IPv4 subnetting reference

In the IPv4 address space certain address blocks are specially allocated or reserved for special uses such as loopback interfaces, private networks (RFC 1918), and state-less autoconfiguration (Zeroconf, RFC 3927) of interfaces. Such addresses may be used without registration or allocation from Regional Internet Registries (RIRs). However, these address ranges must not be routed into the public Internet infrastructure.

The netmask is a bitmask that can be used to separate the bits of the network identifier from the bits of the host identifier. It is often written in the same notation used to denote IP addresses.

Not all sizes of prefix announcement may be routable on the public Internet: see routing, peering.

Class	Leading bits	Start	End	Subnet Mask in dotted decimal and CIDR notation
A	0	0.0.0.0	127.255.255.255	255.0.0.0 or /8
B	10	128.0.0.0	191.255.255.255	255.255.0.0 or /16
C	110	192.0.0.0	223.255.255.255	255.255.255.0 or /24
D	1110	224.0.0.0	239.255.255.255	n/a
E	1111	240.0.0.0	255.255.255.255	n/a

The blocks numerically at the start and end of classes A, B and C were originally reserved for special addressing or future features, i.e., 0.0.0.0/8 and 127.0.0.0/8 are reserved in former class A; 128.0.0.0/16 and 191.255.0.0/16 were reserved in former class B but are now available for assignment; 192.0.0.0/24 and 223.255.255.0/24 are reserved in former class C.

While the 127.0.0.0/8 network is a Class A network, it is designated for loopback and cannot be assigned to a network.

Also see RFC 6890 (Special-Purpose IP Address Registries).

CIDR	Host bits	Netmask	Addresses in subnet	Classful name	Typical usage
/8	24	255.0.0.0	16777216 = 224	Class A (see this list)	Largest block allocation made by IANA
/9	23	255.128.0.0	8388608 = 223
/10	22	255.192.0.0	4194304 = 222
/11	21	255.224.0.0	2097152 = 221
/12	20	255.240.0.0	1048576 = 220
/13	19	255.248.0.0	524288 = 219
/14	18	255.252.0.0	262144 = 218
/15	17	255.254.0.0	131072 = 217
/16	16	255.255.0.0	65536 = 216	Class B
/17	15	255.255.128.0	32768 = 215		ISP / large business
/18	14	255.255.192.0	16384 = 214		ISP / large business
/19	13	255.255.224.0	8192 = 213		ISP / large business
/20	12	255.255.240.0	4096 = 212		Small ISP / large business
/21	11	255.255.248.0	2048 = 211		Small ISP / large business
/22	10	255.255.252.0	1024 = 210
/23	9	255.255.254.0	512 = 29
/24	8	255.255.255.0	256 = 28	Class C	Large LAN
/25	7	255.255.255.128	128 = 27		Large LAN
/26	6	255.255.255.192	64 = 26		Small LAN
/27	5	255.255.255.224	32 = 25		Small LAN
/28	4	255.255.255.240	16 = 24		Small LAN
/29	3	255.255.255.248	8 = 23		Smallest multi-host network
/30	2	255.255.255.252	4 = 22		Point to point links
/31	1	255.255.255.254	2 = 21		Point to point links (RFC 3021)
/32	0	255.255.255.255	1 = 20		Host route

In networks that follow RFC 919, the all-zeroes network address identifies the network, while the all-ones address is used as an all-hosts broadcast address. The remaining addresses within the block are available for hosts. RFC 919 states that, "as a notational convention, we refer to networks (as opposed to hosts) by using addresses with zero fields. For example, 36.0.0.0 means "network number 36" while 36.255.255.255 means "all hosts on network number 36".

RFC 3021 (published December 2000) allows an exception which makes it possible to use /31 networks for point-to-point links, saving two addresses which would be unused in a /30 point-to-point link. RFC 3021 allows "a 1-bit wide host-number field", and requires that "[in] a point-to-point link with a 31-bit subnet mask, the two addresses above MUST be interpreted as host addresses."

See also

• Subnetwork
• Classless Inter-Domain Routing (CIDR)
• IPv6 subnetting reference