Understanding IPv6 address types is fundamental to modern network configuration and troubleshooting. Unlike IPv4's simpler address classification, IPv6 introduces a more sophisticated addressing architecture designed to support the massive scale and diverse requirements of the modern Internet. This article explores the main types of IPv6 addresses, their use cases, and how to identify them.
IPv6 defines three primary address types that replace IPv4's addressing model:
Notably absent from this list is broadcast addressing. IPv6 eliminates broadcast addresses entirely, relying instead on multicast to handle group communication more efficiently. This architectural change reduces network noise and improves overall performance.
Unicast addresses identify a single interface on an IPv6-enabled device. Packets sent to a unicast address are delivered to exactly one recipient. Within the unicast category, IPv6 defines several important subtypes:
Global unicast addresses are the IPv6 equivalent of public IPv4 addresses. These addresses are globally routable on the IPv6 Internet and must be unique worldwide.
Characteristics:
Example: 2001:0db8:85a3:0000:0000:8a2e:0370:7334
Use Cases:
When you visit test-ipv6.run to test your IPv6 connectivity, your device's global unicast address is what gets detected and displayed.
Link-local addresses are automatically configured on all IPv6-enabled interfaces and are used exclusively for communication on a single network segment (link).
Characteristics:
Example: fe80::1ff:fe23:4567:890a
Use Cases:
Link-local addresses are critical for IPv6 network operation. Even if a device has no global unicast address, it will always have a link-local address for local network communication.
Unique local addresses (ULA) are the IPv6 equivalent of IPv4 private addresses (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16).
Characteristics:
Example: fd12:3456:789a:1::1
Use Cases:
While ULAs provide privacy similar to IPv4 NAT, IPv6's design philosophy encourages using global unicast addresses even for internal devices, with firewalls providing security rather than address translation.
Multicast addresses enable efficient one-to-many communication. A single packet sent to a multicast address is delivered to all interfaces that have joined that multicast group.
Characteristics:
Example: ff02::1 (all nodes on the local link)
The 4-bit scope field defines the range for multicast forwarding:
| Scope Value | Scope Name | Description |
|---|---|---|
| 1 | Interface-local | Single interface only |
| 2 | Link-local | Single network link |
| 5 | Site-local | Single site or organization |
| 8 | Organization-local | Multiple sites in an organization |
| E | Global | Entire Internet |
IPv6 defines several important multicast addresses:
ff02::1 - All nodes on the local linkff02::2 - All routers on the local linkff02::1:2 - All DHCP servers and relay agents on the local linkff02::1:ff00:0/104 - Solicited-node multicast addressesSolicited-node multicast addresses are automatically created by mapping the last 24 bits of a device's unicast address to the prefix ff02:0:0:0:0:1:ff00::/104. This special multicast type is used for efficient neighbor discovery.
Use Cases:
Anycast addresses are assigned to multiple interfaces (typically on different devices), with packets routed to the nearest device according to routing protocol metrics.
Characteristics:
Example: 2001:db8::1 assigned to multiple servers in different locations
Use Cases:
The most prominent use of anycast is for DNS infrastructure. All Internet root nameservers are implemented as clusters of hosts using anycast addressing, improving both performance and resilience.
CDNs use anycast to route users to the geographically nearest distribution center, reducing latency and improving user experience.
Enterprise networks deploy anycast for critical services like:
Anycast addresses facilitate transition mechanisms like 6to4, with the well-known anycast address 192.88.99.1 (IPv4) serving as a default gateway.
IPv6 reserves several addresses for specific purposes:
The loopback address allows a device to send packets to itself for testing and inter-process communication.
::1 (or 0:0:0:0:0:0:0:1)The unspecified address indicates the absence of an address.
:: (or 0:0:0:0:0:0:0:0)Reserved for use in documentation and examples, similar to IPv4's 192.0.2.0/24.
| Address Type | Prefix | Example | Scope |
|---|---|---|---|
| Global Unicast | 2000::/3 | 2001:db8::1 | Global |
| Link-Local Unicast | FE80::/10 | fe80::1 | Link-local |
| Unique Local Unicast | FC00::/7 | fd00::1 | Private |
| Multicast | FF00::/8 | ff02::1 | Variable |
| Anycast | 2000::/3 | 2001:db8::1 | Global |
| Loopback | ::1/128 | ::1 | Host |
| Unspecified | ::/128 | :: | N/A |
Understanding these address types becomes practical when testing your network's IPv6 capabilities. Tools like test-ipv6.run can help you verify:
The test distinguishes between different connectivity scenarios, including the problematic "broken IPv6" state where IPv6 is configured but timing out—a situation that's impossible to diagnose without understanding these address types.
IPv6's address architecture represents a significant evolution from IPv4, with distinct address types optimized for specific use cases. The three primary types—unicast, multicast, and anycast—each serve unique purposes in modern networks:
As IPv6 adoption continues to grow (approaching 44% globally in 2025), understanding these address types becomes increasingly important for network administrators, developers, and anyone troubleshooting connectivity issues. The elimination of broadcast, the automatic configuration of link-local addresses, and the elegant simplicity of multicast and anycast all contribute to IPv6's superior design for the modern Internet.
Whether you're configuring enterprise infrastructure, developing networked applications, or simply curious about your connection's IPv6 readiness, recognizing these address types empowers you to make informed decisions and diagnose issues effectively.