Server farm
A server farm, or server cluster, is many computers working together to provide more power and services than one machine can. They can include thousands of servers and need a lot of electricity and cooling, so running them can be expensive. Many farms have backup servers that can take over if a main server fails. The machines, network gear, and power equipment are usually kept in racks in a data center.
Uses include web hosting (a web farm), scientific simulations, and rendering 3D images (render farms). They are also used for cluster computing and are sometimes chosen instead of large mainframes by big companies.
In large farms, individual servers fail from time to time, so farms rely on redundancy, automatic failover, and quick reconfiguration to keep services running.
Power and cooling are the biggest limits, not raw speed. Computers run 24/7 and use a lot of energy, so designers focus on performance per watt and use power-saving features like variable clock speeds or turning off unused parts or whole machines when demand is low.
The internal network is crucial for moving large amounts of data.
Benchmarks exist to predict power efficiency, and power can be tracked at the rack or server. A common measure is PUE, which compares total data center power to the power used by the servers.
A rough rule is that cooling can use about half as much energy as the servers. Location matters: colder climates and cheap, renewable power help. Iceland is building a large site with geothermal power and fiber links to North America and Europe. Other countries such as Canada, Finland, Sweden, and Switzerland are trying to attract data centers by offering favorable conditions, sometimes using server heat to warm buildings.
This page was last edited on 3 February 2026, at 16:38 (CET).