Single point of failure
Any component, path or shared dependency whose loss removes the service, regardless of how much redundant equipment surrounds it.
Also called: SPOF
In practice
The phrase is used loosely, which weakens it. Precisely, a single point of failure is an element with no alternative path around it for the service in question. It is defined relative to a service and a boundary: something that is a single point of failure for one hall may be irrelevant to a service running in another.
Redundant equipment does not remove them by itself. Two supply paths fed from one utility service share a source. Two transfer arrangements driven by one controller share a decision. Two feeds running through the same cable route, the same riser or under the same floor section share an exposure to a physical event. Counting equipment finds none of these; tracing paths finds all of them.
Some of the most durable ones are not equipment at all. A control system, a management network, a firmware version common to both paths, a single supplier for an unusual part, or one person who knows how the transfer scheme is really configured — each can remove a service as effectively as a failed breaker, and none appears on a single-line diagram.
The practical exercise is unglamorous: pick a service, trace every dependency it genuinely has, and ask what happens if each one disappears. Most rooms have at least one single point of failure, which is normal. What is avoidable is not knowing which it is until it acts.
Scope of this definition
This is an educational summary of how the term is used in data center practice. It is not a standard, a specification or engineering advice, and where a real decision depends on it, the current adopted standards, verified site information and qualified professional review are the correct sources.