Design Option 2: DHCP Clusters
Applies To: Windows Server 2008
The DHCP Server service in Windows Server 2008 supports the use of clusters. You can increase reliability by deploying a DHCP server cluster using the Cluster service provided with Windows Server 2008 Enterprise operating system.
Clustering support allows you to implement a local method of DHCP server failover, thereby achieving greater fault tolerance. You can also enhance fault tolerance by combining DHCP server clustering with a remote failover configuration, such as split scope.
Requirements for setting up a DHCP clustered resource
The DHCP Server service, when used in a server cluster, requires a disk resource, an IP address resource, and a name resource.
In this configuration, a virtual IP address is defined for the IP address resource using the Cluster Administrator MMC. This address must be a static IP address and not one obtained from another DHCP server. The DHCP Server service binds to this virtual IP address, which must then be used to authorize the DHCP Server service in AD DS so that it can function correctly and service clients on the network.
The advantages of DHCP clusters include:
Simple administration: For example, you can use the Cluster Administrator MMC to manage DHCP server clusters that are deployed on remote computers running Windows Server 2008 Enterprise.
Online hardware upgrading: A clustered solution allows most hardware upgrades to be carried out while the DHCP server is live.
100 percent scope availability: A DHCP cluster provides for 100 percent scope availability at all times.
The disadvantages of DHCP clusters include:
Incomplete database protection: The Jet database upon which the DHCP Server service relies cannot support clusters. If the clustered DHCP database becomes unavailable to the DHCP Server service, the service will report thousands of ESENT application log errors and the DHCP service will be interrupted.
Single point of failure: There is only a single DHCP database in this solution. Although there is duplicate hardware, there is still a single point of failure. If the database becomes corrupted, the DHCP cluster will fail. You can mitigate this through automatic backups and server-side conflict detection.