A /31 has two addresses: why subtracting two does not always work
The size of an address block and the rules for using it are different questions. Very small blocks make the distinction particularly important.
IPv4 subnet calculator →Calculate the total without assigning roles
A /31 leaves one bit and therefore contains two addresses. A /32 leaves zero bits and contains one address. The calculator’s total comes directly from 2 raised to the power of 32 minus the prefix. That arithmetic does not decide which addresses a device or platform can use.
Interpret the block according to its purpose
RFC 3021 describes using /31 prefixes on IPv4 point-to-point links, where the two addresses can serve as endpoints. A /32 commonly identifies a single address in routing. The tool’s traditional broadcast label describes the block’s highest-address calculation; it should not be used to impose ordinary broadcast-subnet reservations on every /31 use case.
Check the actual platform before deployment
Equipment, providers and cloud services can impose their own supported configurations and reservations. This tool displays mathematical boundaries and does not verify that an interface supports /31 addressing or permits a particular allocation. Keep the prefix, link type and platform requirements with the plan. The bandwidth and unit-conversion tools may assist with separate capacity planning, but they do not resolve address-assignment policy. Avoid reporting a generic total-minus-two result as a universally usable-host count.
Try the calculators
- IPv4 subnet calculator
Find network, broadcast and subnet mask.
- Download time calculator
Estimate download time from file size, connection rate and usable bandwidth.
- Unit converter
Convert length, mass, area, volume, speed and data size.