What Is DDoS Protection? A Plain-English Guide
What it does, how it works, and how to tell real protection from marketing.
The 30-Second Version
A DDoS attack tries to overwhelm your server with junk traffic so legitimate users can't reach it. DDoS protection is a filtering layer — usually upstream of your server — that drops the junk before it gets to you. Real protection means your service stays up during the attack. Fake protection means your IP gets null-routed (taken offline) "to protect the network."
How It Actually Works
Three stages:
- Detection — traffic patterns get analyzed (volume, packet types, source distribution). Real protection runs detection continuously, not after an attack starts.
- Diversion — when an attack is detected, traffic is routed through scrubbing centers — clusters of hardware designed to inspect and filter packets at line rate.
- Mitigation — malicious packets are dropped; clean traffic is forwarded to your server.
The whole loop happens in milliseconds when the network is well-designed. Always-on mitigation (KoDDoS, Path.net) keeps detection running constantly. Reactive mitigation (some cheaper hosts) starts the loop after the attack begins, causing 30–120 seconds of downtime.
Layer 3/4 vs Layer 7
Layer 3/4 (volumetric): UDP floods, SYN floods, amplification attacks. Measured in Gbps. The big, loud attacks.
Layer 7 (application): HTTP floods, Slowloris, game-query floods. Lower bandwidth, much harder to filter without affecting legitimate users. The sneaky attacks.
Modern protection must handle both. Read the deep-dive.
How to Tell Real Protection From Marketing
- Published mitigation capacity in Gbps or Tbps, with ASN evidence.
- Mention of Layer 7 — not just Layer 3/4.
- "Always on" or "always inline" — not reactive.
- Real money. Plans under $5/mo with "unlimited DDoS protection" almost never have real mitigation.
