Microsoft Windows Active Directory is prone to a remote code-execution vulnerability that arises because the application fails to handle specially crafted LDAP or LDAP over SSL (LDAPS) requests in a proper manner. Successfully exploiting this issue would allow an attacker to execute arbitrary code and gain complete access to a vulnerable computer. The attacker may also be able to cause the affected system to stop responding to further requests and restart. This issue affects only Windows 2000 servers configured as Active Directory domain controllers.
Block external access at the network boundary, unless external parties require service.
Filter access to the affected computer and resources at the network boundary if global access isn't needed. Consider blocking incoming and outgoing traffic to TCP ports 389 and 636. Restricting access to only trusted computers and networks might greatly reduce the likelihood of a successful exploit.
Deploy network intrusion detection systems to monitor network traffic for malicious activity.
Deploy NIDS to detect and block attacks and anomalous activity such as requests containing suspicious network traffic.
Run all software as a nonprivileged user with minimal access rights.
If possible, running software as a user with least privileges possible can help mitigate the impact of exploit attempts against latent vulnerabilities in applications.
Implement multiple redundant layers of security.
Various memory-protection schemes (such as nonexecutable and randomly mapped memory segments) may hinder an attacker's ability to exploit this vulnerability to execute arbitrary code.
The vendor released an advisory along with fixes to address this issue. Please see the references for more information.