DNS threat guide

DNS poisoning: protect the integrity of DNS answers.

DNS cache poisoning occurs when incorrect DNS data enters a resolver’s cache and is reused for later requests. Forged responses are one way to attempt it. Defense requires checking which answer belongs to a request and what evidence supports trusting its data.

DNS poisoning versus response spoofing

Response spoofing concerns the acceptance of a forged answer. Cache poisoning concerns the retained state: bad information can affect more than the original lookup while it remains reusable. A wrong address may divert clients away from the intended service, so cache integrity is part of DNS security as well as resolver correctness.

Checking only that a packet looks like DNS is insufficient. Resolvers must associate responses with outstanding questions and carefully handle referrals and additional address records. RFC 5452 describes safeguards against forged answers; cryptographic validation adds a separate basis for trusting signed DNS data.

Raven checks responses and referrals

Raven requires strict response-to-question binding. During iterative resolution, it checks referral progression and bailiwick: attached name-server addresses are trusted only within the relevant scope. Out-of-bailiwick name-server addresses are resolved independently rather than used directly as a trusted destination.

Its cache matches request identity and the active generation, respects expiry, and does not serve stale answers after a lookup failure. These rules support consistent cache behavior, but cache freshness alone is not proof that a DNS answer is authentic.

What DNSSEC validation adds

When enabled, Raven validates DNSKEY, DS, RRSIG, and authenticated denial evidence locally. It does not trust an upstream resolver’s AD flag as proof. A valid chain for signed data can establish authenticity and integrity; unsigned data cannot receive the same cryptographic assurance.

Client flags change the response. With CD=0, a Bogus result returns SERVFAIL, with additional error information when possible. With CD=1 (Checking Disabled), unchecked data may be returned with AD=0. An Insecure result means there is no complete signed chain, not that the domain has been classified as malicious.

Keep authenticity, privacy, and policy distinct

DNSSEC does not encrypt DNS traffic or decide whether a correctly signed domain is safe to visit. DNS over TLS and DNS over HTTPS protect a transport connection. DNS firewall policy evaluates destinations. These controls answer different questions and should be evaluated separately.

Investigate failed lookups using the DNSSEC result, response code, client flags, and upstream conditions. A SERVFAIL can reflect a resolution problem or validation failure. Temporary Negative Trust Anchors in Raven require an exact non-root domain, a reason, and an expiry; they are exceptions to trust validation, not a general repair for poisoning.

DNS security guides

Understand the threat, the control, and the deployment conditions that affect protection.

DNS firewall

Understand DNS firewall filtering with Raven DNS Guard: domain policies, RPZ threat feeds, request and response inspection, and explicit enforcement actions.

DNS attacks

Compare DNS attack types: amplification, cache poisoning, tunneling, rebinding, query floods, and DGA activity, with their distinct controls and limits.

DNS amplification

Learn how DNS amplification and reflection work, why open resolvers matter, and how Raven access controls, rate limits, and response controls reduce abuse.

Review your DNS requirements

See the controls in context.

Walk through Raven’s policies, resolver behavior, and investigation views with our team.

Book a demo