively, "which of these four am I even looking at right now?" before touching anything, because the answer changes everything about what could plausibly be wrong. A dropped Bluetooth connection, a Wi-Fi dead zone, a city fiber ring outage, and a severed transatlantic cable are four completely different failures, with four completely different fixes, that all present to an ordinary user as the exact same symptom: "the internet stopped working."

One more wrinkle is worth carrying with you before you close this chapter: real infrastructure rarely respects these categories as cleanly as a diagram suggests. A corporate VPN can make an office in another country feel, to the software running on it, exactly like it is sitting on the same LAN as headquarters — the applications have no idea they are actually crossing a WAN link at all. Do not be surprised, once you start working with real systems, when the tidy PAN-LAN-MAN-WAN boundaries blur at the edges; that blurring is a sign engineers keep finding clever ways to make one kind of network behave like another when it is convenient to do so.

The best way to make any of this stick is to start noticing it. The next time you connect a pair of headphones, glance at the Wi-Fi icon on your laptop, or watch a video buffer for a moment during an international call, ask yourself which of these four circles you are actually standing inside — and remember that every one of them, from the tightest to the widest, is running on the same three ingredients you met in the first chapter: nodes, links, and an agreed-upon set of rules. Once you begin writing real networked code, starting in the Go Fundamentals part of this book, you will find yourself building software that has to behave differently depending on exactly this distinction — a chat server built for a LAN can make assumptions about latency that the same server, deployed across a WAN, simply cannot afford to make.

Frequently Asked Questions

Are PAN, LAN, MAN, and WAN just arbitrary labels for "small," "medium," and "large," or do they mean something more precise? They're closer to engineering decisions than casual size labels — each boundary marks a point where the priorities that matter most (battery life, trust, ownership, latency) genuinely change, not just the raw distance involved. A PAN gives up range and speed specifically to sip power for an earbud battery the size of a lentil; a WAN gives up low latency specifically because no amount of money buys back the time light needs to physically cross an ocean.

Why do PAN technologies like Bluetooth exist at all instead of everything just using Wi-Fi at short range? Because Wi-Fi is built to reach across a house, and reaching across a house costs far more power than a coin-sized earbud battery can spare. Bluetooth and Bluetooth Low Energy are deliberately engineered to trade away that range in exchange for sipping power instead of gulping it, which is a completely different design goal than what Wi-Fi is optimized for.

Is my home Wi-Fi network "the internet"? No — this chapter calls that out directly as a common misconception. Your home Wi-Fi is a LAN, and everything behind your router (your laptop, your smart TV, your printer) lives on that LAN; the public internet only begins on the other side of your router, which merely introduces your LAN to it.

What's actually going on when a corporate VPN makes a remote office "feel" like it's on the same LAN as headquarters? The software running on those machines has no idea it's actually crossing a WAN link at all — the VPN is deliberately making one kind of network behave like another for convenience. This chapter flags that blurring on purpose: the tidy PAN-LAN-MAN-WAN boundaries in the table are a useful mental model, but real infrastructure regularly disguises one category as another when it's useful to do so.

Why does a MAN get its own category instead of just being called "a really big LAN" or "a small WAN"? Because it sits in a genuinely distinct sweet spot: large enough that a single building's LAN can't cover it, but small enough — usually one to a hundred kilometers — that a university or private company can still justify laying and owning its own dedicated fiber, unlike a WAN's continent-spanning scale. That in-between quality is exactly why it's the category most people benefit from daily without ever consciously naming it.

Fun Facts and Notes

• The first LAN parties were gamers in the 1990s physically lugging heavy desktop towers and tangles of cable to a friend's house so their machines could be wired together directly for a weekend of multiplayer.
• The internet's undersea cables, laid end to end, are long enough to circle the Earth more than thirty times over — and some of them have needed shark-resistant armor.