Combine upstream mitigation, edge caching, application filtering and protected origins with clear service boundaries and a rehearsed incident plan.
DDoS protection for websites starts with layers
A distributed denial-of-service incident threatens availability by exhausting resources from multiple sources. CISA distinguishes volumetric, protocol and application-layer techniques; these can affect bandwidth, network handling or application processing. For a business website, protection must address more than the capacity of its web server. [1]
Start with the customer journeys that matter: browsing products, signing in and completing orders. Map the infrastructure each journey depends on, then assign a defensive layer to each bottleneck. Upstream mitigation filters traffic before it reaches constrained infrastructure, while application controls and capacity planning address requests that still reach the service. No single setting covers every failure mode. [1]
CDN vs WAF: complementary roles
A content delivery network, or CDN, serves eligible content from distributed caches, reducing requests to the origin server that generates or stores it. Caching product images and public pages preserves origin capacity during traffic spikes. Personalized checkout or account responses need separate cache policies; content must be safe to share before it enters a shared cache. [2]
A web application firewall, or WAF, evaluates HTTP and HTTPS requests against rules, allowing, blocking or challenging matching traffic. Rate-based controls can focus on expensive endpoints that cannot be cached. These are different jobs: caching reduces work, while filtering decides which requests proceed. [3] A CDN’s included network protection and a WAF’s features depend on the service; verify both rather than treating the labels as interchangeable.
Protect the origin and its dependencies
Edge protection is incomplete if the origin remains directly reachable outside the intended delivery path. Restrict origin access to the authorized CDN connection using supported network restrictions or authenticated origin access. Official architecture guidance recommends ensuring that only the intended distribution can forward requests to the origin; hiding an address is not the same as enforcing access. [4]
Include DNS and other critical services in the availability review: CISA notes that infrastructure beyond the website itself can be affected. [1] Document who manages those dependencies and how failures are detected. Use the secure web hosting checklist to place DDoS controls alongside patching and administrative security, rather than treating availability as the whole security program.
What DDoS protected hosting actually covers
Treat DDoS protected hosting as a scope to verify, not an unlimited guarantee. Ask which attack layers, protocols, addresses and services are covered. Establish whether mitigation is always active or requires escalation, who changes filtering rules, and what happens when traffic exceeds the agreed protection scope. Obtain written answers before an incident makes those boundaries urgent.
Review bandwidth allowances, request charges, scaling costs and support entitlements together. Ask whether an overwhelmed service may be temporarily isolated, and under what restoration process. Separate mitigation commitments from uptime credits: the guide to uptime SLAs and recovery objectives helps translate contractual language into business expectations. A credit should not be mistaken for a recovery deadline or proof that checkout will remain available.
Prepare the response before traffic spikes
CISA recommends normal-traffic baselines, monitoring, defined incident roles and provider coordination. [1] Record request rates, latency, errors and origin resource usage so responders can compare symptoms with ordinary demand. Assign an incident lead, retain escalation contacts outside the affected website, and define who can approve temporary filtering changes. Rehearse decisions through a tabletop exercise before an outage.
During disruption, activate the approved plan, contact hosting and mitigation teams, and preserve timestamped logs and alerts. Communicate customer impact without promising an unverified restoration time. After service returns, review which controls worked and update the runbook. [1] Include checks for legitimate customers affected by emergency rules; restoring availability means restoring usable service, not merely lowering traffic counters.
Sources & further reading
These resources provide background for the explanations above. External links are included for context, not as endorsements.
Joint defensive guidance on attack categories, monitoring, mitigation-provider coordination, incident response and recovery.
Official implementation guidance supporting the role of caching and separate controls for dynamic endpoints; not a provider recommendation.
Current official documentation describing HTTP and HTTPS request inspection, allow and block rules, challenges and rate-based controls.
Official guidance on restricting origin access to the intended CDN distribution, including private origins and origin access controls.
Written with AI assistance and based on the sources listed above. Published November 16, 2026. No hands-on provider testing is claimed. Read our editorial policy.
