A wave of DNS hijack attacks has been sweeping across Europe, the Middle-East and America according to recently published reports from FireEye and Cisco. While the attacks are creative and sophisticated, the root attack vector is often a simple credential compromise to the DNS control panel of an organisation’s domain name registrar. The scale of the campaign is such that US-Cert has issued an alert and various Federal agencies have been given 10 days to take urgent action to re-secure their infrastructure.
The credentials to access domain registrars or DNS provider control panels are typically not used frequently, and for many organisations, a single login is shared between several administrators. Compromise of these credentials using traditional phishing techniques is allowing attackers to mount sophisticated and almost invisible attacks against many organisations.
Once the attackers are able to access the DNS control panel they edit the DNS records to divert web traffic to a server controlled by the attackers. Control of the DNS records allows an automated service such as LetsEncrypt to generate a valid TLS certificate for the attacker’s server and then present a login page which harvests users login credentials and passwords. The web traffic is then forwarded to the real server and users continue unaware that their credentials have been compromised. A detailed description is available in the following report from FireEye:
Administrators are advised to take steps to protect their domain administration control panels. These steps include:
- Turning on multi-factor authentication if available
- Changing (shared) login passwords for administration portals
- Regularly checking DNS records for unexpected changes
- Reviewing server access logs for high traffic volumes from unexpected IP addresses
One of the checks conducted by SecureTeam during an External Penetration Test identifies the DNS records which are configured for the customers domain name, with specific checks being conducted on the IP addresses which the domain records point to in order to ensure they do not point to a malicious server belonging to an attacker.