CCNA online course

IPv6 Address Types for CCNA

Learn IPv6 address types, prefixes, scopes, multicast, anycast, address assignment, and Cisco IOS verification commands for CCNA.

IPv6 uses 128-bit addresses to identify interfaces and deliver packets. For CCNA, you must recognize address formats, classify addresses by prefix, understand their scope, and know how Cisco IOS displays and uses them.

This lesson assumes familiarity with IPv4 addressing, hexadecimal numbers, CIDR prefix lengths, Layer 2 forwarding, and basic Cisco IOS interface configuration.

IPv6 address fundamentals

128-bit notation and hextets

An IPv6 address contains 128 bits. It is normally written as eight groups of four hexadecimal digits. Each group is a hextet, meaning a 16-bit hexadecimal segment.

2001:0DB8:0001:000A:0000:0000:0000:0025

The eight hextets are separated by colons. Since each hexadecimal digit represents four bits, four hexadecimal digits represent 16 bits, and eight hextets represent 128 bits.

IPv6 abbreviation rules

IPv6 notation can be shortened using two rules:

1. Remove leading zeros from any hextet. A hextet of 000A becomes A, and 0025 becomes 25. Do not remove zeros from the middle or end of a hextet.
2. Replace one consecutive sequence of all-zero hextets with ::. This compression can be used only once in an address.

2001:0DB8:0001:000A:0000:0000:0000:0025
2001:DB8:1:A:0:0:0:25
2001:DB8:1:A::25

The address 2001:DB8:0:0:1:0:0:1 may become 2001:DB8::1:0:0:1 or 2001:DB8:0:0:1::1, depending on which zero sequence is compressed. A single zero hextet may also be written as 0, but :: represents one or more complete zero hextets and must not be used twice.

Prefix length, address, and scope

A prefix length identifies how many leftmost bits belong to the network prefix. For example, /64 means the first 64 bits are the prefix and the remaining 64 bits are available for the interface identifier.

2001:DB8:100:10::25/64
Prefix:              2001:DB8:100:10::/64
Interface identifier:                 ::25

• An IPv6 address is the complete 128-bit value assigned to or used by an interface.
• An IPv6 prefix is the network portion identified by a prefix length, such as 2001:DB8:100:10::/64.
• An interface identifier is the interface or host portion. In a typical /64 subnet, it is the final 64 bits.
• Scope describes where an address is valid or reachable, such as on one local link, inside an organization, or across the public Internet.

Prefix and scope are related but not identical. A prefix indicates address purpose and structure; scope describes the area in which traffic can be delivered.

IPv6 delivery types

The primary IPv6 delivery types are unicast, multicast, and anycast.

• Unicast is one-to-one delivery to a single interface.
• Multicast is one-to-many delivery to interfaces that joined a multicast group.
• Anycast uses one address on multiple interfaces and delivers traffic to the topologically nearest reachable instance according to routing.

IPv6 has no broadcast address type. IPv6 uses multicast groups for functions that commonly used broadcast in IPv4. For example, Neighbor Discovery uses ICMPv6 multicast rather than an ARP broadcast.

IPv6 address types and prefixes

Global unicast addresses

Global unicast addresses normally come from 2000::/3. They are the IPv6 equivalent of publicly routable IPv4 addresses, although an address is reachable from the Internet only when the relevant routes, security policies, and upstream connectivity exist.

A common enterprise /64 can be viewed as three logical sections:

• Global routing prefix: allocated by a provider or registry and used for hierarchical routing.
• Subnet ID: identifies a particular subnet within the organization.
• Interface ID: identifies an interface on that subnet.

2001:DB8:0100:0010:0000:0000:0000:0025/64
|--------- global and subnet prefix --------| interface ID |

For example, 2001:DB8:100:10::25/64 is a global-unicast-format address, and 2001:DB8:100:10::/64 is its subnet. However, 2001:DB8::/32 is reserved for documentation. Use it in labs and examples, not production networks.

A /64 provides 64 bits for interface identifiers and is the typical subnet size for hosts using SLAAC. Enterprises, ISPs, and Internet-connected networks commonly allocate multiple /64 subnets from a larger global routing prefix.

Link-local unicast addresses

Link-local addresses use the FE80::/10 range. An IPv6-enabled interface normally creates a link-local address automatically, even when no global address has been configured.

Link-local traffic remains on the local Layer 2 segment. Routers do not forward a link-local packet beyond that link. Link-local addresses are nevertheless essential for:

• Neighbor Discovery Protocol, including Neighbor Solicitation and Neighbor Advertisement.
• Router Solicitation and Router Advertisement messages.
• IPv6 routing protocols and routing-protocol adjacencies.
• Communication with a directly connected next hop.

Because the same link-local value could exist on multiple interfaces, Cisco commands often require an outgoing interface when a link-local address is used as a next hop.

ipv6 route 2001:DB8:200:20::/64 gigabitEthernet0/0 FE80::2

The route identifies both the destination prefix and the interface-local next hop. A link-local next hop from another Layer 2 segment is not valid.

Unique local addresses

Unique local addresses, or ULAs, use FC00::/7. In common locally assigned deployments, addresses use the FD00::/8 portion. ULAs are intended for internal communication and are not intended to be reachable through the public Internet.

FD12:3456:789A:0001::/64

ULAs resemble IPv4 private addressing such as RFC 1918 space, but they are not identical. IPv4 private ranges are widely reused by unrelated organizations and frequently require NAT for Internet access. ULA design uses a pseudo-random global ID to reduce the chance of collisions when organizations connect networks, and IPv6 hosts can use ULAs without making NAT a required design feature.

ULAs may be used alongside global unicast addresses. For example, an internal service can have a ULA for stable internal access and a global address for approved external access. Routing and firewall policy must still control which paths are permitted.

Loopback and unspecified addresses

The loopback address is ::1/128. It identifies the local host, not a physical or routed interface, and is used to test the local IPv6 stack. It corresponds conceptually to IPv4 127.0.0.1.

The unspecified address is ::/128. It means “no address” and is not assigned to an interface as a normal usable address. A host may use it as a source while determining its address, such as during early Neighbor Discovery and Duplicate Address Detection. It corresponds conceptually to IPv4 0.0.0.0.

IPv6 multicast addresses

Multicast addresses begin with FF00::/8. An IPv6 multicast address includes fields for flags and scope. The scope field indicates how far the multicast traffic is intended to travel. Common scope values include:

• 1: interface-local scope.
• 2: link-local scope.
• 5: site-local scope in multicast-scope terminology.
• 8: organization-local scope.
• E: global scope.

The second byte commonly displays flags and the four-bit scope together. For example, FF02:: indicates a multicast address with link-local scope.

Router Advertisements are sent using IPv6 multicast. Neighbor Discovery uses multicast to limit traffic to relevant nodes instead of interrupting every device with a broadcast.

Solicited-node multicast addresses

Every IPv6 unicast and anycast address has a corresponding solicited-node multicast address. The address is formed by taking the low-order 24 bits of the unicast or anycast address and appending them to FF02::1:FF00:0/104.

Example:

Unicast address:       2001:DB8:1:1::1234:ABCD
Low-order 24 bits:                         34:ABCD
Solicited-node group: FF02::1:FF34:ABCD

The node uses this group as the destination for a Neighbor Solicitation message. The target can respond with a Neighbor Advertisement, allowing address resolution without an IPv4-style ARP broadcast. The same process supports Duplicate Address Detection, which checks whether another node is already using an address.

At Layer 2, IPv6 multicast is mapped to an Ethernet multicast destination. A switch can therefore forward the frame as multicast rather than treating it as a broadcast to every host. Multicast filtering, VLAN operation, and ICMPv6 security policies must not block required Neighbor Discovery traffic.

Anycast addresses

An anycast address is one address assigned to multiple interfaces, usually on different devices. Routing determines which instance receives a packet. The selected instance is generally the nearest according to the routing topology, not necessarily the device with the smallest geographic distance.

Anycast is useful for redundant services. For example, two DNS resolvers can advertise the same anycast-format IPv6 address. A client reaches whichever resolver is topologically closest and reachable. A related reserved form is subnet-router anycast: an address formed from a subnet prefix with an all-zero interface identifier, such as the address represented by a subnet's prefix and ::.

Anycast addresses have unicast format, so their purpose cannot be identified solely by looking at the address. Configuration and routing documentation must indicate that the address is anycast. Do not confuse anycast with multicast: anycast selects one instance, while multicast delivers to all joined members.

Special and reserved IPv6 ranges

IPv4-embedded representations

Some IPv6 notation represents an IPv4 value inside an IPv6 address. IPv4-mapped addresses use the form ::FFFF:w.x.y.z, such as ::FFFF:192.0.2.10. These are commonly used by software APIs to represent an IPv4 peer in an IPv6-capable application and are not ordinary globally routable IPv6 addresses.

Older IPv4-compatible IPv6 concepts also used IPv4 values embedded in IPv6 notation, but that mechanism is deprecated. On an exam, distinguish the representation of an IPv4 value from a normal global IPv6 address.

Documentation and deprecated site-local ranges

2001:DB8::/32 is reserved for documentation. It is ideal for diagrams, configurations shown in lessons, and lab examples, but external connectivity should not be expected.

FEC0::/10 is deprecated site-local addressing. Modern designs should use ULAs from FC00::/7, commonly from FD00::/8, instead.

IPv4 and IPv6 behavior comparison

IPv6 address assignment methods

EUI-64 creates a 64-bit interface identifier from a MAC address. Modern operating systems often use stable-random or temporary privacy identifiers instead, so the address displayed on a host may not reveal its MAC address.

SLAAC depends on Router Advertisement messages. DHCPv6 can be stateful, assigning addresses, or stateless, supplying additional information while SLAAC assigns the address. A host can have multiple IPv6 addresses at once, including a link-local address, one or more global or ULA addresses, and temporary privacy addresses.

Cisco IOS configuration context

Enable IPv6 forwarding

A Cisco router must have IPv6 forwarding enabled before it can route IPv6 packets between interfaces.

configure terminal
ipv6 unicast-routing

Assign global and link-local addresses

interface gigabitEthernet0/0
 ipv6 address 2001:DB8:100:10::1/64
 no shutdown

An interface normally creates a link-local address automatically. You can assign a specific link-local address when predictable next-hop addressing is required.

interface gigabitEthernet0/0
 ipv6 address FE80::1 link-local

Use EUI-64

interface gigabitEthernet0/0
 ipv6 address 2001:DB8:100:10::/64 eui-64

The command supplies the /64 prefix and asks IOS to generate the interface identifier. The resulting address can be viewed with interface verification commands.

Verify IPv6 operation

show ipv6 interface brief
show ipv6 interface gigabitEthernet0/0
show ipv6 route
show ipv6 neighbors
ping ipv6 2001:DB8:100:10::2
traceroute ipv6 2001:DB8:100:10::2

show ipv6 interface brief provides a quick view of interface state and addresses. The detailed interface command shows IPv6 configuration and Neighbor Discovery information. The route and neighbor commands verify Layer 3 paths and local neighbor resolution.

For routing-protocol context, review Configure OSPF; IPv6 routing protocols commonly use link-local addresses for neighbor relationships.

Troubleshooting IPv6 address problems

Only a link-local address is present

If a host has an FE80:: address but no usable global address, check whether a Router Advertisement is arriving, whether the router interface has a global /64 prefix, and whether IPv6 forwarding and interface configuration are complete.

• Verify router addressing with show ipv6 interface.
• Confirm the interface and VLAN are operational.
• Confirm that the intended /64 prefix is advertised.
• Inspect neighbor state with show ipv6 neighbors.
• Check that local multicast and ICMPv6 traffic are not filtered.

A static route with a link-local next hop fails

Specify both the exit interface and the link-local next hop:

ipv6 route 2001:DB8:200:20::/64 gigabitEthernet0/0 FE80::2

Also verify that FE80::2 belongs to a neighbor on that interface, that the neighbor is reachable, and that its interface is up.

Neighbor resolution fails on a LAN

Check for ICMPv6 or Neighbor Discovery filtering, incorrect target addresses, duplicate address detection failures, VLAN problems, and blocked solicited-node multicast traffic. Examine the IPv6 neighbor table and confirm the local link is operational.

An application expects broadcast

IPv6 does not provide broadcast. Identify the application's IPv6 multicast discovery mechanism, confirm IPv6 support, and permit the required multicast and ICMPv6 messages where appropriate.

A documentation address cannot reach the Internet

An address from 2001:DB8::/32 is not a production allocation. Replace it with an ISP- or registry-assigned global prefix and verify both forward and return routes.

Exam-focused summary

• IPv6 addresses are 128 bits, written as eight 16-bit hexadecimal hextets.
• Remove leading zeros per hextet and use :: only once to compress consecutive zero hextets.
• A /64 commonly divides a subnet prefix from a 64-bit interface identifier.
• Unicast is one-to-one, multicast is one-to-many, and anycast is one-to-nearest.
• IPv6 has no broadcast; multicast replaces many broadcast functions.
• 2000::/3 identifies global-unicast-format addresses.
• FE80::/10 identifies link-local addresses, which are not routed beyond the local link.
• FC00::/7, commonly FD00::/8, identifies unique local addresses.
• ::1 is loopback and :: is unspecified.
• FF00::/8 is multicast, with scope encoded in the address.
• Solicited-node multicast uses FF02::1:FF00:0/104 plus the low-order 24 bits of a unicast or anycast address.
• 2001:DB8::/32 is for documentation, and FEC0::/10 site-local addressing is deprecated.
• Link-local next hops in Cisco IOS generally require an exit interface.