Remediating DNS poisoning that redirects traffic

DNS poisoning turns a trusted domain name into a redirection mechanism for an attacker. By injecting forged responses into a recursive resolver or altering authoritative DNS records, the attacker can send users to fraudulent servers while the address bar still displays a legitimate URL. The resulting credential theft, malware delivery, session hijacking and payment diversion can continue long after the first suspicious request appears.

Remediation requires more than replacing a poisoned DNS record. Teams must determine which part of the resolution path was compromised, contain malicious traffic, remove hidden persistence and restore service without destroying evidence. Built by Exclusive Networks to integrate technologies from multiple security vendors, CARM Security supports this coordinated approach by helping enterprises identify, contain, respond to and remediate post-breach activity across the wider attack surface.

Recognising malicious redirection

The first signs are often inconsistent rather than dramatic. Users may reach an unfamiliar login portal, encounter a certificate warning, see one service while another appears offline, or report that an application is connecting to an unexpected IP address. DNS and proxy logs become especially valuable when the problem affects only users sharing a particular resolver, office network or managed security provider.

Signal More consistent with DNS poisoning More consistent with a benign DNS or CDN issue
Resolution results Different addresses across regions or resolvers Stable, documented anycast addresses
Affected users Clusters sharing a resolver or branch network A specific application, browser or device version
Website behaviour Legitimate domain redirects to unrelated infrastructure Correct origin with an expired certificate or temporary CDN fault
Change history Unauthorised record, account or delegation changes Approved migration, maintenance or certificate renewal
Traffic destination New or previously unseen hosting infrastructure Known service provider ranges with normal certificate coverage

Geographic comparison can help narrow the fault. A Sydney office resolving an address differently from Melbourne and Brisbane may indicate regional cache poisoning, whereas identical suspicious results everywhere suggest a change to the authoritative zone. Australian customs brokers and freight operators often depend on shared portals and third-party connectivity, so a seemingly local redirection can affect a chain of organisations across several cities.

A redirect can also resemble a legitimate security feature. Provider sinkholes often send harmful requests to a warning page, while content delivery networks may temporarily route traffic to a maintenance host. Investigators should therefore validate the destination, certificate, hosting history and DNS response path rather than treating every unexpected IP address as proof of an attack.

Establishing what the attacker changed

The resolution chain must be mapped before records are edited. That chain commonly includes the domain registrar, authoritative nameservers, recursive resolvers, service-provider DNS, endpoint configuration and any dynamic update system. The objective is to identify where forged answers entered the chain and whether attacker-controlled accounts could recreate them.

Multiple clean resolvers, passive DNS records and historical configuration data should be compared with the response received by affected users. The fact that a domain ends in .au does not make an incident local in the narrow sense: auDA operates the namespace, while registrars, hosting providers and recursive resolvers remain separate administrative layers. A registrar compromise, modified zone file and poisoned cache can produce very different security consequences even when the visible domain is identical.

Scanner findings require careful interpretation during this triage. A vulnerable service and an exploitable DNS change are not the same issue, and a report may describe a plausible weakness without proving active exploitation. Reviewing scanner false positives can help teams separate noisy automated detections from evidence connected to the redirection. Configuration changes should also be checked against approved change records and administrator access logs.

Containing the redirect without losing evidence

Immediate containment should block the malicious destination while preserving enough information to investigate the intrusion. Security teams can apply DNS, proxy and firewall blocks, isolate affected resolvers and temporarily redirect known-bad destinations to a controlled sinkhole. Logs, resolver caches, firewall rules, endpoint artefacts and cloud audit trails should be captured before devices are restarted or configurations are replaced.

Blocking an IP address alone is rarely durable. Attackers can rotate infrastructure, alter the authoritative response or compromise another internal resolver. Temporary controls should therefore operate alongside investigation of the registrar, DNS hosting account, nameserver delegation and internal network configuration. Hard-coded resolver addresses may provide short-term stability, but they do not replace root-cause remediation.

If the redirection involved a proxy or convincing login page, credentials and active sessions must be treated as exposed. Security teams should revoke suspicious tokens, reset privileged accounts, enforce phishing-resistant multifactor authentication and review recent authentication from affected users. Containment records should include exact timestamps and the scope of each action so later analysis can distinguish attack activity from changes made by defenders.

Organisations should also establish clear notification and coordination procedures. Australian entities can use the ASD-run Joint Cyber Security Centre network to coordinate with government and industry partners when an incident has broader significance. Customers, payment providers, suppliers and telecommunications carriers may need early warnings, while legal, regulatory and insurance obligations should be assessed through the appropriate Australian channels.

Removing poisoned records and persistence

Authoritative DNS should be restored from a known-good source rather than copied blindly from an affected resolver. The registrar account needs its malicious sessions terminated, passwords rotated and multifactor controls checked. Zone files, delegation records, wildcard entries, dynamic updates and administrative APIs should be compared with the last trusted version. Where DNSSEC is used, signatures and the chain of trust must also be validated.

Recursive caches remain a risk after the authoritative configuration is repaired. Providers should purge affected zones, and organisations should check whether local resolvers, smartphones, VPN clients and managed devices retained forged answers. In distributed environments, validation from separate Australian and international networks can expose a regional cache that internal testing misses. Restoration should proceed only when resolution and routing produce the expected result from each relevant path.

The DNS change may be one stage of a larger compromise. Investigators should inspect endpoint management platforms, DHCP services, browser policies, scheduled tasks, OAuth applications, mail rules and cloud identities for persistence or secondary access. CARM Security can combine vendor telemetry during this wider review, linking the visible redirection to suspicious activity elsewhere and preventing responders from treating DNS restoration as the end of the breach.

Coordinating cleanup across teams and providers

Large incidents often involve registrar engineers, network teams, identity specialists, application owners, legal advisers, external forensic teams and managed security providers. One incident commander should control priorities, while a shared evidence register records what each party observed and changed. Duplicate remediation, especially repeated password resets or cache purges, can obscure the original sequence of events.

Australian organisations frequently rely on local managed security service providers for 24-hour monitoring and incident support. A provider may operate the recursive resolver used by multiple branches or customers, making provider-level investigation essential. For enterprises engaging an MSSP after a breach, CARM Security’s guidance on coordinated post-breach cleanup illustrates how shared responsibilities, evidence transfer and defined ownership can accelerate restoration without weakening security controls.

Completion criteria should be measurable rather than based on a single successful webpage load. They can include correct authoritative records, clean responses from independent resolvers, revocation of malicious sessions, restoration of affected applications, confirmation that persistence mechanisms are absent and continuous monitoring for renewed unauthorised changes. Independent validation by a party that did not perform the repair adds assurance, particularly when shared DNS infrastructure affects critical services.

Restoring service and preventing recurrence

Service restoration should be staged. Low-risk services can return first after validation, followed by sensitive applications once identity, routing and logging checks are complete. Teams should compare traffic destinations with approved provider ranges, examine certificate transparency records and watch for a sudden return of the original redirect. A stable resolution result alone does not prove that the underlying domain account remains secure.

Preventive controls should align with the Australian Signals Directorate’s Essential Eight. Relevant measures include regular patching, phishing-resistant multifactor authentication, application control, secure configuration, centralised logging, backups and tested incident-response procedures. DNS administration should receive the same protection as cloud consoles or financial systems, including restricted access, approval workflows, tamper-resistant logging and alerts for delegation or record changes.

Long-term assurance depends on continuous monitoring rather than a one-time cleanup. Passive DNS comparisons, certificate and registrar alerts, resolver-health checks and regular access reviews can identify unauthorised changes before users are redirected again. The practical takeaway is straightforward: restore DNS from verified records, invalidate every contaminated cache and credential, validate the correction from independent Australian and overseas resolvers, and monitor the domain until normal resolution and service behaviour remain stable.