Octet

Load share over a 200 Mbps and a 100 Mbps WAN circuit with EIGRP variance

Cisco IOSExpert

Tasks

  1. Set the real circuit bandwidth on both ends of each WAN link.
  2. Run EIGRP AS 100 on all four devices with fixed router IDs.
  3. Before adding variance, confirm R3 is a feasible successor for 10.100.0.0/24 on R1.
  4. Add the smallest variance that installs both paths, and verify the routing table and CEF.

Solution

  1. 1WAN interfaces on R1 with the circuit bandwidthR1
    R1 on Cisco IOS
    interface GigabitEthernet0/0/0
     description WAN to R2 - 200M circuit
     bandwidth 200000
     ip address 10.0.12.1 255.255.255.252
     no shutdown
    interface GigabitEthernet0/0/1
     description WAN to R3 - 100M circuit
     bandwidth 100000
     ip address 10.0.13.1 255.255.255.252
     no shutdown

    bandwidth changes only the numbers routing and QoS use. It does not shape or limit traffic.

  2. 2EIGRP on R1R1
    R1 on Cisco IOS
    router eigrp 100
     network 10.0.12.0 0.0.0.3
     network 10.0.13.0 0.0.0.3
     eigrp router-id 10.255.0.1
  3. 3R2 and R3, with matching bandwidth on their WAN sideR2
    R2 on Cisco IOS
    interface GigabitEthernet0/0/0
     description WAN to R1 - 200M circuit
     bandwidth 200000
     ip address 10.0.12.2 255.255.255.252
     no shutdown
    interface GigabitEthernet0/0/1
     ip address 10.0.2.1 255.255.255.252
     no shutdown
    !
    router eigrp 100
     network 10.0.12.0 0.0.0.3
     network 10.0.2.0 0.0.0.3
     eigrp router-id 10.255.0.2
    !
    ! R3 is the same with 10.0.13.2, 10.0.3.1, bandwidth 100000 and router ID 10.255.0.3

    Set bandwidth on both ends. Otherwise traffic from headquarters back to the branch is calculated as if both circuits were gigabit.

  4. 4CORE1CORE1
    CORE1 on Cisco IOS
    interface GigabitEthernet1/0/1
     no switchport
     ip address 10.0.2.2 255.255.255.252
    interface GigabitEthernet1/0/2
     no switchport
     ip address 10.0.3.2 255.255.255.252
    interface Vlan100
     ip address 10.100.0.1 255.255.255.0
    !
    router eigrp 100
     network 10.0.2.0 0.0.0.3
     network 10.0.3.0 0.0.0.3
     network 10.100.0.0 0.0.0.255
     passive-interface Vlan100
     eigrp router-id 10.255.0.10
  5. 5Variance on R1R1
    R1 on Cisco IOS
    router eigrp 100
     variance 2

    maximum-paths (default 4) still caps the total number of paths. traffic-share balanced is the default and splits traffic in inverse proportion to the metrics.

Verify

R1 output
R1# show ip eigrp topology 10.100.0.0/24
EIGRP-IPv4 Topology Entry for AS(100)/ID(10.255.0.1) for 10.100.0.0/24
  State is Passive, Query origin flag is 1, 1 Successor(s), FD is 13568
  Descriptor Blocks:
  10.0.12.2 (GigabitEthernet0/0/0), from 10.0.12.2, Send flag is 0x0
      Composite metric is (13568/3072), route is Internal
      Vector metric:
        Minimum bandwidth is 200000 Kbit
        Total delay is 30 microseconds
        Reliability is 255/255
        Load is 1/255
        Minimum MTU is 1500
        Hop count is 2
        Originating router is 10.255.0.10
  10.0.13.2 (GigabitEthernet0/0/1), from 10.0.13.2, Send flag is 0x0
      Composite metric is (26368/3072), route is Internal
      Vector metric:
        Minimum bandwidth is 100000 Kbit
        Total delay is 30 microseconds
        Reliability is 255/255
        Load is 1/255
        Minimum MTU is 1500
        Hop count is 2
        Originating router is 10.255.0.10

Run this before step 5. The numbers match the hand calculation: minimum bandwidth 200000 and 100000 Kbit, total delay 30 microseconds on both, RD 3072 from each neighbor. R3's RD of 3072 is below the FD of 13568, which is what makes it a feasible successor.

R1 output
R1# show ip route eigrp | section 10.100.0.0
D        10.100.0.0/24 [90/13568] via 10.0.12.2, 00:00:41, GigabitEthernet0/0/0
                      [90/26368] via 10.0.13.2, 00:00:41, GigabitEthernet0/0/1

After variance 2, both next hops are installed, each with its own metric.

R1 output
R1# show ip cef 10.100.0.0
10.100.0.0/24
  nexthop 10.0.12.2 GigabitEthernet0/0/0
  nexthop 10.0.13.2 GigabitEthernet0/0/1

CEF has both next hops, so forwarding really uses them. Sharing is per flow by default.

Mistakes people make

  • Setting variance and expecting a path that is not a feasible successor to be used. Check the RD against the FD first.
  • Rounding: 26368 / 13568 is 1.94, so variance 2 works here, but a ratio of 2.05 would need variance 3.
  • Setting bandwidth on one end only, which gives different metrics in each direction and asymmetric traffic.
  • Expecting an exact 2 to 1 split. CEF hashes flows, so a few large flows can land mostly on one circuit.

On other platforms

Named mode puts variance under topology base.

Cisco IOS
router eigrp BRANCH
 address-family ipv4 unicast autonomous-system 100
  topology base
   variance 2
  exit-af-topology
  network 10.0.12.0 0.0.0.3
  network 10.0.13.0 0.0.0.3
  eigrp router-id 10.255.0.1
 exit-address-family

Say it in the interview

"First I check eligibility, not the variance number: R3 reports 3072, below the FD of 13568, so it is a feasible successor. Then the ratio: 26368 over 13568 is about 1.94, so variance 2 installs it. I set the real circuit bandwidth on both ends, because EIGRP otherwise sees two gigabit ports, and I warn that per-flow hashing makes the split approximate."

Book 3.4 Routing