server1: scope: 192.168.1.10 -> 192.168.1.50
server2: scope: 192.168.1.51 -> 192.168.1.90
All the advice your getting is good, so my .02
From what you have shown, this is how I build my scopes.
Server 1:
Scope 1 - 192.168.1.1 -> 192.168.1.254
Ip Exclusions - 192.168.1.1 -> 192.168.1.9; 192.168.1.51 -> 192.168.1.254
Server 2:
Scope 2 - 192.168.1.1 -> 192.168.1.254
IP Exclusions - 192.168.1.1 -> 192.168.1.50; 192.168.1.91 -> 192.168.1.254
The reason I do my scopes this way is that using exclusion ranges allows for more flexibility with scope changes.
The 80/20 split rule is a Microsoft rule and should work fine. The only time I've seen issues are with short lease times like one-day. The standard 80/20 rule and the default of lease of 8 days is plenty.
I understand spirit's thinking regarding scope redundancy but I just disagree with the logic. A lease time of 8 days would allow plenty of time to restore and in worse case rebuild a DHCP server. Clients will maintain their IP Address while resolving the issue, at least maintain them until 100% of the lease expires.
Just to keep the terminology from getting confusing, DHCP is not Load Balanced. You can cluster DHCP if high avail is required.