Cisco IOS — IPv6 Configuration

Overview

IPv6 was developed to address the exhaustion of IPv4’s 32-bit address space. With approximately 4 billion addresses, IPv4 was insufficient for global growth, and by 2011 IANA had exhausted its /8 allocations. IPv6 expands addressing to 128 bits, providing roughly 3.4 × 10^38 unique addresses. On Cisco IOS, IPv6 can run alongside IPv4 in a dual-stack configuration, allowing a gradual migration rather than an abrupt cutover.

This article covers the IOS commands required to configure IPv6 on routers and switches, including interface addressing methods, static routes, OSPFv3, and verification tools.

Enabling IPv6 Routing

Before a Cisco router will forward IPv6 packets on behalf of other hosts, you must issue a single global command. Without it, the router handles IPv6 only for its own stack — it will not route transit traffic.

Router(config)# ipv6 unicast-routing

This command activates the IPv6 forwarding engine and enables the router to participate in IPv6 routing protocols. It is the IPv6 equivalent of the default behaviour of IPv4 routing (which is always on). Forgetting this command is one of the most common causes of IPv6 connectivity failures on IOS.

Interface IPv6 Address Configuration

Static Global Unicast Address

The most direct method is assigning a full 128-bit address to an interface:

interface GigabitEthernet0/0
 ipv6 address 2001:DB8:1::1/64
 no shutdown

The /64 prefix length is the recommended default for all deployed subnets. When you assign a global unicast address (GUA), IOS also automatically generates a link-local address (LLA) using EUI-64 from the interface MAC address.

Link-Local Address (Manual)

Every IPv6 interface requires a link-local address in the FE80::/10 range. IOS generates one automatically, but it can be overridden for clarity and to produce a more memorable address:

interface GigabitEthernet0/0
 ipv6 address FE80::1 link-local

Using a simple, predictable LLA (like FE80::1) is common in lab and production environments because LLAs appear as next-hop addresses in routing tables and are easier to reference when troubleshooting.

EUI-64 Address Generation

EUI-64 lets the router auto-generate the 64-bit interface ID (IID) portion of an IPv6 address from the interface’s MAC address. You provide only the 64-bit prefix:

interface GigabitEthernet0/0
 ipv6 address 2001:DB8::/64 eui-64

The IOS EUI-64 process works as follows:

1. Take the 48-bit MAC address and split it into two 24-bit halves.
2. Insert FFFE in the middle, creating a 64-bit value.
3. Invert the 7th bit (Universal/Local bit) of the result.

For example, a MAC of 0200.1111.1111 becomes IID 0000:11FF:FE11:1111, and the full address would be 2001:DB8::11FF:FE11:1111/64. EUI-64 is useful when you want consistent addressing tied to hardware identity without manually specifying each interface ID.

SLAAC (Stateless Address Autoconfiguration)

Interfaces can learn their IPv6 address dynamically from a router’s RA (Router Advertisement) message via SLAAC:

interface GigabitEthernet0/0
 ipv6 address autoconfig

The router listens for RA messages, extracts the on-link prefix, and combines it with a locally generated IID (either EUI-64 or a random privacy extension value). SLAAC is commonly used on client-facing interfaces.

Link-Local Only (ipv6 enable)

If you want to activate IPv6 on an interface without assigning a routable global address — for instance on a WAN link where only a next-hop relationship is needed — use:

interface GigabitEthernet0/1
 ipv6 enable

This generates a link-local address only, with no GUA. Routing protocols such as OSPFv3 can use LLAs as next-hops, so a fully-addressed subnet is not required on every link.

Dual-Stack Operation

Dual-stack means both ip address and ipv6 address are configured on the same interface simultaneously. The router processes IPv4 and IPv6 packets independently using separate routing tables:

interface GigabitEthernet0/0
 ip address 192.168.1.1 255.255.255.0
 ipv6 address 2001:DB8:1::1/64
 no shutdown

Dual-stack is the standard transition strategy. It requires no tunnelling overhead and allows hosts to use whichever protocol they prefer. Both stacks operate independently, so an issue on one does not affect the other.

Verification Commands

Interface Status

show ipv6 interface brief

Displays all interfaces with their IPv6 addresses in condensed form. For each interface you see the GUA, the automatically generated or manually configured LLA, and the interface state.

show ipv6 interface GigabitEthernet0/0

Provides detailed output including all unicast and multicast addresses assigned to the interface, including solicited-node multicast addresses used by NDP.

Routing Table

show ipv6 route
show ipv6 route connected
show ipv6 route static

The IPv6 routing table is separate from the IPv4 table. Route codes include:

• C — Connected
• L — Local (/128 host route for each interface address)
• S — Static
• O — OSPFv3 intra-area
• OI — OSPFv3 inter-area
• ND — NDP-derived default route

NDP Neighbor Table

NDP (Neighbor Discovery Protocol) replaces ARP for IPv6. The neighbor table stores IPv6-to-MAC mappings:

show ipv6 neighbors

This is the IPv6 equivalent of show arp. Entries move through states: INCOMPLETE → REACHABLE → STALE → DELAY → PROBE.

Static IPv6 Routes

IPv6 static routes follow the same logic as IPv4 but use the ipv6 route command and support link-local next-hops (which require specifying an outgoing interface because LLAs are not globally unique).

Network Route

ipv6 route 2001:DB8:2::/64 GigabitEthernet0/1 FE80::2

Using a link-local address as the next-hop is the recommended practice on Ethernet links. The outgoing interface must be specified alongside the LLA.

GUA Next-Hop (without interface)

ipv6 route 2001:DB8:3::/64 2001:DB8:1::2

If the next-hop is a GUA, the interface is determined by recursive lookup and does not need to be specified explicitly.

Default Route

ipv6 route ::/0 GigabitEthernet0/1 FE80::1

The IPv6 default route uses ::/0 (all zeros, prefix length zero). Unlike IPv4, IOS does not display a “Gateway of Last Resort” line — only the ::/0 entry appears in the routing table.

Host Route (/128)

ipv6 route 2001:DB8:1::100/128 GigabitEthernet0/0 FE80::2

A /128 host route targets a specific device and has the highest specificity in longest-prefix matching.

Floating Static Route (Backup)

ipv6 route 2001:DB8:2::/64 cellular0/1 130

Setting an administrative distance higher than the primary protocol (OSPFv3 uses 110) creates a backup route that enters the routing table only when the preferred route is gone.

OSPFv3 Configuration

OSPFv3 is the IPv6-capable version of OSPF (RFC 5340). On IOS, OSPFv3 is configured primarily at the interface level rather than through a network statement.

Global Process Configuration

ipv6 router ospf 1
 router-id 1.1.1.1

A router ID must be explicitly configured for OSPFv3 because there may be no IPv4 addresses on the router. The RID is still a 32-bit dotted-decimal value used only as an identifier — it does not need to match any actual address.

Interface Activation

interface GigabitEthernet0/0
 ipv6 ospf 1 area 0

This single interface command enrolls the interface in OSPFv3 process 1, area 0. Unlike OSPFv2, there is no network command under the process. Each interface is enabled individually, giving precise control.

Passive Interface

ipv6 router ospf 1
 passive-interface GigabitEthernet0/2

A passive interface advertises its connected prefix into OSPF but does not send or receive Hello messages. Use this on LAN interfaces facing end hosts to prevent unnecessary adjacency attempts.

Default Route Origination

ipv6 router ospf 1
 default-information originate always

Advertises ::/0 into OSPFv3. The always keyword causes advertisement even if the router has no default route itself — useful on edge routers where the upstream default is known to exist.

OSPFv3 Verification

show ipv6 ospf neighbor
show ipv6 ospf interface brief
show ipv6 ospf database
show ipv6 route ospf

The show ipv6 ospf neighbor output shows neighbor RIDs, states (Full/DR, Full/BDR, 2WAY/DROTHER), and interface. The neighbour must reach Full state for route exchange to complete.

OSPFv3 vs OSPFv2 Key Differences

Common Troubleshooting Steps

When IPv6 connectivity fails, work through these checks in order:

1. Confirm ipv6 unicast-routing is configured globally.
2. Verify interface IPv6 addresses with show ipv6 interface brief — check both GUA and LLA are present.
3. Check the routing table with show ipv6 route — confirm the expected prefixes appear.
4. Use show ipv6 neighbors to confirm NDP is resolving MAC addresses on local segments.
5. For OSPFv3 issues, verify show ipv6 ospf neighbor shows Full adjacency and that router IDs are unique.
6. For static route issues, remember that LLA next-hops require the outgoing interface to be specified; a bare LLA without an interface is rejected by IOS.

References

• RFC 8200 — Internet Protocol Version 6 Specification
• RFC 4861 — Neighbor Discovery for IP version 6
• RFC 4862 — IPv6 Stateless Address Autoconfiguration
• RFC 5340 — OSPF for IPv6
• CCNA 200-301 Official Cert Guide Vol. 1 (Odom) — Chapters 25–29