Answers: EIGRP for IPv6 1

By certskills April 14, 2016 09:05


This lab has some repetitive configuration elements for EIGRPv6. Check out the original lab first, and then come here to check out your answers. The main task in this case: Get EIGRP working for IPv6 in a three router pod.



Figure 1: Three Router Triangle with Unicast IPv6 Addresses


Example 4: R1 Config


Example 5: R2 Config


Example 6: R3 Config



EIGRP for IPv6 is very similar to its IPv4 variant, it still uses the bandwidth and delay of the route path to calculate metric and still maintains topology and neighbor tables. There are, however, some unique parts of the configuration that are different from legacy EIGRP configuration. First, EIGRP for IPv6 uses an interface style of configuration (i.e. using ipv6 eigrp commands). Second, unlike EIGRP for IPv4, EIGRP for IPv6 requires either an interface on the device to be configured with an IPv4 address OR that the router-id is statically configured as part of the EIGRP for IPv6 configuration; this is because an IPv4 style of router-id is still used.

For this lab, the bandwidth of the link between R2 and R3 is altered to make it less preferred over its faster alternative. This change will affect the routing tables of R2 and R3; specifically, it will affect how many routes R2 will have for the 2001::/64 network and how many routes R3 will have for the 2000::/64 network.

All three routers use the ipv6 router eigrp 20 command, which creates an EIGRPv6 process on each router, with ASN 20, a number which must match on the three routers. Then each router has a command like R1’s eigrp router-id to define the EIGRP router ID. Note that as shown, this router has no IPv4 addresses at all, so the router ID must be explicitly configured as shown here.

Then look at all the interfaces on all three routers. Each interface has an ipv6 eigrp 20 subcommand, enabling the use of the ASN 20 EIGRPv6 process on that interface.

Finally, look at R2’s G0/1 interface and R3’s G0/2 interface. They both use the bandwidth 100000 subcommand, setting the speed down to match the actual link speed. As a result, EIGRP on R2 and R3 will choose routes that run through R1 rather than directly between R2 and R3. Also, note that this bandwidth 100000 command is not actually needed, because the bandwidth setting on router LAN interfaces will adjust to the negotiated speed assuming no bandwidth command is configured on the interface.

Answers: Local DHCP Server 1
Remote DHCP Server 1
By certskills April 14, 2016 09:05
Write a comment


  1. Viktor February 17, 02:52

    Hello. Thanks a lot for the material.
    Have a question. Studuing your 200-105 book, now – OSPFv3 and EIGRP for IPv6. You briefly mention that routers may become neighbors even if their link addresses are not at the same subnet… I suppose that this brief mention sometimes plays a great role, but I break my mind and can not imagine the situation with EIGRP for v6 implemented with routers in different subnets. Could you please provide some explanation for this?
    Thankfully, Viktor

    Reply to this comment
    • certskills Author February 17, 09:25

      Hi Viktor,
      Ponder the packet delivery process. With a link local address available (which isn’t available in IPv4), the IPv6 routers are able to send a packet with source of link local and destination of an IPv6 multicast. And that’s what they use. The IPv6 EIGRP processes don’t even see the neighboring router’s global unicast address, for instance, so there’s no ability to even do a check to make sure they are in the same subnet.

      Could the people that creates OSPF and EIGRP for IPv6 have chosen to use global unicast or unique local addresses as the source of those packets… and therefore then also required the addresses to be in the same subnet? Sure. They just didn’t. Some of the reasons why:
      Addressing flexibility
      EG, you don’t even have to use global unicasts or unquiet locals on WAN links – just take advantage of the way the routing protocols and IPv6 routes use the link local addresses.

      Also, in the case of OSPF, they changed the structure of some of the LSAs, which removed an OSPF restriction for IPv4 that required the IPv4 addresses on neighbors to be in the same subnet.

      That’s some of it… may be worth looking further if you get into CCNP ROUTE?

      Reply to this comment
View comments

Write a comment

Comment; Identify w/ Social Media or Email