EIGRP Overview – Enhanced Interior Gateway Routing Protocol

1. What Is EIGRP?

EIGRP (Enhanced Interior Gateway Routing Protocol) is a Cisco-developed interior gateway protocol (IGP) designed for routing within a single autonomous system. It was originally introduced in 1992 as a proprietary successor to IGRP, and in 2013 Cisco published an informational RFC (7868) allowing other vendors to implement it — though it remains most commonly found in Cisco-only deployments.

EIGRP is classified as an advanced distance-vector (or hybrid) protocol. It shares routing information only with directly connected neighbours (distance-vector characteristic), but also maintains a topology database and uses a loop-free path-selection algorithm (link-state characteristic). This combination gives it the simplicity of distance-vector protocols and much of the scalability of link-state protocols.

Why "hybrid"? EIGRP takes the best of both worlds: from distance-vector it inherits simplicity (no area design required, no full topology flood) and from link-state it borrows the topology database and the DUAL algorithm for guaranteed loop-free convergence. It does not compute a full Shortest Path Tree like OSPF, but it does avoid the count-to-infinity problem and slow convergence of traditional distance-vector protocols.

2. Why EIGRP Exists — The Problems It Solves

To understand EIGRP, it helps to understand the limitations of the protocols that came before it:

Protocol	Limitation	How EIGRP Addresses It
RIP	Maximum 15 hops — cannot scale to large networks. Slow convergence (up to 3 minutes). Uses only hop count — ignores bandwidth completely.	No hop count limit. Sub-second convergence via DUAL. Uses bandwidth and delay for metric — a 10 Mbps link is preferred over a 56 kbps link. See: RIP Concepts
IGRP	Classful only (no VLSM/CIDR). Still slow convergence using Bellman-Ford. No loop prevention beyond split-horizon.	Full classless support. DUAL provides guaranteed loop-free convergence. Feasible Successor enables instant failover without re-computation.
OSPF	Complex area design required for large networks. All routers must compute the full SPF tree. Equal-cost load balancing only.	No area structure needed. DUAL computes paths without full topology. Unique unequal-cost load balancing via variance. See: OSPF Overview

3. Core EIGRP Features

Feature	Detail	Significance
DUAL Algorithm	Diffusing Update Algorithm — maintains loop-free paths and enables instant failover using pre-computed backup routes	Core differentiator from all other distance-vector protocols
Rapid Convergence	When a Feasible Successor exists, failover is instantaneous — no queries, no waiting. Even without FS, DUAL converges faster than RIP or OSPF SPF recalculation in most topologies	Critical for enterprise networks with high availability requirements
Classless Routing	Supports VLSM (Variable-Length Subnet Masking) and CIDR — subnet mask is included in all routing updates	Essential in modern IPv4 networks with complex addressing
Composite Metric	Default: bandwidth + delay. Optional: load, reliability, MTU. More representative of real path quality than simple hop count	Selects genuinely better paths, not just shorter ones
Unequal-Cost Load Balancing	Via `variance` command — traffic can be split across paths with different metrics, proportional to each path's quality	Unique to EIGRP — no other IGP supports this natively
Partial and Bounded Updates	Updates are sent only when topology changes occur (not periodically), and only to affected neighbours — not broadcast to the whole network	Dramatically reduces bandwidth consumption vs RIP's periodic full-table broadcasts
Multiprotocol Support	Supports IPv4, IPv6 (EIGRPv6), and legacy protocols via PDMs (Protocol Dependent Modules)	Single protocol family for dual-stack environments
RTP (Reliable Transport Protocol)	EIGRP's own transport layer running over IP (protocol 88). Provides reliable ordered delivery for Update, Query, Reply packets without using TCP	Reliable delivery without the overhead of TCP connection establishment

4. The Three EIGRP Tables

EIGRP maintains three separate tables in memory. Understanding each and what it contains is essential for both CCNA exams and real-world troubleshooting.

  ┌──────────────────────────────────────────────────────────────┐
  │  1. NEIGHBOUR TABLE                                          │
  │     Adjacent routers and their current status               │
  │     Built and maintained via Hello packets                  │
  │     show ip eigrp neighbors                                 │
  └────────────────────────┬─────────────────────────────────────┘
                           │ feeds into
  ┌────────────────────────▼─────────────────────────────────────┐
  │  2. TOPOLOGY TABLE                                           │
  │     ALL paths to ALL destinations learned from neighbours    │
  │     Includes Successors and Feasible Successors              │
  │     DUAL runs here to select the best paths                  │
  │     show ip eigrp topology                                   │
  └────────────────────────┬─────────────────────────────────────┘
                           │ best paths installed
  ┌────────────────────────▼─────────────────────────────────────┐
  │  3. ROUTING TABLE                                            │
  │     Only the Successor routes (lowest FD per destination)    │
  │     Used for actual packet forwarding decisions              │
  │     show ip route eigrp (shown as "D" in routing table)      │
  └──────────────────────────────────────────────────────────────┘

What Each Table Contains

Table	Key Contents	Verification Command
Neighbour Table	Neighbour IP, interface, hold time, uptime, SRTT, RTO, Q count, sequence number	`show ip eigrp neighbors`
Topology Table	All learned prefixes, each with FD, and via-entries showing each neighbour's RD and the local FD via that neighbour. State (P/A).	`show ip eigrp topology`
Routing Table	Best path (Successor) per destination — installed for forwarding. Shows "D" for internal EIGRP, "D EX" for external EIGRP routes.	`show ip route eigrp` — see show ip route for output interpretation

5. The DUAL Algorithm — How EIGRP Achieves Loop-Free Convergence

DUAL (Diffusing Update Algorithm) is the mathematical engine at the heart of EIGRP. It was designed by Dr. J.J. Garcia-Luna-Aceves specifically to provide guaranteed loop-free routing at every instant during convergence — something traditional distance-vector protocols cannot achieve.

DUAL Key Concepts

Term	Definition	Where Used
Feasible Distance (FD)	The best known metric from this router to a destination — the lowest total cost ever computed for this prefix. Used as the benchmark for the Feasibility Condition.	Stored per-prefix in the topology table
Reported Distance (RD)	The metric a neighbour advertises for a destination — what the neighbour says it costs to reach that prefix from the neighbour's perspective.	Received in EIGRP Update packets; stored in topology table
Successor	The neighbour with the lowest total cost (FD) to a destination. The Successor's route is installed in the routing table as the primary forwarding path.	Routing table + topology table
Feasible Successor (FS)	A backup neighbour whose RD is strictly less than the current FD — it is mathematically guaranteed not to create a loop, because it cannot be routing traffic through us.	Topology table only — used immediately when Successor fails
Feasibility Condition (FC)	A neighbour's Reported Distance must be strictly less than the current Feasible Distance: RD < FD. This is the mathematical guarantee of loop freedom.	Evaluated by DUAL for every candidate backup path

Why the Feasibility Condition Guarantees Loop Freedom

The logic is elegant: if a neighbour's cost to reach a destination (RD) is less than our total cost to that destination (FD), then that neighbour cannot possibly be routing traffic through us to get there. If it were routing through us, its metric would be greater than ours — not less. Therefore, making that neighbour our next-hop creates no loop.

  Topology: R1 → R2 → R3 → 10.0.0.0/24

  R1's view:
  ┌──────────────────────────────────────────────────────┐
  │ Route to 10.0.0.0/24 via R2:                        │
  │   R2's Reported Distance (RD) = 50                  │
  │   R1's link cost to R2        = 20                  │
  │   R1's Feasible Distance (FD) = 50 + 20 = 70        │
  │                                                      │
  │ Backup path via R4:                                  │
  │   R4's RD = 40                                       │
  │   Is R4's RD (40) < R1's FD (70)? YES ✓             │
  │   → R4 is a Feasible Successor (loop-free backup)   │
  │                                                      │
  │ Another candidate via R5:                            │
  │   R5's RD = 80                                       │
  │   Is R5's RD (80) < R1's FD (70)? NO ✗              │
  │   → R5 is NOT a Feasible Successor                  │
  │   (R5 might be routing through R1 — using it        │
  │    as a next-hop could create a loop)                │
  └──────────────────────────────────────────────────────┘

DUAL Convergence States

State	Meaning	What Happens
Passive (P)	Route is stable — a Successor is known and installed	Normal operation. Packets forwarded via Successor.
Active (A) — with FS	Successor failed, but a Feasible Successor exists	Instantaneous failover — FS promoted to Successor without any queries. Route returns to Passive immediately.
Active (A) — no FS	Successor failed, no Feasible Successor — DUAL runs distributed computation	Router sends QUERY packets to all neighbours. Must wait for all REPLY packets before convergence. Route stays Active. If no reply in 3 min → SIA.

6. EIGRP Metric — Composite Calculation

EIGRP's composite metric combines multiple path characteristics, weighted by K-values. By default only bandwidth and delay contribute (K1=1, K3=1; K2=K4=K5=0).

Full formula:
Metric = [K1 × (10⁷ / min_bandwidth_kbps) + K3 × (cumulative_delay_µs / 10)] × 256

Default (K1=1, K3=1, all others=0):
Metric = [(10⁷ / min_bandwidth_kbps) + (cumulative_delay_µs / 10)] × 256

How Each Component Is Measured

Component	How Measured Along a Path	Interface Command to View
Bandwidth	Takes the minimum (bottleneck) bandwidth across all links in the path — one slow link limits the entire path's metric contribution	`show interfaces Gi0/0` → "BW" field (kbps)
Delay	Takes the cumulative sum of all interface delays across the entire path — unlike bandwidth, delay accumulates at every hop	`show interfaces Gi0/0` → "DLY" field (µs)

Worked Example

Path: R1 → R2 → R3 → 10.0.0.0/24

Link R1–R2: GigabitEthernet (BW 1,000,000 kbps, delay 10 µs)
Link R2–R3: FastEthernet    (BW 100,000 kbps,   delay 100 µs)
Link R3–destination: Serial (BW 1,544 kbps,     delay 20,000 µs)

min_bandwidth  = 1,544 kbps          (Serial link is the bottleneck)
cumulative_delay = 10 + 100 + 20,000 = 20,110 µs

Metric = [(10,000,000 / 1,544) + (20,110 / 10)] × 256
       = [6,476 + 2,011] × 256
       = 8,487 × 256
       = 2,172,672

! The slow Serial link dominates the metric through both bandwidth and delay

Tuning tip: Because bandwidth is the single most impactful component (minimum on path), engineers often adjust the configured bandwidth on interfaces to influence EIGRP path selection without changing actual link speed. Use bandwidth <kbps> under the interface. Never set it lower than the actual speed for QoS purposes, but it can be set to reflect WAN circuit speeds on serial interfaces that default to 1.544 Mbps regardless of contracted speed.

7. Neighbour Discovery and Adjacency

EIGRP neighbours discover each other by sending Hello packets out EIGRP-enabled interfaces using multicast address 224.0.0.10. A neighbour relationship forms when all conditions are satisfied.

Adjacency Requirements Summary

Requirement	Must Match?	What Happens if Wrong
Autonomous System (AS) number	Yes — must be identical	Hellos silently ignored — no adjacency, no error logged
K-values	Yes — must be identical	"K-value mismatch" logged — adjacency rejected
Shared subnet	Yes — must be Layer 3 reachable neighbours	Hellos don't reach each other
Authentication	Yes — if enabled, must match on both sides	Hellos silently dropped — no adjacency
Hello and Hold timers	No — do not need to match	Each router uses its neighbour's advertised hold time

Hello and Hold Timer Defaults

Interface Type	Hello Interval	Hold Time	Notes
LAN (GigabitEthernet, FastEthernet)	5 seconds	15 seconds	Hold time = 3 × hello by default
WAN / NBMA (Serial, Frame Relay)	60 seconds	180 seconds	Longer timers for slower, less reliable links

Timer mismatch is allowed: This is a common exam trap comparing EIGRP to OSPF. In OSPF, Hello and Dead timers must match for adjacency. In EIGRP, they do not — each router advertises its own hold time in Hello packets, and the neighbour honours the received hold time for that specific adjacency.

8. EIGRP Packet Types and RTP

EIGRP uses its own transport protocol, RTP (Reliable Transport Protocol), running directly over IP (protocol number 88). RTP provides selective reliability — some packet types are delivered reliably (with acknowledgment), others are not.

Packet Type	Direction	Purpose	Reliable (ACK required)?
Hello	Router → All EIGRP neighbours (multicast)	Discover and maintain neighbour relationships; heartbeat mechanism	No — sent periodically without confirmation
Update	Router → Neighbours (multicast initially, then unicast retransmit)	Carry routing information — sent only when changes occur, not periodically	Yes — must be acknowledged
Query	Router → All neighbours (multicast)	Ask "do you have a path to X?" when a route is lost and no FS exists	Yes — router waits for Reply from every neighbour queried
Reply	Neighbour → Querying router (unicast)	Respond to Query with either a path or "I have no path either"	Yes — querying router acknowledges each Reply
Acknowledgment (ACK)	Unicast back to sender	Confirm receipt of reliable packets; sent as a Hello with no data	N/A — ACKs are not themselves acknowledged

9. Administrative Distance

Administrative Distance (AD) is used to select between routes learned from different sources when they point to the same destination. Lower AD = more trusted = preferred. EIGRP distinguishes between internally learned routes and redistributed (external) routes.

Route Source	AD	Routing Table Code	Notes
EIGRP Internal	90	`D`	Routes learned natively within the EIGRP AS
EIGRP External	170	`D EX`	Routes redistributed into EIGRP from another protocol (OSPF, static, etc.)
OSPF	110	`O`	Internal EIGRP (90) beats OSPF (110)
RIP	120	`R`	EIGRP internal beats RIP by a wide margin
Static route	1	`S`	Static beats everything except directly connected (0)

Practical AD example: A router runs both EIGRP and OSPF, and both protocols learn a route to 10.100.0.0/24. EIGRP offers it at AD 90 and OSPF at AD 110. EIGRP wins — 90 < 110 — and the EIGRP route is installed. If EIGRP were to redistribute that route from OSPF (making it EIGRP external), it would have AD 170 and lose to the native OSPF route (AD 110). The AD difference between internal (90) and external (170) is intentional — it prevents redistributed routes from displacing natively learned routes.

10. EIGRP vs OSPF vs RIP — Comparison

Feature	EIGRP	OSPF	RIP v2
Protocol type	Advanced distance-vector (hybrid)	Link-state	Distance-vector
Algorithm	DUAL	Dijkstra SPF	Bellman-Ford
Metric	Composite (BW + Delay by default)	Cost (based on bandwidth)	Hop count (max 15)
Administrative Distance	90 (internal), 170 (external)	110	120
Convergence speed	Very fast — instant when FS exists	Fast (SPF recalculation)	Slow (up to 3 minutes)
Update type	Partial, bounded (change-triggered only)	Partial LSA flooding	Full table every 30 seconds
Load balancing	Equal AND unequal-cost (via variance)	Equal-cost only	Equal-cost only
Area design required?	No — flat AS structure	Yes — backbone area 0 required for multi-area	No
Topology database?	Yes — topology table (not full link-state database)	Yes — full LSDB	No — only routing table
Transport	IP protocol 88 (RTP)	IP protocol 89	UDP port 520 — see Common Port Numbers
Multicast address	224.0.0.10	224.0.0.5 (all OSPF) / 224.0.0.6 (DR/BDR)	224.0.0.9
Vendor support	Primarily Cisco (RFC 7868 for others)	Open standard — universal	Open standard — universal

11. Key EIGRP Terminology Reference

Term	Definition	Exam Significance
AS (Autonomous System)	A logical group of routers sharing the same EIGRP process number. Routers with different AS numbers cannot become neighbours.	Must match on both sides of every adjacency
RTP (Reliable Transport Protocol)	EIGRP's own delivery mechanism running over IP protocol 88. Provides reliable delivery for Update/Query/Reply without using TCP.	Explains how EIGRP achieves reliable updates without TCP
DUAL	Diffusing Update Algorithm — selects loop-free paths and provides instant failover when a Feasible Successor exists.	The core differentiator from all other distance-vector protocols
FD (Feasible Distance)	Total metric from this router to a destination — our link cost plus the neighbour's cost. The lowest FD is the Successor.	Basis for the Feasibility Condition comparison
RD (Reported Distance)	The metric a neighbour advertises for a destination — the neighbour's own cost to reach it. Also called Advertised Distance (AD).	Compared against FD to evaluate Feasibility Condition
Successor	The neighbour providing the best (lowest FD) path to a destination. Installed in the routing table. Only one Successor per destination.	Primary forwarding path in the routing table
Feasible Successor	A backup neighbour satisfying the Feasibility Condition (RD < FD). Stored only in the topology table — used instantly if Successor fails.	Enables sub-second convergence without queries
Feasibility Condition (FC)	RD must be strictly less than the current FD. Guarantees the backup path is loop-free.	The single most frequently tested EIGRP concept on CCNA exams
Passive state (P)	Route is stable — a Successor is known. Normal operating condition.	All routes should be Passive in a stable network
Active state (A) / SIA	Route is being recomputed — DUAL sent queries and is waiting for replies. Stuck In Active (SIA) if no reply after 3 minutes.	SIA indicates a problem — missing stub config or unreachable neighbours
Variance	Multiplier that allows Feasible Successors with FD ≤ variance × best FD to be installed for unequal-cost load balancing.	Unique EIGRP capability — no other IGP supports this

12. When to Use EIGRP

EIGRP is the right choice in specific scenarios. Understanding when to choose it over OSPF is a common CCNA exam topic:

Scenario	EIGRP Preferred?	Reason
All-Cisco network requiring fastest convergence	Yes	DUAL with pre-computed FS provides near-instant failover; no SPF recalculation needed when FS exists
Hub-and-spoke WAN with unequal bandwidth links	Yes	Variance enables traffic distribution proportional to link speed — a 100 Mbps link carries more traffic than a 10 Mbps link automatically
Multi-vendor network environment	No — prefer OSPF	EIGRP's RFC 7868 is informational — not all vendors implement it. OSPF is universally supported
Large enterprise campus with complex topology	Consider OSPF	OSPF's area structure provides better scalability and summarisation control at network boundaries. EIGRP works but requires careful route summarisation to limit query scope
Simple Cisco-only branch with single exit	Yes	Easy configuration, fast convergence, minimal overhead for single-exit (stub) sites

13. Common Misconceptions About EIGRP

"EIGRP timers must match between neighbours."
False — this is OSPF. EIGRP Hello and Hold timers do not need to match. Each router advertises its own hold time in Hello packets, and the neighbour respects the received value. Mismatched timers do not prevent adjacency.
"A Feasible Successor is the second-best path by metric."
Not necessarily. A path qualifies as a Feasible Successor based on the Feasibility Condition (RD < FD), not simply by having the second-lowest FD. A path with the second-lowest FD may fail the FC, and a path with the third-lowest FD may pass it. It is possible to have no Feasible Successor at all, or multiple FSs.
"EIGRP floods updates to all routers."
EIGRP sends partial, bounded updates — only to directly connected neighbours, only when something changes, and only the changed prefixes (not the full routing table). This is far more efficient than RIP's periodic full-table broadcasts and more targeted than OSPF's LSA flooding.
"A higher variance number means better routing."
Variance simply expands the set of FS paths eligible for load balancing. A high variance can include paths with significantly worse metrics, potentially causing traffic to traverse poor links. Use variance conservatively — variance 2 is typical; higher values should be used with care.
"EIGRP and OSPF can share the same network without issues."
Running both simultaneously requires redistribution, which introduces AD considerations and potential routing loops if not carefully designed with route-maps and filters. Mutual redistribution between EIGRP and OSPF is a complex advanced topic — for CCNA, understand that EIGRP internal (AD 90) beats OSPF (AD 110) on the same router without redistribution.

14. Key Points & Exam Tips

EIGRP = Cisco-proprietary advanced distance-vector (hybrid) protocol; uses DUAL algorithm for loop-free convergence.
IP protocol 88; multicast 224.0.0.10; transport = RTP (not TCP/UDP).
Three tables: Neighbour (adjacent routers), Topology (all paths), Routing (best path — Successor).
FD = total metric from this router to destination. RD = neighbour's advertised cost to that destination.
Feasibility Condition: RD < FD → qualifies as Feasible Successor. Guarantees loop-free backup path.
Successor fails + FS exists → instant failover (no queries). No FS → Active state, DUAL sends Queries, waits for Replies.
Default metric: bandwidth (bottleneck) + delay (cumulative) × K-values. K-values must match on both sides for adjacency.
AD: 90 internal, 170 external. Beats OSPF (110) and RIP (120).
Hello/Hold timers: LAN = 5s/15s; WAN = 60s/180s. Do not need to match (unlike OSPF).
Variance = unequal-cost load balancing — unique to EIGRP among IGPs.
Routing table shows EIGRP internal as D; external as D EX.

15. EIGRP Overview Quiz

1. A route to 172.16.0.0/24 in EIGRP has the following topology table entries:
FD = 300,000
Via R2: (300,000 / 200,000)
Via R3: (350,000 / 250,000)
Via R4: (400,000 / 280,000)
Which neighbours qualify as Feasible Successors?

A. R3 and R4 — because they have higher FD than R2 and can serve as backups B. R3 and R4 both fail the Feasibility Condition; neither is a Feasible Successor. Only R2 is the Successor. No FS exists for this route. C. R3 qualifies (RD 250,000 < FD 300,000) but R4 does not (RD 280,000 < FD 300,000 — wait, both qualify) D. All three are Feasible Successors because all have RD less than the FD of 300,000

Correct answer is B. Wait — let's re-evaluate carefully. The Feasibility Condition is: RD < current best FD. The Successor is R2 with FD=300,000. Via R3: RD=250,000 — is 250,000 < 300,000? YES. Via R4: RD=280,000 — is 280,000 < 300,000? YES. So actually both R3 and R4 satisfy the FC. The correct answer should be that R3 and R4 both qualify as Feasible Successors. Reviewing the options: option D says "all three are FS" — R2 is the Successor, not an FS. Option C is partially correct in reasoning but the wording trails off. Option B is the trap answer claiming none qualify. The actual correct interpretation: R3 (RD 250,000) and R4 (RD 280,000) both satisfy RD < FD (300,000), making them both Feasible Successors. Re-read: answer B is wrong. The correct response is that R3 and R4 are both FSs — closest to the intent of option C's partial reasoning. For exam purposes: apply the FC rigorously to each candidate — if RD < best FD, it qualifies as FS.

1. (Corrected) A route has FD = 300,000 and Successor via R2 (FD/RD = 300,000/200,000). Candidate via R5 has FD/RD = 320,000/310,000. Does R5 qualify as a Feasible Successor?

A. Yes — R5's FD (320,000) is close to the best FD (300,000) so it qualifies B. Yes — R5 qualifies because it has a backup path to the same destination C. Yes — configure variance 2 to include R5 D. No — R5's RD (310,000) is NOT less than the current FD (300,000), so the Feasibility Condition fails; R5 is not a Feasible Successor

Correct answer is D. The Feasibility Condition requires RD < FD (strictly less than). R5's Reported Distance is 310,000. The current Feasible Distance is 300,000. Is 310,000 < 300,000? NO — the FC fails. R5 is NOT a Feasible Successor. This means R5 could potentially be routing traffic through R1 to reach the destination (R5's metric is worse than R1's, which is consistent with R5 using R1 as a transit hop). Using R5 as a next-hop in this case could create a routing loop. Variance does not override the Feasibility Condition — variance only allows FSs (that already satisfy FC) to participate in load balancing.

2. What makes EIGRP's convergence faster than OSPF in a network where Feasible Successors exist?

A. EIGRP uses shorter Hello timers than OSPF by default B. EIGRP sends full routing table updates more frequently than OSPF C. When a Feasible Successor exists, EIGRP instantly promotes it to Successor without sending any queries — no SPF recalculation required D. EIGRP uses TCP for faster reliable delivery than OSPF's IP protocol 89

Correct answer is C. EIGRP's convergence advantage over OSPF comes specifically from the pre-computed Feasible Successor. When a Successor fails and a FS exists, the FS is immediately promoted to Successor in the local topology table without sending any Query packets to neighbours and without any distributed computation. The route goes from Active immediately back to Passive. OSPF, by contrast, must wait for LSA flooding to propagate the failure information and then run the full Dijkstra SPF algorithm on every router in the area. In a large network, EIGRP with FS is typically faster. Without FS, EIGRP's Active state (sending queries, waiting for replies) can actually be slower than OSPF's SPF recalculation.

3. A network engineer notices that `show ip eigrp topology` shows a route marked "A" (Active) for several minutes. What is the most likely cause and risk?

A. The Successor failed and no Feasible Successor exists — DUAL is waiting for Query replies. If a neighbour doesn't reply within 3 minutes, EIGRP will declare it Stuck In Active (SIA) and reset the adjacency B. Active means the route is actively being used for forwarding — this is normal C. The route is being summarised and the Active state shows the summarisation is in progress D. Active means EIGRP authentication is verifying the route's origin

Correct answer is A. In EIGRP topology table notation, "P" = Passive (stable, Successor known — normal) and "A" = Active (DUAL running — problem). A route goes Active when its Successor is lost AND no Feasible Successor exists. DUAL floods Query packets to all neighbours asking for alternative paths. The router stays in Active state until all queried neighbours reply. If a neighbour fails to reply within 3 minutes (the SIA timer), EIGRP tears down the adjacency to that neighbour. Use show ip eigrp topology active to see which neighbours haven't replied. Fix by configuring stub routing on edge routers to limit query scope.

4. Why does EIGRP use the minimum bandwidth along a path (rather than the average or sum) as part of its metric calculation?

A. Using minimum is faster to calculate than average B. The minimum is the most recent bandwidth measured on the path C. The slowest link (bottleneck) determines the maximum throughput a flow can achieve on that path — advertising anything higher would be misleading about actual path capacity D. Minimum bandwidth is used because delay uses the sum, and using both sums would double-count the path cost

Correct answer is C. A path's actual throughput is limited by its slowest link — the bottleneck. If a path goes through a 1 Gbps link and a 1 Mbps link, no flow on that path can exceed 1 Mbps regardless of the high-speed segment. Using the minimum bandwidth in the metric accurately reflects this real-world capacity constraint, allowing EIGRP to prefer paths with genuinely higher throughput capacity. Delay, by contrast, accumulates at every hop (queuing, propagation, serialisation) so it is correctly summed. This asymmetry — minimum for bandwidth, sum for delay — is deliberate and accurate modelling of path characteristics.

5. What happens to EIGRP Hello timers if two neighbours have different Hello interval settings — for example, R1 sends Hellos every 5 seconds and R2 sends Hellos every 10 seconds?

A. The adjacency fails — both routers must use the same Hello interval for EIGRP to work (same as OSPF) B. The adjacency forms normally — EIGRP timers do not need to match. Each router sends Hellos at its own interval; each respects the hold time received in the neighbour's Hello packets C. The router with the faster timer (R1) adopts the slower timer (10s) to synchronise D. The adjacency forms but operates at the slower timer rate (10s) for both routers

Correct answer is B. This is a frequently tested contrast between EIGRP and OSPF. OSPF requires Hello and Dead timers to match on both sides of an adjacency — mismatched timers prevent adjacency formation. EIGRP does not have this requirement. Each router sends Hellos at its own configured interval and advertises its own hold time value in the Hello packet header. The neighbour reads the hold time from the received Hello and uses that value as the timeout for that specific adjacency. R1 with 5s Hellos and a 15s hold time can form a fully functional adjacency with R2 using 10s Hellos and 30s hold time.

6. An engineer needs to choose a routing protocol for a new enterprise network that includes Cisco, Juniper, and Arista routers. Which protocol is the best choice and why?

A. EIGRP — it is the fastest protocol and supports unequal-cost load balancing B. EIGRP — RFC 7868 made it fully open-standard so all vendors support it equally C. RIP — the simplest protocol that all vendors support D. OSPF — it is a true open standard universally supported by all vendors including Cisco, Juniper, and Arista, unlike EIGRP which remains primarily Cisco-centric despite RFC 7868

Correct answer is D. While RFC 7868 (2015) published EIGRP specifications, it is an informational RFC rather than a standards-track RFC, and adoption by other vendors has been limited. Juniper, Arista, and most non-Cisco vendors do not implement EIGRP. OSPF (RFC 2328 for OSPFv2, RFC 5340 for OSPFv3) is a true open standard supported uniformly across all major routing vendors. In any multi-vendor environment — which is extremely common in enterprise networks — OSPF is the correct choice. EIGRP should only be chosen when the network is entirely or predominantly Cisco, and when specific EIGRP features (unequal-cost load balancing, simpler configuration) justify the vendor lock-in.

7. A router running EIGRP has a route for 10.0.0.0/24 with FD = 200,000 via Successor R2. Another router R3 advertises this same prefix with RD = 180,000, giving a local FD via R3 of 220,000. Does R3 qualify as a Feasible Successor, and can it be used for unequal-cost load balancing with `variance 2`?

A. Yes to both — R3 satisfies FC (RD 180,000 < FD 200,000) making it an FS, and its FD (220,000) is within variance 2 of best FD (2 × 200,000 = 400,000), so it can participate in unequal-cost load balancing B. R3 is an FS but cannot be used for load balancing — variance only applies to equal-cost paths C. R3 fails the FC because its FD (220,000) is higher than R2's FD (200,000) D. R3 cannot be used — variance only works with equal-metric paths

Correct answer is A. Two conditions must be checked: (1) Feasibility Condition: R3's RD (180,000) must be < best FD (200,000). 180,000 < 200,000 — YES, R3 is a Feasible Successor. (2) Variance check: R3's FD (220,000) must be ≤ variance × best FD = 2 × 200,000 = 400,000. 220,000 ≤ 400,000 — YES, R3 qualifies for unequal-cost load balancing. With variance 2 configured, both R2 (FD 200,000) and R3 (FD 220,000) are installed in the routing table. Traffic is distributed inversely proportional to metric — R2 carries more traffic because it has a lower (better) metric.

8. What is the routing table code for an EIGRP external route, and what does "external" mean in this context?

A. Code "E" — a route learned from a neighbour outside the local subnet B. Code "D" — same as internal routes; there's no distinction C. Code "D EX" with AD 170 — a route that was redistributed into EIGRP from another routing protocol (e.g., OSPF, static, or connected routes on another router) D. Code "EX" with AD 90 — routes learned across WAN links are classified as external

Correct answer is C. EIGRP distinguishes between internal routes (learned natively within the EIGRP autonomous system, AD 90, shown as "D") and external routes (redistributed into EIGRP from another source — OSPF, RIP, static routes, connected interfaces on a redistributing router, AD 170, shown as "D EX"). The high AD (170) for external routes is intentional — it makes them less trusted than natively learned EIGRP routes, preventing redistributed routes from displacing better internal paths. The "D" in both codes stands for DUAL/EIGRP (EIGRP's predecessor was called IGRP, which used "I"; EIGRP uses "D" for DUAL).

9. Which statement correctly describes how EIGRP differs from traditional distance-vector protocols like RIP in terms of update behaviour?

A. EIGRP sends full routing table updates every 90 seconds instead of RIP's 30 seconds B. EIGRP sends partial, bounded, change-triggered updates only to directly connected neighbours — not periodic full-table broadcasts like RIP's 30-second full table updates C. EIGRP floods updates to all routers in the AS using the same LSA mechanism as OSPF D. EIGRP does not send routing updates — it uses only Hello packets to maintain connectivity

Correct answer is B. This is one of EIGRP's core efficiency advantages. RIP sends its complete routing table to all directly connected RIP neighbours every 30 seconds — wasteful, especially for large tables on slow WAN links. EIGRP sends Update packets only when a topology change occurs (triggered updates), contains only the changed prefixes (partial updates), and sends them only to the directly connected neighbours that need them (bounded updates). After initial convergence, if no changes occur, EIGRP sends no routing updates at all — only periodic Hello packets for neighbour maintenance. This dramatically reduces bandwidth consumption and CPU load.

← Back to Home

EIGRP Overview – Enhanced Interior Gateway Routing Protocol

1. What Is EIGRP?

2. Why EIGRP Exists — The Problems It Solves

3. Core EIGRP Features

4. The Three EIGRP Tables

What Each Table Contains

5. The DUAL Algorithm — How EIGRP Achieves Loop-Free Convergence

DUAL Key Concepts

Why the Feasibility Condition Guarantees Loop Freedom

DUAL Convergence States

6. EIGRP Metric — Composite Calculation

How Each Component Is Measured

Worked Example

7. Neighbour Discovery and Adjacency

Adjacency Requirements Summary

Hello and Hold Timer Defaults

8. EIGRP Packet Types and RTP

9. Administrative Distance

10. EIGRP vs OSPF vs RIP — Comparison

11. Key EIGRP Terminology Reference

12. When to Use EIGRP

13. Common Misconceptions About EIGRP

14. Key Points & Exam Tips

15. EIGRP Overview Quiz

1. A route to 172.16.0.0/24 in EIGRP has the following topology table entries: FD = 300,000 Via R2: (300,000 / 200,000) Via R3: (350,000 / 250,000) Via R4: (400,000 / 280,000) Which neighbours qualify as Feasible Successors?

1. (Corrected) A route has FD = 300,000 and Successor via R2 (FD/RD = 300,000/200,000). Candidate via R5 has FD/RD = 320,000/310,000. Does R5 qualify as a Feasible Successor?

2. What makes EIGRP's convergence faster than OSPF in a network where Feasible Successors exist?

3. A network engineer notices that show ip eigrp topology shows a route marked "A" (Active) for several minutes. What is the most likely cause and risk?

4. Why does EIGRP use the minimum bandwidth along a path (rather than the average or sum) as part of its metric calculation?

5. What happens to EIGRP Hello timers if two neighbours have different Hello interval settings — for example, R1 sends Hellos every 5 seconds and R2 sends Hellos every 10 seconds?

6. An engineer needs to choose a routing protocol for a new enterprise network that includes Cisco, Juniper, and Arista routers. Which protocol is the best choice and why?

8. What is the routing table code for an EIGRP external route, and what does "external" mean in this context?

9. Which statement correctly describes how EIGRP differs from traditional distance-vector protocols like RIP in terms of update behaviour?

1. A route to 172.16.0.0/24 in EIGRP has the following topology table entries:
FD = 300,000
Via R2: (300,000 / 200,000)
Via R3: (350,000 / 250,000)
Via R4: (400,000 / 280,000)
Which neighbours qualify as Feasible Successors?

3. A network engineer notices that `show ip eigrp topology` shows a route marked "A" (Active) for several minutes. What is the most likely cause and risk?