Software Engineering · Distributed Systems
The system keeps operating even when a network failure splits it into isolated groups that can no longer communicate.
Example: Two data centers keep functioning even after losing connection to each other.
In practice: In practice, network partitions are treated as inevitable in any distributed system, which is why CAP theorem debates usually come down to consistency vs. availability, not whether to tolerate partitions at all.