AI EncyclopediaChips & deep techData centers & power
Server
서버
A computer that takes requests from other computers and does work for them. AI services run on servers too.
Overview
When you watch a video on your phone or ask an AI a question, the request travels over the internet to a server far away. The server does the work and sends the result back. It works on the same principle as a computer at home, but it is a flat box with no screen or keyboard, built tough enough to stay on without a break.
In a data center, servers are slotted into shelves called racks, stacked in rows. An AI server is a server with several AI chips such as GPUs inside. It costs much more than an ordinary server, uses a lot of electricity and gives off a lot of heat. So how much power one rack uses and how to cool it have become central problems in designing data centers.
An everyday example
At a restaurant, a customer (your device) places an order and the kitchen (the server) makes the food and sends it out. When more customers come, the kitchen has to grow, and when more people use a service, servers have to be added.
Why it matters
You often hear "the server went down" and "orders for AI servers are rising." Servers and racks are the units in which AI chips are actually sold and used, so it helps to know them when you read chip news.