# EIGRP: Basics and Advanced Configuration

> Chapter 5 of the Courseiva ENCOR curriculum — https://courseiva.com/learn/encor/eigrp-basics-and-advanced

**Official objective:** 3.1 — Configure and verify EIGRP (classic and named mode) for IPv4 and IPv6

## Introduction

Enhanced Interior Gateway Routing Protocol (EIGRP) is an advanced distance-vector routing protocol that Cisco routers use to advertise network reachability data and automatically compute the fastest, loop-free path to every destination. In an enterprise, dozens of routers must constantly share which networks they can reach, and EIGRP does this lightning-fast with minimal chatter, using its DUAL algorithm to avoid routing loops without requiring a central map of the whole network. For the 350-401 ENCOR exam, mastering both classic and named EIGRP for IPv4 and IPv6 is non-negotiable because you will be asked to configure, verify and troubleshoot a protocol that still anchors thousands of real-world corporate networks.

## The Route Optimisation Board Analogy

The Route Optimisation Board at MetroParcel’s central depot is a live map of the entire city’s delivery zones. This board receives constant updates from delivery vans about road closures, traffic jams and newly discovered shortcuts. When Van 7 reports a jam on Oak Street, the board instantly recalculates alternative routes for all other vans heading that way, guaranteeing that no two vans take a path that creates a loop. The board’s internal algorithm, called the Delivery Update Loop Avoidance (DUAL) engine, ensures that every suggested detour is safe and never sends a van back toward the start of the jam. Even better, the board pre-approves several backup routes it calls feasible shortcuts; the moment Oak Street blocks, a van already on the road can switch to a pre-vetted alternative without waiting for a full route recomputation.

In the old ‘Classic’ setup, a technician adjusted physical dials for each delivery zone independently, tweaking speed and road preferences one neighbourhood at a time. The newer ‘Named’ mode replaces all those dials with a single touchscreen tablet, where the fleet manager defines every zone’s policy inside one unified profile. Under the hood, exactly the same DUAL logic runs, but the configuration experience is far cleaner. This is precisely how Enhanced Interior Gateway Routing Protocol (EIGRP) operates inside a business network: classic mode configures each address family separately via global commands, while named mode wraps all settings under one elegant hierarchy.

## Core explanation

EIGRP is a routing protocol that falls into the category of advanced distance-vector protocols, though Cisco often calls it a hybrid protocol because it borrows characteristics from both distance-vector and link-state designs. A routing protocol is the set of rules a router uses to talk to its neighbours and learn about remote networks. In a distance-vector protocol, routers only know the direction (which interface) and distance (metric) to a destination, not the full map of every link in the network. EIGRP enhances that model by maintaining a topology table of all learned routes and by running the Diffusing Update Algorithm (DUAL) to find the shortest path while guaranteeing a loop-free network at all times.

EIGRP uses a composite metric to pick the best path. By default, the metric is calculated from the minimum bandwidth along the path and the cumulative delay. Bandwidth is measured in kilobits per second, and delay is measured in tens of microseconds. The full formula involves several K values; out of the box, K1 and K3 are set to 1, while K2, K4 and K5 are zero, meaning only bandwidth and delay matter. If you change these K values they must match on every router in the same autonomous system, or neighbour relationships will flap.

Every router running EIGRP belongs to an autonomous system (AS), a numeric ID that identifies the group of routers sharing the same EIGRP domain. Routers with different AS numbers will not form neighbour relationships. Neighbour discovery happens through hello packets sent to the multicast address 224.0.0.10 for IPv4 and to FF02::A for IPv6. When two routers agree on the AS number, authentication credentials (if any) and hello/hold timers, they become neighbours. All discovered neighbours appear in the neighbour table. Hello packets contain parameters like the hold time, which tells the neighbour how long to wait before declaring the sender dead. The default hello interval is 5 seconds on LAN links and 60 seconds on low-speed WAN links, with the hold time typically three times the hello interval.

Once neighbours are up, they exchange full routing updates. Every prefix a router learns from a neighbour is stored in the topology table along with the metric it received. The topology table is more than a simple list; it holds all paths to each destination, regardless of whether they are currently used. From this table DUAL selects the best path—called the successor—based on the lowest composite metric. To avoid loops, DUAL requires that any backup path (a feasible successor) must satisfy the feasibility condition: the reported distance (RD, the metric the neighbour says it has to the destination) must be strictly less than the feasible distance (FD, the current best metric from the local router to that destination). If no feasible successor exists, the router enters active state and sends query packets to neighbours to find a new path, temporarily suspending traffic for that prefix until a response arrives.

The routing table (often called the forwarding information base, or FIB) only receives the successor path, plus any feasible successors that the administrator has allowed for unequal-cost load balancing. EIGRP supports equal-cost load balancing by default. To use paths with different metrics, you must configure the variance command: if the feasible successor’s metric is within variance times the FD, that path is also installed.

Classic mode configuration is the traditional way. You enter global configuration with ‘router eigrp 100’, then use network statements to tell the router which interfaces participate in EIGRP. A network statement like ‘network 10.0.0.0’ matches any interface with an IP address in that major classful network, which sometimes leads to unintended interfaces being included unless you use a wildcard mask such as ‘network 10.1.1.0 0.0.0.255’. In classic mode, you must also explicitly disable automatic summarisation with ‘no auto-summary’ if you are running a discontiguous network.

Named mode was introduced to simplify multi-address-family configuration. Everything lives inside a single ‘router eigrp my-process’ block, and you then enter an address-family section—‘address-family ipv4 unicast autonomous-system 100’ for IPv4, or ‘address-family ipv6 unicast autonomous-system 100’ for IPv6. Inside the address family, you can still use network statements, or you can activate EIGRP directly under the interface with the command ‘eigrp my-process’. This mode is the recommended way for new deployments and is the focus of the ENCOR exam.

EIGRP for IPv6 has some quirks. You must first enable IPv6 unicast routing globally with ‘ipv6 unicast-routing’. Then, under each interface, you configure ‘ipv6 eigrp 100’. Inside the named-mode address family, you may need to manually set a router ID with the ‘eigrp router-id’ command if the router has no IPv4 address, because EIGRP traditionally uses a 32-bit router ID that defaults to the highest IPv4 loopback address.

Verification commands differ between the two modes:

- For classic mode: ‘show ip eigrp neighbors’ shows neighbour relationships; ‘show ip eigrp topology’ lists the topology table; ‘show ip route eigrp’ displays routes learned via EIGRP.
- For named mode: the equivalent commands are ‘show eigrp address-family ipv4 neighbors’, ‘show eigrp address-family ipv4 topology’, and the routing table is still queried with ‘show ip route’.

EIGRP also supports MD5 and SHA-256 authentication for neighbour relationships. In named mode, authentication is configured under the interface inside the address-family section with a key chain, making it cleaner than the older ‘ip authentication mode eigrp’ and ‘ip authentication key-chain eigrp’ interface commands.

Other advanced features include stub routing, which limits the types of routes a spoke router advertises, and route summarisation, which can reduce the size of the topology table by sending a single summary prefix to neighbours instead of many specific subnets. In named mode, summarisation is set per interface inside the address-family.

All of these mechanisms make EIGRP an efficient, fast-converging protocol that is well suited for enterprise campus networks where quick failover and flexible design are required.

## Real-world context

A mid-sized retail chain operates three branch offices connected to a head office through a service provider’s MPLS network. The IT department had originally configured static routes on each branch router, pointing traffic toward the head office’s gateway for internet access and shared services. One Tuesday afternoon, the MPLS link between Branch 2 and the head office begins flapping due to a fibre issue. Because static routes do not automatically fail over, the point-of-sale terminals at Branch 2 lose connectivity to the central inventory database until a network engineer manually logs in and changes the next-hop. The resulting 40-minute outage halts sales and triggers a cascade of support tickets.

To prevent a recurrence, the network team decides to deploy EIGRP across all four sites. The engineer logs into each Cisco router via SSH and performs the following steps:

- Enables EIGRP nominated mode with a descriptive process name: ‘router eigrp RETAIL-CORE’.
- Creates an address family with the corporate AS number: ‘address-family ipv4 unicast autonomous-system 100’.
- Activates EIGRP on the appropriate interfaces using the ‘eigrp RETAIL-CORE’ interface command, or supplies a network statement with a wildcard mask to precisely match the LAN and WAN subnets.
- Disables auto-summary and applies an MD5 key chain to the WAN interfaces for secure neighbour authentication.

After committing the configuration, the engineer verifies neighbour adjacencies with ‘show eigrp address-family ipv4 neighbors’. All branch routers now show the head office router as established, and each branch also forms a neighbour relationship with at least one other branch via the MPLS cloud. To test resilience, the engineer intentionally deactivates the primary WAN interface on Branch 2. Within seconds, the routing table on Branch 2 switches to a path through Branch 1 to reach the head office, and point-of-sale sessions resume without human intervention. The topology table viewed with ‘show eigrp address-family ipv4 topology’ confirms that a feasible successor was precomputed and installed because it met the feasibility condition.

In the weeks that follow, the team gradually migrates the configuration to EIGRP for IPv6. They add the ‘ipv6 eigrp’ command on each interface, create the IPv6 address family under the same named process, set a manual router ID where needed, and open the relevant firewall rules for multicast address FF02::A. Now all internal applications that speak IPv6 benefit from the same dynamic failover. The result is a hands-off routing design that shrinks mean time to repair from tens of minutes to zero, directly protecting revenue at the stores.

## Exam focus

The 350-401 ENCOR exam loves EIGRP because it lets Cisco test fundamental routing understanding alongside Cisco-proprietary details. Expect multiple-choice questions, drag-and-drop tasks that ask you to match commands to their output, and configuration-based simulations that require you to spot a misconfigured neighbour or a missing address-family statement.

Key concepts the exam repeatedly tests include the feasibility condition itself. You must be able to analyse a topology table excerpt and answer: ‘Which is the feasible successor?’ The correct answer is always the path whose reported distance (RD) is less than the feasible distance (FD) of the current successor. If all RDs are equal to or greater than the FD, no feasible successor exists, and the network would go active on that prefix. Many candidates confuse FD and RD, so trap answers will swap the definitions.

Another heavily tested area is the difference between classic and named mode verification commands. The exam might show the output of ‘show ip eigrp neighbors’ and then ask why a particular router does not appear—when in fact the configuration uses named mode and the administrator should have used ‘show eigrp address-family ipv4 neighbors’. Know that classic commands like ‘show ip eigrp neighbors’ will not return expected results if the process is named.

IPv6 EIGRP traps are common. The exam will check if you remember that an IPv6 EIGRP process without a router ID cannot form neighbours. A drag-and-drop might ask you to place the necessary commands in order: enable ‘ipv6 unicast-routing’, activate ‘ipv6 eigrp 100’ on the interfaces, create the named process with the IPv6 address family, and optionally set a router ID. Forgetting the ‘ipv6 unicast-routing’ step is a classic error the exam will exploit.

Authentication configuration exhibits subtle differences. In classic mode, you apply the key chain and authentication mode directly on the interface. In named mode, under the interface section inside the address-family, you use the ‘authentication mode md5’ and ‘authentication key-chain MY-KEYS’ commands. The exam expects you to pick the correct syntax for the given mode.

Stub routing is another favourite. You need to know which stub types (connected, summary, static, redistribute) limit what the stub router advertises, and why you would configure a branch router as ‘eigrp stub connected summary’ to preserve bandwidth.

Finally, pay attention to default metric manipulation. The exam may present a scenario where an engineer wants a specific WAN link to be preferred and asks how to change the metric. Since bandwidth and delay are the defaults, you will often adjust the delay value under the interface with ‘delay’ in tens of microseconds. The correct answer pattern is always: modify K-consistent parameters on the local interface and clear the neighbour if needed to force reconvergence.

To summarise the exam danger zones:

- Confusing FD and RD, and misapplying the feasibility condition.
- Using classic verification commands on a named mode configuration.
- Overlooking the mandatory router ID for IPv6 when no IPv4 address exists.
- Misplacing authentication commands between classic and named mode.
- Forgetting to enable ‘ipv6 unicast-routing’ before attempting IPv6 EIGRP.
- Assuming EIGRP load balances over unequal-cost paths without a variance command.

## Step by step

1. **Plan the EIGRP autonomous system number and router ID** — Pick a consistent AS number across all routers that will share EIGRP routes. Decide on a stable router ID (usually the highest IPv4 loopback address) to avoid flapping during IPv6 configurations. This prevents misconfigurations that would block neighbour relationships.
2. **Enter the EIGRP process (classic or named) and address family** — For named mode, issue ‘router eigrp CampusNet’, then ‘address-family ipv4 unicast autonomous-system 100’. For classic mode, use ‘router eigrp 100’. The address family is where you place all network statements and interface instructions; choosing the correct mode shapes which verification commands later work.
3. **Advertise networks using network statements or interface activation** — In classic mode, ‘network 10.0.0.0 0.255.255.255’ enables EIGRP on any interface with a matching IP. In named mode, you can either use the same network command inside the address family, or go to the interface and apply ‘eigrp CampusNet’. This step tells the router which links to use for forming neighbours and exchanging prefixes.
4. **Verify neighbour adjacencies and the topology table** — Run ‘show eigrp address-family ipv4 neighbors’ (named) or ‘show ip eigrp neighbors’ (classic) to confirm hellos are accepted. Then inspect ‘show eigrp address-family ipv4 topology’ to see all learned routes, the FD, and whether any feasible successors exist. This is the moment you catch mismatched AS numbers, authentication failures or incorrect timers.
5. **Apply advanced features such as authentication or stub routing** — Under the address family, configure a key chain and assign it to the interface with ‘authentication mode md5’ and ‘authentication key-chain MyKeys’ so that only trusted peers join. If a branch router has a single exit path, configure ‘eigrp stub connected summary’ to reduce query traffic and prevent it from being used as a transit node. These steps harden and optimise the design.

## Comparisons

### Classic Mode vs Named Mode

**Classic Mode:**
- Entered with ‘router eigrp 100’ at global level; IPv4 and IPv6 require separate processes.
- Verification uses ‘show ip eigrp neighbors’ and similar legacy commands.
- Authentication is applied directly on the interface outside the router process.

**Named Mode:**
- Entered with ‘router eigrp CampusNet’; all address families live inside one process.
- Verification uses ‘show eigrp address-family ipv4 neighbors’ and equivalent family-specific commands.
- Authentication is configured under the interface section inside the address family for a cleaner, unified view.

### Feasible Distance (FD) vs Reported Distance (RD)

**Feasible Distance (FD):**
- The local router’s own best metric to a destination network.
- Used as the benchmark in the feasibility condition.
- Is the value against which all reported distances are compared.

**Reported Distance (RD):**
- The metric the neighbour advertises for the same destination.
- Must be strictly less than the FD for the path to qualify as a feasible successor.
- Shown in the topology table next to the neighbour’s IP address.

### Successor vs Feasible Successor

**Successor:**
- The primary loop-free path with the lowest metric, installed in the routing table.
- Does not require the feasibility condition check because it is already the best path.
- When it fails, EIGRP immediately switches to a feasible successor without going active.

**Feasible Successor:**
- A backup path that satisfies the feasibility condition (RD < FD).
- Also installed in the routing table if variance is configured or equal-cost paths exist.
- If no feasible successor exists, the router must actively query neighbours, delaying convergence.

### EIGRP for IPv4 vs EIGRP for IPv6

**EIGRP for IPv4:**
- Uses multicast address 224.0.0.10 for hello packets.
- Can be enabled with network statements and does not require a forced router ID if an IPv4 interface exists.
- Classic mode commands like ‘show ip eigrp neighbors’ are fully sufficient.

**EIGRP for IPv6:**
- Uses multicast address FF02::A for hello packets.
- Requires ‘ipv6 unicast-routing’ and interface-level ‘ipv6 eigrp <AS>’; a manual router ID may be needed.
- Even in classic mode, verification uses ‘show ipv6 eigrp neighbors’; named mode uses ‘show eigrp address-family ipv6 neighbors’.

## Diagram

_A simplified workflow of EIGRP neighbour formation and route calculation leading to the routing table._

```mermaid
graph LR
    A[Interface IP 10.1.1.1] -->|Hello multicast| B[Neighbour Discovery]
    B --> C[Neighbour Table]
    C --> D[Topology Table (DUAL)]
    D --> E{Feasibility Condition?}
    E -->|Yes| F[Feasible Successor Backup]
    E -->|No| G[Only Successor]
    F --> H[Install in Routing Table]
    G --> H
```

## Common misconceptions

- **Misconception:** EIGRP is a pure link-state protocol like OSPF. **Reality:** EIGRP is an advanced distance-vector protocol that uses the DUAL algorithm; it does not build a complete network map the way OSPF does. (Cisco calls EIGRP a hybrid protocol, leading beginners to assume it works like OSPF. The exam wording reinforces this confusion if you do not grasp the neighbour-table/topology-table model.)
- **Misconception:** Feasible distance and reported distance are the same thing. **Reality:** Feasible distance (FD) is the local router’s best metric to a destination; reported distance (RD) is the metric the next-hop neighbour claims it has to that destination. (The terms sound similar and are often mixed up because both appear in ‘show ip eigrp topology’ output. The feasibility condition hinges on comparing them correctly.)
- **Misconception:** Named mode EIGRP does not require an autonomous system number. **Reality:** You must still specify an AS number inside the address-family command, e.g. ‘address-family ipv4 unicast autonomous-system 100’, even though the router process itself is named. (The ‘named’ aspect refers to the process label, not the AS. Many learners skip the AS part and then wonder why neighbours never form.)
- **Misconception:** EIGRP for IPv6 uses network statements like IPv4 to enable interfaces. **Reality:** IPv6 EIGRP is activated on interfaces with ‘ipv6 eigrp <AS>’ and does not use network statements under the address family. (Most IPv4 EIGRP examples rely on the network command, so when moving to IPv6, the same habit leads to configuration that silently does nothing.)

## Key takeaways

- EIGRP uses the DUAL algorithm to guarantee loop-free paths by enforcing the feasibility condition (reported distance < feasible distance).
- A successor is the primary loop-free path with the lowest metric; a feasible successor is a precomputed backup that also satisfies the feasibility condition.
- Classic EIGRP is configured with ‘router eigrp <AS>’ and network statements, while named mode nests all parameters under ‘router eigrp <name>’ with address-family sections.
- To verify EIGRP neighbours in named mode for IPv4, use ‘show eigrp address-family ipv4 neighbors’ — classic commands will not display the expected output.
- EIGRP for IPv6 requires ‘ipv6 unicast-routing’ and the interface command ‘ipv6 eigrp <AS>’, plus an explicit router ID if no IPv4 address exists on the router.
- Unequal-cost load balancing only happens when you configure the variance command; without it, EIGRP uses only equal-cost paths.

## FAQ

**What is the difference between EIGRP classic mode and named mode?**

Classic mode uses ‘router eigrp <AS>’ at the global level and requires separate configuration for IPv4 and IPv6 on different lines. Named mode packages everything under ‘router eigrp <name>’ with distinct address-family sections, making multi-address-family management cleaner and providing a single entry point for verification commands.

**How do I configure EIGRP for IPv6 and what are the key commands?**

First enable IPv6 routing with ‘ipv6 unicast-routing’. On each interface, apply ‘ipv6 eigrp <AS>’. Then create a named process (‘router eigrp NAME’), enter ‘address-family ipv6 unicast autonomous-system <AS>’ and, if needed, set a router ID with ‘eigrp router-id’. Verify with ‘show eigrp address-family ipv6 neighbors’.

**What is the feasibility condition and why does it matter?**

The feasibility condition states that a route can be a feasible successor only if the reported distance (the neighbour’s metric) is less than the feasible distance (the local best metric) of the current successor. It guarantees a loop-free backup path without any recalculation, so EIGRP can fail over instantly without sending queries.

**How do I check if EIGRP neighbours are formed?**

For classic mode, use ‘show ip eigrp neighbors’. For named mode, use ‘show eigrp address-family ipv4 neighbors’ or the IPv6 equivalent. Look for the ‘address’ of the neighbour and the uptime; if the state is not established, check AS numbers, authentication keys, and whether interfaces are in the same subnet.

**What does the ‘variance’ command do in EIGRP?**

Variance multiplies the feasible distance of the successor, and any feasible successor with a metric less than that product becomes eligible for unequal-cost load balancing. For example, if the successor has FD=100 and variance 2, a feasible successor with a metric up to 200 will also be installed in the routing table.

**Can I run multiple EIGRP autonomous systems on the same router?**

Yes, you can run separate EIGRP processes—either classic or named—with different AS numbers. Routes are not automatically shared between processes unless you explicitly redistribute them. This lets you segment routing domains, for instance keeping a corporate AS distinct from a test lab AS.

## Check your understanding

1. **Which command displays EIGRP neighbours in IPv4 for a named mode process called BranchNet?**

   Answer: ’show eigrp address-family ipv4 neighbors’ will show the neighbour table for the BranchNet named process.

2. **A topology table excerpt shows a route to 10.0.0.0/24 with FD=3072 and a neighbour advertising RD=2816. Is this neighbour a feasible successor?**

   Answer: Yes, because the reported distance (2816) is less than the feasible distance (3072), satisfying the feasibility condition.

3. **What multicast address does EIGRP for IPv6 use for hello packets?**

   Answer: FF02::A.

4. **In classic mode, what global command starts an EIGRP process using autonomous system 200?**

   Answer: router eigrp 200

5. **Which two values make up the default EIGRP composite metric?**

   Answer: Minimum bandwidth along the path (in kbps) and cumulative interface delay (in tens of microseconds).

---

Interactive version with quiz and diagrams: https://courseiva.com/learn/encor/eigrp-basics-and-advanced