Design Option 1: Split Scopes
Updated: November 10, 2008
Applies To: Windows Server 2008
The split scope method refers to the use of two separate DHCP servers that are configured with duplicate scopes. One part of the scope is excluded on the first server, the remaining part of the scope is excluded on the other server; you can use the 80/20 rule to divide the scope addresses between the two DHCP servers. If Server 1 is configured to make available most (approximately 80 percent) of the addresses in a range, then Server 2 can be configured to make the remaining addresses (approximately 20 percent) available to clients. For more information about the 80/20 rule and other DHCP best practices, see the DHCP Server Operations Guide.
The advantages of split scopes include:
Cost: Although split scope designs require the purchase of twice as many servers as non-split scoped designs, this method does not require the purchase of an external disk system for clustering.
Protection from database corruption: Unlike DHCP clusters which offer no redundancy in the event of database corruption, the split scope method allows the other DHCP server to service clients while any database issues are being resolved on the affected server.
The disadvantages of split scopes include:
Limited IP availability during outages: If a server is unavailable, the remaining servers must service the clients with the addresses available to it. As each scope covers just a portion of the entire range, the number of free addresses is significantly reduced. This reduction could be crucial when a scope is heavily utilized and one DHCP server is unavailable for a long period of time. During a long-term failure, the smaller scope might not be able to service all of the client requests.
Significantly higher total cost of ownership (TCO): If using a single DHCP server or a cluster, only one scope has to be created and documented and one UDP port (port 67) forwarding address has to be configured on routers or relay agents for each subnet. Split scope designs double these numbers. Troubleshooting in a large environment becomes more difficult because of the multiple servers, scopes, and forwarding addresses to be examined.
Downtime introduced through server upgrades: Any hardware upgrades performed on servers hosting split scopes will make the DHCP server unavailable for client requests. Clustered DHCP services do not have this drawback, as a rolling upgrade can be performed.