192.168.1.0/24. The slash and number tell you the size and shape of the subnet — how many addresses it holds, where the network ends and hosts begin, and what mask to plug into a router. One number, all the math.
The slash is a bit count
IPv4 addresses are 32 bits long. The number after the slash — the prefix — tells you how many of those bits identify the network. Whatever’s left identifies individual hosts. So /24 = 24 network bits + 8 host bits. /16= 16 + 16. /30 = 30 + 2.
Host count is 2^(32 − prefix)
- /8 — 16,777,216 total / 16,777,214 usable (Class A size)
- /16 — 65,536 / 65,534 (Class B)
- /24 — 256 / 254 (typical home network)
- /27 — 32 / 30
- /28 — 16 / 14
- /29 — 8 / 6
- /30 — 4 / 2 (point-to-point links)
- /31 — 2 / 2 (RFC 3021 point-to-point, no broadcast)
- /32 — 1 / 1 (a single host)
Subtract two from total because the first address is the network ID and the last is the broadcast. Both are reserved; you can’t assign either to a host.
The mask is the same info, longer
A subnet mask like 255.255.255.0 is just /24 written as four dotted octets. The 1-bits cover the network portion; the 0-bits cover the host portion. Most modern tools accept either notation. CIDR is shorter, so newer docs prefer it.
Network address, broadcast, mask, wildcard, and host range — with a binary view that shows exactly which bits are which.
Why CIDR replaced classes
Before 1993, IPv4 was “classful” — every address fell into Class A (/8), Class B (/16), or Class C (/24). If your org needed 500 hosts, your only choice was a full Class B with 65,000 addresses, wasting 64,500 of them. CIDR let registries hand out arbitrary prefixes (“here’s a /22, that’s 1,022 hosts”) which slowed the IPv4 exhaustion crisis by years.
Reading a CIDR fast
Two shortcuts: (1) the smaller the slash number, the bigger the subnet (yes, backwards from intuition); (2) every step up in prefix halves the size — /24 has 256, /25 has 128, /26 has 64, and so on down to /32 (just one).