Computational cost is the time, CPU cycles, memory, and other processing resources required to complete a calculation. In cyber security, it matters because every security mechanism has a cost, and attackers often try to force systems to spend more resources than they should. A task that is cheap for a single user can become expensive at scale, turning a valid feature into a bottleneck.
Defenders use computational cost as part of design choices. Strong password hashing, encryption, malware scanning, and anomaly detection all consume resources, so engineers balance protection against latency and server load. Attackers exploit that balance with resource-exhaustion tactics, expensive login attempts, or inputs that trigger slow parsing and verification. Good security engineering looks for methods that are expensive for the attacker but affordable for the defender.