EIGRP fundamentals: DUAL, successors, feasible successors and the composite metric

Where EIGRP fits

Enhanced Interior Gateway Routing Protocol (EIGRP) was created by Cisco and later published as an informational RFC (RFC 7868). It behaves like a distance-vector protocol, learning routes from neighbours, but adds features that make it fast and efficient: partial updates only when something changes, and pre-computed backup paths. Its administrative distance is 90 for internal routes and 170 for external (redistributed) routes. In practice it appears mostly in networks built on Cisco equipment; multi-vendor networks usually choose OSPF.

Neighbours and the three tables

EIGRP routers send hellos to 224.0.0.10 (every 5 seconds on fast links, with a 15-second hold time) and use IP protocol 88. To become neighbours they must share the same autonomous system number, K values (metric weights), subnet and authentication settings. Each router then keeps three tables:

• Neighbour table: adjacent EIGRP routers.
• Topology table: every route learned from every neighbour, with its metrics.
• Routing table: the best routes chosen from the topology table.

The metric

By default EIGRP uses two values: the slowest bandwidth along the path and the total delay of all outgoing interfaces. The classic formula is 256 multiplied by (10,000,000 divided by the slowest bandwidth in kbit/s, plus the total delay in tens of microseconds). Load and reliability exist as options but are off by default because they make routes unstable. Newer "named mode" configurations use 64-bit wide metrics so that very fast links are still told apart.

Because the metric uses interface settings, the bandwidth and delay commands change path selection. They do not change the real link speed.

DUAL: successors and feasible successors

Two distances are key to the Diffusing Update Algorithm (DUAL):

• Reported distance (RD): the neighbour's own metric to the destination, as it advertised it.
• Feasible distance (FD): your total metric through that neighbour; the lowest FD is your best path.

The neighbour giving the lowest FD is the successor. Another neighbour is a feasible successor if its RD is lower than the successor's FD: the feasibility condition. That proves the neighbour's path does not loop back through you, so it can be used immediately if the successor fails. If there is no feasible successor, the route goes "active" and the router queries its neighbours for a new path.

Example: R1 reaches 10.9.0.0/16 via R2 with FD 3000. R3 reports an RD of 2500 and would give R1 an FD of 4000. Because 2500 is less than 3000, R3 is a feasible successor and is ready the moment R2 fails.

Configuration and load balancing

router eigrp 100
 network 10.1.12.0 0.0.0.3
 network 192.168.10.0 0.0.0.255
 passive-interface default
 no passive-interface GigabitEthernet0/1
 variance 2

show ip eigrp neighbors
show ip eigrp topology
show ip route eigrp

EIGRP shares traffic across equal-cost paths by default. The variance command also allows unequal-cost load balancing: with variance 2, any feasible successor whose metric is up to twice the best is used, in proportion to its metric. Only feasible successors qualify, so loops cannot form.

Real-world example: a company with two WAN links to each branch, 100 Mbit/s and 20 Mbit/s, wanted to use both. With OSPF only the faster link carried traffic. Their Cisco-only network ran EIGRP, so setting variance on the branch routers spread traffic across both links in proportion to their metrics, while keeping each link as an instant backup for the other.

Remember that both ends of a link must agree on the autonomous system number.