Description

📊 Get more market insights HERE

🤝 Speak with someone from our global data team HERE 

🖥️ Book a platform demo HERE

In Podcast 33, we continue our Data Center Fundamentals series and cover the basics of data center infrastructure.

Electrical (Power): Electricity moves through a “power chain” from the utility to the server—typically substation → building transformer → switching station → UPS → PDU → RPP → racks/servers, with on-site generators backing up the facility if utility power is interrupted. Think of a data center like a giant laptop: wall power gets transformed into usable power, and a “battery” (backup systems) keeps things running during failures.

Mechanical (Cooling): A data center’s usable power is often limited by how much heat it can remove per rack (density). Many facilities cool around 5–10 kW per rack, though some go much higher. The most common setup uses a raised floor: cold air is pushed under the floor and up through perforated tiles, hot exhaust air is pulled away, cooled via chillers, then recirculated. In some climates, “free cooling” uses outside air to reduce costs and improve efficiency.

Connectivity: Without strong connectivity, a data center is just a building of computers. Multiple fiber providers enable low-latency access and broader customer reach. Fiber typically enters through secured vaults and routes to a meet-me-room (where carriers interconnect) or directly to customer servers.

Redundancy: Redundancy is described as N plus extras (e.g., N+1) or multiples (e.g., 2N). It applies across power, generators, cooling, UPS, and even diverse utility feeds and fiber paths. Because data centers underpin the internet, they need robust, redundant infrastructure to avoid service interruptions.