What Is a Server? A Computer That Stays On When Yours Doesn't

TL;DR: A server is not a special machine. It is a computer that isn’t yours and stays on for you all the time. Because it keeps running even when you step away or close your laptop, it keeps doing the job it was set up to do. This article confirms that by having you run your own computer as a server for a moment, and by looking at a photo of a real server rack.

When you ask AI to build an automation, you often hear, “This needs to go on a server to keep running.” That’s where people get stuck. Ask what a server is and you’ll hear “a computer that handles requests,” but then the words “request” and “handle” are unfamiliar too. In the end, the single word “server” becomes a wall that keeps you from starting.

This article gets past that wall not with an analogy but with direct verification. A server isn’t a concept that exists only in computer science textbooks. Your laptop can briefly become one too. After reading, you’ll know exactly what “put it on a server” means, and you’ll have turned a server on and off yourself.

Contents

1. A Server Is Not a Kind of Machine, but a Machine That Stays On
2. This Is What One Looks Like
3. Why the Word “Server” Is Confusing
4. Let’s Briefly Turn Your Computer Into a Server
5. Two Ways to Use a Server
6. If You Want to Read More

A Server Is Not a Kind of Machine, but a Machine That Stays On

Whether it’s a laptop or a desktop, a computer is just a computer. It’s easy to mistake a server for a different kind of machine, but what separates a server from an ordinary computer isn’t its appearance. It’s whether it stays on.

If you run automation on your own computer, that automation lives only as long as the computer is on. Close the laptop or restart it, and the automation stops at that moment. But if you want automation to keep working at dawn or while you’re asleep, something has to stay on. The computer that takes on that role, staying on for you, is called a server. You can use your own computer as a server, but usually you rent a computer in someone else’s data center. When usage is low, many services let you rent one for free.

When you run automation on your own computer

1. Now Your laptop is on The automation runs normally
2. A little later You close the lid or shut it down The power takes the automation with it
3. Result The automation stops too It resumes only when you turn it back on

Your computer's power is the automation's power

When you run automation on a server

1. Now The server is on It stays on, separate from your computer
2. A little later You close the lid or shut down your laptop That has nothing to do with the server's power
3. Result The automation keeps running It runs as scheduled even when you're away

Your power and the automation's power are separated

This Is What One Looks Like

Servers feel abstract because you can’t see them. The word “cloud” sounds like something without substance, but behind it is a machine that is really plugged in and running. Below is a photo, taken from behind, of a rack with several servers stacked inside a data center.

Why the Word “Server” Is Confusing

The same word can point to different things depending on the sentence, which makes it confusing. You only need to tell apart three meanings.

Where you see “server” in a sentence	What it actually refers to	Covered in this article?
“Put it on a server and it keeps running”	A computer that is always on and not yours	This is the meaning this article covers
“The server handles that request”	A role that receives requests and returns responses	Covered in What Is an API?
“Create a Discord server”	A name for grouping chat rooms	It’s a separate word, not a computer

The first and second are actually the same computer viewed from different angles. Looked at by what it does, it’s the role of handling requests. Looked at by where it is, it’s a computer that stays on. This article covers the “where” side.

Let’s Briefly Turn Your Computer Into a Server

If terminals are new to you, read What Is a Terminal? first. This experiment is for macOS. Windows uses different commands and steps, so this article doesn’t cover it. Even if you only read along, you can still understand the idea that “it disappears when you turn it off.”

You don’t need Python installed beforehand. On a Mac, typing just python3 in the terminal often brings up an install prompt. Follow it to install Python, then continue from there.

1. Open Terminal and move to any empty folder. For example, to create a new folder on the desktop and enter it, type mkdir ~/Desktop/server-test && cd ~/Desktop/server-test.
2. Create a short page file inside this folder.

   echo '<meta charset="utf-8">' > index.html
   echo '<h1>내가 만든 페이지</h1>' >> index.html

   Without the first line, the Korean text breaks. The server you’re about to start does not tell the browser which character encoding (the text format) the file was written in, so the file itself has to declare upfront that it is UTF-8.
3. In the same terminal, type python3 -m http.server and press Enter. The cursor may look stuck, but that’s normal. This computer has just become a server, and it is now waiting for requests while staying on.
4. Open a browser and type http://localhost:8000 in the address bar. localhost is the address that means “this computer itself,” and the 8000 after it is a port number, which keeps multiple programs inside one computer from getting confused. The heading you just wrote appears with its text intact. Refresh and the same screen appears each time, because the computer running in the terminal responds to every refresh.
5. When you’re done checking, return to the terminal and press Ctrl + C (hold the Control key and press C; it’s a common shortcut that immediately stops whatever program is running) to shut down the server you just started.
6. Refresh the browser again. This time the page that loaded fine won’t appear, and you’ll see an error saying the connection can’t be made. The wording varies by browser, but they all say essentially the same thing: the connection failed.

Two Ways to Use a Server

When people say they “upload” something to a server, what they actually do falls into two kinds. One is running code each time a request comes in and producing the result. The other is simply leaving a prepared file open as it is.

Type	What the server does	Example
Runs code	Runs the set code each time a request comes in and handles it	Deploying a Supabase Edge Function
Publishes files	Keeps a prepared file open as it is. It doesn’t process anything, it just shows it	Opening a Site with GitHub Pages

Calling publishing “processing” is inaccurate. A file uploaded to GitHub Pages always returns the same content to whoever asks, without calculating anything differently per request. The reason you still need a server is the same, though. The file has to live on some computer, so that even when you turn off your computer or step away, that address stays open.

Within the runs-code type there are differences too. Some options, like Supabase Edge Functions, wake up only briefly when a request arrives (serverless). Others stay always on and wait. Either way, the core fact is the same: the computer is not your own.

When people think of a server, they tend to picture a company’s IT room or a monthly fee first, but that’s not necessarily the case. There are already servers you can use for free with just a Google account, and the GitHub Pages and Supabase Edge Functions covered in this article are free at low usage. However, free limits differ by service and change over time, so the most accurate way to check before you use one is the pricing page of each service.

If You Want to Read More

This article focused on pinning down what a server is. If you want to look into choosing an actual cloud server, here are two resources. Both are in Korean and free to read.

• AWS “What Is a Cloud Server?” (Korean-language page) https://aws.amazon.com/ko/what-is/cloud-server/ Covers the definition and how cloud servers work, and how they differ from physical servers.
• Samsung SDS Cloud Glossary: “What Is a Cloud Server?” (Samsung SDS is a large South Korean IT services company.) Gives a short summary of the concept of cloud servers and their advantages over traditional servers.