IPv4 addresses fall into two main categories: public addresses, directly accessible on the Internet, and private addresses, intended for internal use within a local network.

A public IPv4 address is globally unique and routable across the Internet. It is assigned by official authorities and required for public-facing services such as websites, email servers, or cloud infrastructure. The worldwide uniqueness of these addresses is essential to avoid any routing conflicts or collisions.

The IANA (Internet Assigned Numbers Authority), operating under the ICANN (Internet Corporation for Assigned Names and Numbers), manages the distribution of these IP ranges. In concrete terms, IANA divides the IPv4 space into 256 blocks of size /8, according to CIDR notation. Each block represents just over 16.7 million addresses (2³² / 2⁸).

These unicast address blocks are entrusted by IANA to the Regional Internet Registries (RIRs). These RIRs are responsible for redistributing the addresses at regional level, according to the real needs of access providers, companies or administrations. The unicast address space extends from blocks 1/8 to 223/8, with portions either reserved for special uses (research, documentation, testing), or allocated directly to a network or RIR for redistribution.

To check who owns a public IP address, you can consult the RIR databases using the whois command, or by using the web interfaces provided by each registry. These tools can be used to trace the address back to the organization or provider that declared it.

Conversely, there are private IPv4 addresses, a practical response to the shortage of public addresses. These addresses, which are not routable on the Internet, are reserved for local environments: corporate networks, home LANs, datacenters or computing clusters. They are not unique worldwide: many private networks can reuse the same IP ranges without interference, as long as they remain isolated or use a network address translation device to access the internet.

To allow a device with a private IP address to access the Internet, networks use NAT (Network Address Translation). NAT works by dynamically replacing the private address with a public one, enabling dozens (or even hundreds) of devices to share a single public IP address. This method optimizes the use of IPv4 space and also adds a layer of security by hiding the internal network structure.

Another special category is unspecified addresses. The IPv4 notation 0.0.0.0 or its IPv6 version ::/128 means "no specific address". Such an address is invalid as a network address destination, but it can be used locally by a host to indicate "all interfaces" or "no address assigned yet". This is common in DHCP dynamic assignment or for listening on all server interfaces.

IPv6 also supports private addressing, but the standard generally recommends public addressing to avoid stacking multiple NAT layers. The site-local addresses (site-local) of the fec0::/10 block were deprecated by RFC 3879 for consistency and security reasons. They were replaced with Unique Local Addresses (ULA) located in the fc00::/7 block. ULAs allow the creation of private IPv6 networks with clean internal routing, using a randomly generated 40-bit identifier to ensure local uniqueness.

IPv4 exhaustion was officially confirmed in 2011. To extend its lifespan, the Internet community adopted several strategies:

These measures show that IP addressing is not just a technical challenge, but also a matter of global governance, central to the Internet's ongoing expansion.