EMZETT.Power Distribution
In short: The distribution of electrical energy from a source (e.g. a power supply or power strip) to several consumers — from a single PC to an entire server room.
In more detail: In a PC, the power supply distributes power to the motherboard, CPU, graphics card and drives via various cable strands. In data centres, power distribution is its own field: redundant supply lines (UPS, several feeds) are meant to cushion the failure of individual components, so servers keep running continuously.
In Depth
Distribution inside a PC
Inside a single PC, the power supply distributes different voltages (typically 12V, 5V and 3.3V) to different consumers via separate cable strands — power-hungry components like the graphics card often get their own, dedicated power connectors directly from the power supply, instead of sharing supply through the motherboard, to avoid voltage drops under high load. How many such separate connectors (e.g. several 8-pin PCIe connectors for high-end graphics cards) a power supply provides is an important selection criterion when putting together powerful systems. Modern, high-quality power supplies are also modular: only the cables actually needed are plugged in, which reduces cable clutter in the case and improves airflow.
Efficiency classes and 80 PLUS certification
When converting mains power (AC) into the DC voltages a PC needs, some energy is always lost as heat. The 80 PLUS certification system (levels from Standard through Bronze, Silver, Gold to Platinum and Titanium) states how efficiently a power supply works — an 80 PLUS Gold power supply loses less than 12% of the energy drawn under typical load, a simple, uncertified model considerably more. Over a system’s lifetime, this difference adds up to noticeable electricity costs, especially under continuous operation.
Data centres as their own discipline
At large scale, such as in data centres, power distribution becomes its own engineering discipline with several layers of redundancy. Uninterruptible power supplies (UPS) are used, which instantly (within a few milliseconds) switch to battery backup in the event of a power outage, while at the same time a diesel backup generator spins up, which typically takes over from the batteries after a few minutes and can sustain operation for hours or days. Redundant feeds from different power grids or substations are also meant to ensure that even a large-scale regional grid outage doesn’t interrupt operation.
N+1 and 2N redundancy
Data centres often classify their power redundancy according to standardised levels: N+1 means an additional reserve unit exists beyond the actual need (if a component fails, the reserve takes over seamlessly); 2N means complete duplication of all critical components across two fully independent paths. This classification is closely related to the so-called tier levels (Tier I through IV) for data centres overall, and is crucial so servers can keep running independently of external power problems — especially for services that depend on continuous availability (often contractually guaranteed as an SLA with 99.9% or higher availability).
See also: Power