Question 1

What is actually inside a data centre?

Accepted Answer

Far more **power and cooling infrastructure than computing** — by floor area, by cost and by engineering attention. The servers are the visible part and rarely the difficult part.

**What is physically there:**

**Racks of servers**, storage and networking equipment, arranged in rows.

**Hot aisle / cold aisle containment.** Racks face each other so cold air is drawn in from one aisle and hot air exhausted into the other, which is then contained and returned to cooling. **Mixing hot and cold air is the fundamental inefficiency** this layout exists to prevent.

**Cooling** — computer room air handlers, chillers, and increasingly **free cooling** using outside air, evaporative cooling, or liquid cooling directly to the chip, which is now necessary for dense AI hardware that air cannot cool.

**Power infrastructure**, which is the bulk of it: utility feeds from more than one substation, **uninterruptible power supplies** with batteries or flywheels to bridge the seconds before generators start, and **diesel generators** with fuel for days, tested under load regularly.

**Distribution** through switchgear and busbars, with dual paths so any component can fail or be maintained without interruption.

**Fire detection and suppression**, using gas or mist rather than water where possible.

**Physical security** — mantraps, biometrics, cages, cameras.

**Network connectivity** from multiple carriers entering through separate physical routes.

**How efficiency is measured.** **PUE** — power usage effectiveness — is total facility power divided by IT power. **1.0 would mean every watt reached a computer**; typical modern facilities achieve somewhere around 1.1 to 1.5, against much worse historically.

**The tier classification** describes redundancy, from a single path with no redundancy up to fully fault-tolerant with concurrent maintainability — which determines expected availability.

**Why location matters:** power availability and price, which is now the binding constraint on new capacity; climate, since cool dry air reduces cooling cost; network connectivity; latency to users; land; and jurisdiction for data residency.

**Water use** for evaporative cooling has become a significant local issue.

Question 2

How does the internet actually cross oceans?

Accepted Answer

Through submarine fibre optic cables lying on the seabed — not satellites, which carry a negligible share of intercontinental traffic despite what most people assume. The scale. Well over 95% of

Question 3

Why do I need to restart my router so often?

Accepted Answer

Needing an occasional restart is normal. Needing one every few days is a symptom, and there are only a handful of realistic explanations. The most common is resource exhaustion. A router is a small

Question 4

What is a subnet, and why are networks divided up?

Accepted Answer

A subnet is a logically separate section of a network, and networks are divided for three reasons that are worth distinguishing: reducing unnecessary traffic, improving security, and making

Question 5

What is a point of presence?

Accepted Answer

A point of presence (PoP) is a physical location where a network has equipment that other networks or customers can connect to — the places where an abstract "network" actually touches the ground.

Question 6

What is Power over Ethernet?

Accepted Answer

Power over Ethernet (PoE) delivers electrical power along the same cable that carries data, so a device needs only one connection rather than a network cable and a mains adapter. Why it is useful.

Question 7

How does fibre actually carry data?

Accepted Answer

As pulses of light down a strand of extremely pure glass, guided by a physical effect that keeps the light from escaping even as the cable bends. The physical mechanism. A fibre has a core of glass

Related Questions