Option to exclude .0 and .255 addresses automatically from IPv4 subnets larger than /24
This is a feature request.
Having encountered twice recently (one using Kea and once using ISC DHCP) a scenario where a large IPv4 subnet had its pools defined as many small ranges, specifically to avoid the .1 and .255 addresses, I wondered if it might be a good idea to have a configuration option to make this happen automatically.
The reason for excluding these addresses is that some broken clients don't handle them well, even though they should, if the netmask is shorter than 255.255.255.0 - thus leaving operators in a quandary - particularly in environments where they either don't have control over which clients are allowed addresses on their network, or where they have known broken equipment that cannot be assigned fixed addresses or easily upgraded/replaced.
The reason for suggesting it as an option is to ask whether if would be more efficient operationally to run with one large pool (and a logic switch that excludes these addresses from the pool) versus many small pools (the usual configuration workaround).
It would also make configuration much simpler for those who need this for their subnet definitions.
I'm asking specifically for ISC DHCP, but this could also be a feature to consider for Kea DHCP.
The potential performance improvements may be too negligible for this to be worthwhile to implement, but combined with the simplification of configuration, could a reasonable case be made to implement something like this.
What do other DHCP server implementations do?