EIGRP Overview

EIGRP Facts


Enhanced IGRP is a Cisco-proprietary balanced hybrid routing protocol that combines the best features of distance vector and link state routing. EIGRP:
  • Sends the subnet mask in the routing update. It supports route summarization and VLSM.
  • Supports automatic classful route summarization at major network boundaries (this is the default in EIGRP). Unlike IGRP and RIP, manual route summarization can also be configured on arbitrary network boundaries to reduce the routing table size.
  • Is not susceptible to routing loops. Instead, EIGRP uses built-in loop avoidance techniques. Under certain conditions, EIGRP will use split horizon, but not hold downs or flush timers.
  • Is scalable and does not have the 16 hop limitation of RIP.
  • Uses hello packets to discover neighbor routers. Hello intervals on EIGRP routers do not need to match.
  • Exchanges the full routing table at startup, and then partial routing updates thereafter.
  • Uses unicasts or multicasts to 224.0.0.10 for routing updates. Hello packets always use the multicast address.
  • Uses bandwidth, delay, reliability, and load for the route metric. The metric is expressed as the number of microseconds.
    • The degree to which each value is used to calculate the metric can be customized by modifying one of five K values.
    • By default, K1 and K3 are set to 1, while K2, K4, and K5 are set to 0. These settings mean that with the default configuration, only delay and bandwidth have an effect on the metric.
    • On serial links, a default bandwidth of 1544 is used. EIGRP does not detect the actual bandwidth on the link. You must manually configure bandwidth values for accurate metric calculations.
  • Uses an autonomous system (AS) number to identify routers that are to share EIGRP information. The AS number on both routers must match.
  • Maintains partial network topology information in addition to routes.
  • Supports load balancing on equal-cost and unequal cost links. This means that EIGRP can keep multiple paths to a single network, even if they have a different cost. With IOS 12.4 and above, EIGRP supports up to 16 paths (earlier versions supported up to 6), with the default being 4 equal-cost paths.
  • Minimizes network bandwidth usage for routing updates. During normal operation EIGRP transmits only hello packets across the network. EIGRP does not send periodic routing updates like RIP and IGRP. When change occurs, only routing table changes are propagated in EIGRP not the entire table.
  • Requires less processing and memory than link state protocols.
  • Converges more quickly than distance vector protocols. In some cases, convergence can be almost instantaneous because an EIGRP router stores backup routes for destinations. If no appropriate route or backup exists in the routing table, EIGRP will query neighbor routers to discover an alternate route. In this manner, EIGRP can quickly adapt to alternate routes when changes occur.
  • Uses the DUAL link-state algorithm for calculating routes.
  • Supports multiple protocols. EIGRP can exchange routes for IP, AppleTalk and IPX/SPX networks.
  • Uses a neighbors table to keep track of neighbor routers. The neighbors table includes the following for each neighbor:
    • A hold time. Each hello packet includes a hold time that identifies how long the hello information is valid. If the hold time expires without receipt of a hello packet, the neighbor is assumed to be unreachable.
    • Round-trip timers that help the router identify cost values to reach the neighbor router.
  • Uses a topology database to keep track of all known networks.
    • The topology table has a list of each destination network and all neighbor routers that reported routes to that network.
    • The best routes that will be used for routing packets are copied from the topology table into the routing table.
    • The topology table holds up to 16 known routes (previously up to 6 before IOS version 12.4).
To understand how EIGRP can provide load balancing and fast recovery for failed links, you need to understand the following concepts:
Term Definition
Advertised Distance (AD) The advertised distance (AD) is the cost to the destination network as reported by the neighbor router. The AD is also called the reported distance (RD).
Feasible Distance (FD) The feasible distance (FD) is the lowest total cost to a destination network. The feasible distance is identified for each destination network, and is determined as follows: # For each neighbor, a total cost to the network through the neighbor is calculated by adding the AD to the cost required to reach the neighbor router (the cost of the link used to reach the neighbor router).
  1. The router compares the total cost of all routes. The lowest total cost to the destination network is the feasible distance to the network.
Note: Sometimes the total cost for each neighbor route is referred to as a feasible distance. However, the term more correctly identifies the lowest known cost to the network, not the total cost for each reported (possible) route.
Successor A successor is the route to a destination network with the lowest total cost. * When a new route is first learned, the total cost to the successor route is used as the feasible distance to that network.
  • The successor route is copied from the topology table into the routing table.
  • You can have multiple successor routes if multiple routes to the same network exist with the same lowest metric.
Feasible Successor A feasible successor is an alternate route to a destination network. The total cost to the route through the feasible successor is higher than the total cost of successor routes. A route must meet the following condition to qualify as a feasible successor route:
The advertised distance of the route through that neighbor must be less than the feasible distance used for that network (AD < FD).
Be aware of the following regarding feasible successors: * Satisfying the AD < FD condition ensures that the route is loop free. In other words, the router knows for sure that the route does not include itself in the path if the AD is lower than the FD. Note: Successor routes must also meet this condition.
  • Feasible successor routes are kept in the topology table but are not copied to the routing table.
  • Successor routes can also be classified as feasible successor routes.
  • When all successor routes to a network are lost, the router can immediately begin to use the next best feasible successor route. This provides for rapid recovery in the event of a topology change.
Be aware of the following regarding the EIGRP and routes:
  • All known routes to a destination are kept in the topology table. Only successor routes are copied to the routing table.
  • If the successor route goes down and there are no feasible successors, routes whose advertised distance is greater than the feasible distance for the route are not used because they might be routes that include loops.
  • When the last feasible successor route to a network is lost, the router recalculates all routes for the lost neighbor. Instead of using other routes that are not feasible successor routes, it first communicates with neighbor routers. If necessary, the router recalculates the feasible distance for the route.
  • A route whose AD is greater than the FD does not prove that a loop exists, only that a loop might exist. After the last feasible successor route is lost, a previously unacceptable route could be identified as a feasible successor route as long as its AD is less than the newly-calculated FD.
  • By default, EIGRP uses equal-cost load balancing. To use unequal-cost load balancing, configure the variance value. The variance is a multiplier that identifies the degree to which alternate paths can be used.
    • The variance value ranges from 1 to 255.
    • The default variance is 1, meaning that only routes that match the best route can be used.
    • Setting the variance to 2 allows alternate routes to be used whose total costs are within a factor of 2 (double or less) of the best cost route.
    • Only feasible successor routes can be used. This means that a route whose AD is greater than the FD cannot be used as an alternate route, even if its total cost is within the variance amount.
For an EIGRP router to share information with a neighbor, the following conditions must be met:
  • Both routers are on the same subnet with the same subnet mask.
  • If used, authentication checks must pass.
  • Both routers must be configured with the same AS number.
  • Metric weight values (K values) must match on both routers.


EIGRP Command List

You configure EIGRP just the same as you would configure IGRP. The following table lists the applicable commands.
Command Function
Router(config)#router eigrp number Defines an EIGRP process.
The number must match between routers for information to be shared.
-->
Router(config-router)#network n.n.n.n
Router(config-router)#network n.n.n.n w.w.w.w

Identifies a network that participates in the routing process.
Networks can be specified with or without the wildcard mask. If you do not use a wildcard mask, the network address you add will be automatically truncated based on classful network boundaries.
You must use a wildcard mask to identify VLSM subnets.

-->
Router(config-router)#no auto-summary

Turn off automatic route summarization.
With automatic route summarization, subnets are summarized based on classful boundaries when advertising routes on networks with a different class boundary. You must disable automatic summarization if you have a network address (such as 10.0.0.0) subnetted into smaller subnets and separated by a network with a different classful network address (such as 12.0.0.0).
Example
The following commands enable EIGRP on a router and define three networks that participate in the routing process.

Router(config)#router eigrp 2
Router(config-network)#network 172.16.1.0 0.0.0.255
Router(config-network)#network 172.16.2.0 0.0.0.255
Router(config-network)#network 172.16.3.0 0.0.0.255

Use the following commands to manage and monitor EIGRP.
Command Features
-->
show ip route

View EIGRP-learned routes.

-->
show eigrp neighbors

View neighboring routers from which EIGRP routes can be learned. Lists the IP address of the connected router.

-->
show eigrp interfaces

View the interfaces that are running EIGRP and the number of connected routers.

LAB

You have two routers connected as shown in the network diagram. Router Jujuy has already been configured to share route information using EIGRP with an autonomous system number of 100. Your task in this lab is to configure the Salta router to share routing information using EIGRP with the Jujuy router.
All interfaces have been configured and enabled. Your task is to:
  • Configure the Salta router to share information about all directly-connected routes with the Jujuy router.
  • When you are finished, save your changes.
Tip: To check your work, view the routing table on each router. If successful, each router will have learned about two networks through EIGRP.


Task SummaryActions you were required to perform
  • Run EIGRP on Fa0/0 for AS 100
  • Run EIGRP on Fa0/1 for AS 100
  • Run EIGRP on s0/1/1 for AS 100
Explanation
When configuring EIGRP, all routers that share information must use the same autonomous system number. In this scenario, use router eigrp 100 to configure the Salta router. Add network statements to identify networks on which to run EIGRP.
Use the following commands:

Salta>enable
Salta#config t
Salta(config)#router eigrp 100
Salta(config-router)#network 192.168.1.0
Salta(config-router)#network 192.168.2.0
Salta(config-router)#network 172.17.150.140 0.0.0.3

(Press Ctrl + Z)
Be aware of the following when configuring EIGRP:
  • Using a network statement without a wildcard mask makes an entry using classful network boundaries.
  • To run EIGRP on the Salta s0/1/1 interface, you could have also used the following command: network 172.17.0.0. This classful network entry would match the IP address assigned to the s0/1/1 interface, thereby enabling EIGRP on that interface.
  • You can enable EIGRP on all interfaces on a router using a single command as follows: network 0.0.0.0 255.255.255.255. This wildcard mask value matches every possible network, enabling EIGRP on all IP interfaces.