denial of service

Confidentiality and integrity are about keeping secrets and detecting tampering — but there's a third security goal, availability, and a whole family of attacks aimed squarely at it. Imagine someone jamming a shop's only phone line by calling it nonstop, so no real customer can ever get through. The shop's data is safe and unread, but the service is effectively dead. A denial-of-service (DoS) attack does exactly this to a computer or network: it makes a service unavailable to its legitimate users, not by breaking in, but by overwhelming or jamming it.

DoS attacks work by exhausting some finite resource the target needs to function. That might be network bandwidth (flood the link with so much junk traffic that real packets can't get through), connection state (open thousands of half-finished connections to fill a server's connection table — the SYN flood), CPU or memory (send requests that are cheap to make but expensive to answer), or an application-specific limit. The unifying idea is asymmetry: the attacker spends a little effort to consume a lot of the victim's capacity, tipping the system into overload so it can no longer serve anyone.

Why it matters: availability is a real security property, and DoS attacks the property directly — they don't steal or alter data, they deny access to it, which can be just as damaging (an unreachable bank, hospital, or emergency line). Defenses include over-provisioning capacity, filtering or rate-limiting abusive traffic, and detecting attacks early. The hard part, and the honest difficulty: a single attacking source is relatively easy to identify and block, so attackers escalate to thousands of sources at once — the distributed denial-of-service (DDoS) — which is far harder to defend against because the flood looks like it comes from everywhere.