InInternet networking, aprivate network is acomputer network that uses a privateaddress space ofIP addresses. These addresses are commonly used forlocal area networks (LANs) in residential, office, and enterprise environments. Both theIPv4 and theIPv6 specifications defineprivate IP address ranges.[1][2]
MostInternet service providers (ISPs) allocate only a single publiclyroutable IPv4 address to each residential customer, but many homes have more than onecomputer,smartphone, or other Internet-connected device. In this situation, anetwork address translator (NAT/PAT) gateway is usually used to provide Internet connectivity to multiple hosts. Private addresses are also commonly used incorporate networks which, for security reasons, are not connected directly to theInternet. Often aproxy,SOCKS gateway, or similar devices are used to provide restricted Internet access to network-internal users.
Private network addresses are not allocated to any specific organization. Anyone may use these addresses without approval fromregional or local Internet registries.Private IP address spaces were originally defined to assist in delayingIPv4 address exhaustion.IP packets originating from or addressed to a private IP address cannot be routed through the public Internet.
Private addresses are often seen as enhancingnetwork security for the internal network since use of private addresses internally makes it difficult for an external host to initiate a connection to an internal system.
TheInternet Engineering Task Force (IETF) has directed theInternet Assigned Numbers Authority (IANA) toreserve the following IPv4 address ranges for private networks:[1]: 4
| RFC 1918 name | IP address range | Number of addresses | LargestCIDR block (subnet mask) | Host ID size | Mask bits | Classful description[Note 1] |
|---|---|---|---|---|---|---|
| 24-bit block | 10.0.0.0 – 10.255.255.255 | 16777216 | 10.0.0.0/8 (255.0.0.0) | 24 bits | 8 bits | single class A network |
| 20-bit block | 172.16.0.0 – 172.31.255.255 | 1048576 | 172.16.0.0/12 (255.240.0.0) | 20 bits | 12 bits | 16 contiguous class B networks |
| 16-bit block | 192.168.0.0 – 192.168.255.255 | 65536 | 192.168.0.0/16 (255.255.0.0) | 16 bits | 16 bits | 256 contiguous class C networks |
In practice, it is common to subdivide these ranges into smallersubnets.
In April 2012, IANA allocated the100.64.0.0/10 block of IPv4 addresses specifically for use incarrier-grade NAT scenarios.[4]
| IP address range | Number of addresses | LargestCIDR block (subnet mask) | Host ID size | Mask bits |
|---|---|---|---|---|
| 100.64.0.0 – 100.127.255.255 | 4194304 | 100.64.0.0/10 (255.192.0.0) | 22 bits | 10 bits |
This address block should not be used on private networks or on the public Internet. The size of the address block was selected to be large enough to uniquely number all customer access devices for all of a single operator'spoints of presence in a large metropolitan area such asTokyo.[4]
The concept of private networks has been extended in the next generation of theInternet Protocol,IPv6, and special address blocks are reserved.
The address blockfc00::/7 is reserved by IANA for unique local addresses (ULAs).[2] They areunicast addresses, but contain a 40-bit random number in the routing prefix to prevent collisions when two private networks are interconnected. Despite being inherentlylocal in usage, theIPv6 address scope of unique local addresses is global.
The first block defined isfd00::/8, designed for/48 routing blocks, in which users can create multiple subnets, as needed.
| RFC 4193 Block | Prefix/L | Global ID (random) | Subnet ID | Number of addresses in subnet |
|---|---|---|---|---|
| 48 bits | 16 bits | 64 bits | ||
| fd00::/8 | fd | xx:xxxx:xxxx | yyyy | 18446744073709551616 |
Examples:
| Prefix/L | Global ID (random) | Subnet ID | Interface ID | Address | Subnet |
|---|---|---|---|---|---|
| fd | xx:xxxx:xxxx | yyyy | zzzz:zzzz:zzzz:zzzz | fdxx:xxxx:xxxx:yyyy:zzzz:zzzz:zzzz:zzzz | fdxx:xxxx:xxxx:yyyy::/64 |
| fd | 12:3456:789a | 0001 | 0000:0000:0000:0001 | fd12:3456:789a:1::1 | fd12:3456:789a:1::/64 |
A former standard proposed the use ofsite-local addresses in thefec0::/10 block, but because of scalability concerns and poor definition of what constitutes asite, its use has been deprecated since September 2004.[5]
Another type of private networking uses the link-local address range. The validity of link-local addresses is limited to a single link; e.g. to all computers connected to aswitch, or to onewireless network. Hosts on different sides of anetwork bridge are also on the same link, whereas hosts on different sides of anetwork router are on different links.
InIPv4, the utility of link-local addresses is inzero-configuration networking whenDynamic Host Configuration Protocol (DHCP) services are not available and manual configuration by a network administrator is not desirable. The block169.254.0.0/16 was allocated for this purpose.[6][7] If a host on an IEEE 802 (Ethernet) network cannot obtain a network address via DHCP, an address from169.254.1.0 to169.254.254.255[Note 2] may be assignedpseudorandomly. The standard prescribes that address collisions must be handled gracefully.
InIPv6, the blockfe80::/10 is reserved for IP address autoconfiguration.[8]The implementation of these link-local addresses is mandatory, as various functions of the IPv6 protocol depend on them.[9]
A special case of private link-local addresses is theloopback interface. These addresses are private and link-local by definition since packets never leave the host device.
IPv4 reserves the entire class A address block127.0.0.0/8 for use as private loopback addresses. IPv6 reserves the single address::1.
Some are advocating reducing127.0.0.0/8 to127.0.0.0/16.[10]
It is common for packets originating in private address spaces to be misrouted onto the Internet. Private networks often do not properly configureDNS services for addresses used internally and attemptreverse DNS lookups for these addresses, causing extra traffic to the Internetroot nameservers. TheAS112 project attempted to mitigate this load by providing specialblack holeanycast nameservers for private address ranges which only return negative result codes (not found) for these queries.
Organizational edge routers are usually configured to drop ingress IP traffic for these networks, which can occur either by misconfiguration or from malicious traffic using a spoofed source address. Less commonly, ISP edge routers drop such egress traffic from customers, which reduces the impact to the Internet of such misconfigured or malicious hosts on the customer's network.
Since the private IPv4 address space is relatively small, many private IPv4 networks unavoidably use the same address ranges. This can create a problem when merging such networks, as some addresses may be duplicated for multiple devices. In this case, networks or hosts must be renumbered, often a time-consuming task or a network address translator must be placed between the networks to translate or masquerade one of the address ranges.
IPv6 definesunique local addresses,[2] providing a very large private address space from which each organization can randomly or pseudo-randomly allocate a 40-bit prefix, each of which allows 65536 organizational subnets. With space for about one trillion (1012) prefixes, it is unlikely that two network prefixes in use by different organizations would be the same, provided each of them was selected randomly, as specified in the standard. When two such private IPv6 networks are connected or merged, the risk of an address conflict is therefore virtually absent.