CIDR and subnet masks explained

A prefix is an address boundary

In 192.168.10.77/26, the /26 means that the first 26 bits identify the IPv4 network. Six bits remain, giving 64 addresses in the block. The dotted-decimal mask is another representation of the same boundary: 255.255.255.192. A larger prefix number gives a smaller address block. CIDR is described in RFC 4632.

Think of the calculation as identifying the block containing an address. It does not discover the configuration of a router. If you enter the wrong prefix, the arithmetic can be internally correct while describing the wrong network.

Worked example: a /26 office segment

Use 192.168.10.77 with prefix /26 in the calculator. The last-octet block boundaries are 0, 64, 128, and 192. Because 77 lies between 64 and 127, the network is 192.168.10.64/26.

For 50 planned devices, 62 conventional addresses leave 12 before your own reservations. Account for gateways, infrastructure, growth, and platform-specific reserved addresses before deciding that the block is large enough. A cloud provider may reserve more addresses than this generic calculation.

Why /31 and /32 need different interpretation

The familiar “subtract two” rule describes conventional IPv4 subnets; it is not universal. A /31 has two endpoint addresses on supported point-to-point links, as specified in RFC 3021. A /32 describes a single address or host route. Neither should be treated as an ordinary multi-host LAN with a broadcast address.

The calculator labels these special cases explicitly. Before applying /31 to equipment, confirm that both ends and the intended interface type support it. Numerical capacity alone is not an interoperability test.

IPv6 capacity is a different calculation

IPv6 has 128 bits. A /64 therefore contains 2^64 addresses, and IPv6 does not use IPv4-style broadcast. For example, 2001:db8:abcd:1234::42/64 normalizes to 2001:db8:abcd:1234::/64. The documentation prefix in this example is for explanation, not deployment. See RFC 4291.

The current tool accepts hexadecimal IPv6 notation, including compressed groups. It does not accept IPv4-embedded dotted notation as an IPv6 input. Use an equivalent hexadecimal representation. A displayed capacity is not a recommendation to allocate every address or use a particular prefix on every interface.

Check a plan before changing a network

1. Write down the actual address and prefix from the intended interface configuration.
2. Calculate the normalized network and its boundaries.
3. Compare the entire range with neighboring subnets and VPN routes; different starting addresses can still overlap.
4. Verify DHCP scope, gateway, static reservations, routing, and firewall assumptions separately.
5. Keep the previous configuration available and test connectivity after an approved change.

Two addresses in the same calculated subnet do not prove that their devices can communicate. VLAN isolation, routing, host firewalls, and duplicate addresses can still prevent traffic. Use the calculator to check the addressing premise, then investigate the next layer.

Sources and review

• CIDR
• /31 point-to-point links
• IPv6 addressing

Examples are illustrative. Reviewed September 29, 2026 against the listed references and the related tool behavior. Methodology · Report a correction.