In a previous post, I had discussed How to use Split-Scopes in DHCP for a Robust Deployment. Some of you wrote back asking why is a 80:20 split better than a 60:40 or 50:50 split. Let me try to answer how the ratio of the split should be determined.
An 80:20 split of the available address range between the primary and the secondary DHCP servers is most commonly used, but of course you can use any ratio appropriate to your deployment.
A good rule of thumb for determining the ratio is (0.5*Lease Time for the Subnet) : (Amount of time it will take you to restore a server). For instance, if the address lease time on your DHCP server is 8 days, then the clients will renew their lease every (0.5 * 8 = 4) days. Say, it will take you max 1 day to restore a server in case it is down. Then the appropriate ratio would be 4:1 or 80:20. You can vary this based on your requirements/deployment.
Ideally, of course, if you have a lot of free address space available (especially if you are using one of the private address ranges specified by RFC 1918) then you can forget about the above rule and use a 50:50 split. Note that in this case the max number of clients on the network should correspond to around 50% of the available address range. So if you are expecting around 250 clients, you should use a /23 address range for the subnet.
Hope this helps you fine-tune your DHCP split-scope deployment.
Windows Enterprise Networking
[This posting is provided "AS IS" with no warranties, and confers no rights.]
"Ideally, of course" would be that the DHCP service finally implement draft-ietf-dhc-failover / RFC3074 to eliminate the need for this kind of silly scope splitting. Is there any hope for this, or will Microsoft and their customers still be spending time documenting and discussing scope split ratios until v6 deployment is pervasive?
I agree. When will RFC3074 be implemented in Microsoft DHCP?