IP Subnet Calculator

What a subnet calculator hands back

An IPv4 address is 32 bits, and a subnet calculator splits those bits in two. The left part names the network and the right part names the host. Every answer on this page comes from that one split: the network address, the broadcast address, the subnet mask, the wildcard mask, the first and last usable host, and the count of hosts in between.

The split is written as a slash and a number, such as 192.168.10.37/26. The number after the slash is the prefix length, and it counts how many of the 32 bits belong to the network. A /26 leaves 6 bits for hosts. Two of those host patterns are reserved, one for the network itself and one for the broadcast, which is why the usable count is lower than the size of the block.

One habit makes the arithmetic easy. Work in the octet where the prefix ends, because the octets before it are fixed and the octets after it are entirely host bits.

CIDR: the prefix is the mask

Classless Inter-Domain Routing replaced the old class A, B and C system, and RFC 4632 (August 2006) is the current specification. It defines the /n notation and the rule that any prefix length from 0 to 32 is legal, so a network can be any size that is a power of two.

The mask is just the prefix written out in full. A /26 means 26 ones followed by 6 zeros, which is 255.255.255.192, because the last octet holds 128 + 64 = 192. The wildcard mask is the bitwise opposite, 0.0.0.63, and the two always add to 255 in every octet.

The host count follows from the number of host bits h = 32 - p. The block holds 2 to the power of h addresses, and 2 to the power of h minus 2 of them can be assigned to devices. A /24 has 8 host bits, so 256 addresses and 254 usable ones. A /28 has 4 host bits, so 16 addresses and 14 usable ones.

Every step of one prefix halves the block. That is why the masks move in the sequence 128, 192, 224, 240, 248, 252, 254, 255, and why a /29 is half the size of a /28.

A worked example: 192.168.10.37/26

The prefix 26 sits inside the fourth octet, since 24 bits cover the first three octets and 2 more land in the last one. Those 2 network bits leave 6 host bits, so the block size is 2 to the 6th, which is 64. The last octet therefore breaks into blocks that start at 0, 64, 128 and 192.

The address ends in 37, and 37 falls between 0 and 63, so the network address is 192.168.10.0 and the broadcast address is 192.168.10.63. The mask is 255.255.255.192 and the wildcard is 0.0.0.63. The host range runs from 192.168.10.1 to 192.168.10.62, which is 62 addresses.

Check the count two ways. The block holds 64 addresses, and 64 - 2 = 62 are usable. Or count the host bits: 2 to the 6th is 64, and removing the network and broadcast patterns leaves 62. Both routes land on the same number, and the calculator prints that number directly.

Notice what did not change. The first three octets stayed put because all 24 of their bits are network bits. Only the octet that holds the split moves, which is why a quick mental check on a /26 only needs the last number.

Why two addresses disappear, and the /31 and /32 cases

The lowest address in a block has all host bits set to zero, and it names the subnet itself, so a router uses it in its table rather than on an interface. The highest address has all host bits set to one, and it is the directed broadcast that reaches every host in the block. Neither can belong to a single device, so a /24 with 256 addresses serves 254 hosts.

Point to point links waste half of a /30, and RFC 3021 (December 2000) fixed that. On a link with exactly two ends, a /31 is legal and both addresses are usable, one at each end. The catch is that the block has no directed broadcast address, so broadcast traffic on such a link uses 255.255.255.255 instead.

A /30 shows the waste in numbers. It holds four addresses and only two are usable, so a link between two routers burns four addresses to connect two interfaces. Swap in a /31 and the same link uses two, which halves the space that segment needs, and RFC 3021 notes that vendors had already tested the change with OSPF, IS-IS, BGP and EIGRP before it was published.

A /32 names a single host. It shows up as a host route in a routing table and as a loopback alias on a router, and it has no network or broadcast address at all, because there is only one pattern in the block.

Private ranges and the class history

RFC 1918 (February 1996) set aside three blocks for private use: 10.0.0.0/8 with 16,777,216 addresses, 172.16.0.0/12 with 1,048,576, and 192.168.0.0/16 with 65,536. Together they hold 16,777,216 + 1,048,576 + 65,536 = 17,891,328 addresses, and none of them is routed on the public internet, so a home or office network needs network address translation to reach the outside.

The three blocks line up like this, and the host bits column shows why the middle block is not a plain multiple of the other two:

The 172 block carries 20 host bits rather than 16 or 24, which is why its size falls between the other two. Its first octet is fixed at 172 and its second octet runs from 16 to 31, so the block covers 16 values of 65,536 addresses each.

The block sizes are worth memorizing because they explain the odd middle range. The 172 block starts at 172.16 and stops at 172.31, which is 16 values in the second octet, and 16 x 65,536 = 1,048,576. The 192.168 block is a single second octet value, which is why it is the smallest of the three and the one most home routers use.

Before CIDR, the first octet decided the class: 1 to 126 was class A, 128 to 191 class B, and 192 to 223 class C. Class D covered multicast from 224 to 239 and class E reserved 240 to 255. The fixed 8, 16 and 24 bit splits wasted space, and RFC 4632 retired them in favor of any prefix length.

Other blocks carry special meanings, and the IANA IPv4 special purpose registry lists all of them and was last updated in 2025. 127.0.0.0/8 is loopback, 169.254.0.0/16 is link local, 100.64.0.0/10 is shared space used by carriers, 198.18.0.0/15 is for benchmarking, and 192.0.2.0/24, 198.51.100.0/24 and 203.0.113.0/24 are documentation ranges that never appear on a live network.

Planning capacity across subnets

Capacity planning runs the host formula backward. Decide how many devices a segment needs, add room for growth, then pick the smallest block whose usable count covers it. A site with 100 devices needs a /25, since 126 is the first usable count above 100. A site with 50 needs a /26 at 62, and one with 10 needs a /28 at 14.

Split a /24 for five sites that need 50, 25, 10, 6 and 2 hosts. The sizes are /26 at 62 usable, /27 at 30, /28 at 14, /29 at 6 and /30 at 2. The blocks take 64, 32, 16, 8 and 4 addresses, so 64 + 32 + 16 + 8 + 4 = 124 of the 256 are in use and 132 are left. That remainder is enough for one more /25 of 128 addresses plus 4 spare.

The alignment rule explains the addresses in that plan. 192.168.10.64/27 starts on 64 because a /27 has a block size of 32, and 192.168.10.96/28 starts on 96 because a /28 has a block size of 16. A block placed off its own boundary overlaps the one next to it.

Two rules keep the plan clean. Assign the largest block first, because a /26 has to start on a multiple of 64 and a /27 on a multiple of 32, and starting with the big ones avoids the gaps that appear when a small block lands in the middle of the space. Then keep each block on its own boundary: 192.168.10.0/26, 192.168.10.64/27, 192.168.10.96/28, 192.168.10.112/29 and 192.168.10.120/30 fit end to end with no overlap.

Laid out largest block first, that five site plan reads as follows:

The usable column adds to 114 addresses for 93 devices, which is the 21 slots of headroom the plan carries. Site D is the tight one: a /29 holds 6 usable addresses and the site needs 6, so any growth there forces a move to a /28.

Mask, wildcard and where each one appears

The wildcard mask is the mask flipped bit by bit, so 255.255.255.192 becomes 0.0.0.63. Router access lists use the wildcard form because it reads as a pattern: a zero means the bit must match and a one means it can be anything. An entry for 192.168.10.0 with wildcard 0.0.0.63 matches every address in that /26 and nothing outside it.

Interface configuration uses the plain mask, so a /26 is entered as 255.255.255.192. OSPF and EIGRP use the wildcard form in their network statements, which is why the two numbers are worth keeping straight. The pair always sums to 255.255.255.255, so one can be derived from the other in a single step.

IPv6 drops the mask entirely and writes only the prefix, and RFC 4291 (2006) sets the addressing architecture that defines it. A /64 leaves 64 bits for the interface identifier, which is why almost every IPv6 LAN uses that length. The same arithmetic applies, with 128 bits in place of 32, and the calculator on this page stays with IPv4.