Routing protocols are how routers learn where networks are without an engineer typing every route by hand. EIGRP is the Cisco-originated option that is known for fast convergence, simple configuration and the ability to share traffic across paths of different cost. This guide explains the logic that makes it work, gives you the metric formula with a number you can check by hand, shows the IOS configuration, and compares it honestly with OSPF. It also states clearly how much of it the CCNA 200-301 asks of you today, because that affects how you should spend study time.
What is EIGRP and how does it work?
EIGRP is an interior gateway protocol: it runs between routers inside one organization or autonomous system. Cisco originally kept it proprietary, then published the basic protocol as informational RFC 7868 in 2016. Unlike a classic distance vector protocol such as RIP, which periodically re-sends its whole table, EIGRP sends full updates only when a neighbor first appears and afterwards sends partial, triggered updates when something changes. Unlike a link-state protocol, a router does not hold a full map of the network; it knows the routes its neighbors advertise and the metrics they report.
That is why it is called advanced distance vector (or sometimes hybrid): it advertises routes like a distance vector protocol, yet it has neighbor relationships and loop-free, fast failover logic closer to link-state. The behaviors worth remembering:
- Neighbors are discovered and maintained by small multicast Hello packets to 224.0.0.10.
- Updates are sent reliably, with acknowledgements, and only to affected neighbors.
- The DUAL algorithm guarantees loop-free paths and pre-computes backup routes.
- Metrics are based on bandwidth and delay by default and can include load and reliability.
- It supports VLSM, manual summarization on any interface, IPv4 and IPv6, and unequal-cost load balancing.
What are the three EIGRP tables and packet types?
| Table | Contents | Show command |
|---|---|---|
| Neighbor table | Directly connected EIGRP routers with an adjacency | show ip eigrp neighbors |
| Topology table | All routes learned from neighbors, with successors and feasible successors | show ip eigrp topology |
| Routing table | The best routes installed (successors), marked D for EIGRP | show ip route eigrp |
| Packet | Purpose |
|---|---|
| Hello | Discover and keep neighbors; unreliable multicast |
| Update | Carry route information; reliable, unicast to a new neighbor or multicast to all |
| Query | Ask neighbors whether they have a path when a route is lost and no feasible successor exists |
| Reply | Answer a query |
| Acknowledgement (ACK) | Confirm reliable packets; a Hello without data |
EIGRP does not ride on TCP or UDP. It sits directly on IP as protocol 88 and implements its own Reliable Transport Protocol, a point that appears in comparisons with TCP and UDP-based protocols.
What are DUAL, successors and feasible successors?
The Diffusing Update Algorithm (DUAL) chooses best paths and proves that backup paths cannot form loops. It uses four terms you must be able to define:
| Term | Definition |
|---|---|
| Reported distance (RD) or advertised distance (AD) | The metric to the destination as advertised by a neighbor. |
| Feasible distance (FD) | The lowest metric from this router to the destination, including the link to the neighbor. It is recorded for the best route. |
| Successor | The neighbor on the best loop-free path. It is installed in the routing table. |
| Feasible successor (FS) | A backup neighbor whose reported distance is lower than the current feasible distance. It is kept in the topology table. |
| Feasibility condition | RD of the alternative path < FD of the current successor. If true, the path cannot loop through this router. |
Why does the condition prove there is no loop? If a neighbor can reach the destination with a metric lower than my own best, then that neighbor is closer to the destination than I am, so its path cannot go through me. A neighbor with a higher reported distance might send traffic back through me, so DUAL does not trust it until it checks.
Worked example: which neighbors are feasible?
Router R1 has three neighbors that can reach 10.9.0.0/24. The cost of the link from R1 to each neighbor is added to that neighbor's reported distance:
| Neighbor | Reported distance | Link cost | Total (FD via this neighbor) | Role |
|---|---|---|---|---|
| R2 | 1000 | 2000 | 3000 | Successor (lowest total; FD = 3000) |
| R3 | 2500 | 2000 | 4500 | Feasible successor (2500 < 3000) |
| R4 | 3500 | 1500 | 5000 | Not feasible (3500 is not < 3000) |
The numbers are simplified for clarity, not real interface values. If R2 fails, R1 promotes R3 immediately, with no queries, and the route barely flaps. If both R2 and R3 failed, R4 would not qualify: R1 would put the route into the Active state, send queries to its neighbors, and wait for replies before choosing a new path. A route that is stable is in the Passive state.
What is the EIGRP metric formula?
The classic EIGRP metric is 256 times the sum of a bandwidth term (10,000,000 divided by the slowest link in kb/s) and a delay term (total delay in tens of microseconds). A lower metric is better. Bandwidth uses only the slowest link on the path, while delay is added up across every outgoing interface. EIGRP can combine five values (bandwidth, load, delay, reliability and MTU), weighted by constants K1 to K5. With the default K-values (K1 = 1, K2 = 0, K3 = 1, K4 = 0, K5 = 0) the load and reliability terms vanish, and the classic formula reduces to:
Classic EIGRP metric (default K-values)
Metric = 256 x ( 10,000,000 / minimum bandwidth in kb/s + sum of delays in tens of microseconds )
Minimum bandwidth is the slowest link along the path. Delay is cumulative, added at each outgoing interface.
Worked example: calculating a metric
R1 reaches a network through R2. R1 to R2 is Gigabit Ethernet (delay 10 microseconds). R2 reaches the network over a Fast Ethernet link (100 Mb/s = 100,000 kb/s, delay 100 microseconds):
- Minimum bandwidth = 100,000 kb/s, so 10,000,000 / 100,000 = 100.
- Total delay = 10 + 100 = 110 microseconds = 11 tens of microseconds, so 11.
- Metric = 256 x (100 + 11) = 256 x 111 = 28,416.
R2's own metric to the network is 256 x (100 + 10) = 28,160, which is the reported distance R2 advertises; R1 adds its link's delay and arrives at 28,416 as the feasible distance. Note the interface delay values shown in show interfaces are in microseconds, while the formula uses tens of microseconds. A common arithmetic mistake is using the highest bandwidth instead of the lowest, or forgetting the multiplication by 256.
Wide metrics and named mode
The classic metric has little resolution for fast links: 1 Gb/s gives a bandwidth component of 10 and 10 Gb/s gives 1, and above 10 Gb/s every link gets the same bandwidth component (1), so classic EIGRP cannot rank fast links well. Wide metrics use 64-bit values and picosecond-scale delay, and they are used by EIGRP named mode. Both routers in a neighborhood understand each other's metric style; for exam purposes remember that the classic formula above is the one asked for.
What must match to form an EIGRP neighbor adjacency?
| Requirement | Must match? | Notes |
|---|---|---|
| Autonomous system number (classic mode) | Yes | The number after router eigrp. Different AS means no adjacency. |
| K-values | Yes | Default K-values must be the same on both sides. |
| Common subnet / reachable primary address | Yes | Hellos come from the interface primary address, which must be in the neighbor’s subnet. |
| Authentication | Yes, if configured | Key and type must match. |
| Hello and hold timers | No | Can differ, unlike OSPF. Mismatches can cause flapping though. |
| Router ID | Must be unique | Duplicate router IDs cause problems with external routes. |
| Passive interface | Not set on the link | A passive interface sends no hellos and forms no neighbors. |
Practice this for the CCNA
Try the free Cisco CCNA 200-301 practice exam: 30 questions per session in practice mode, no signup needed. The full 500-question bank with exam mode is on the pricing page.
How do I configure EIGRP on Cisco IOS?
Two routers, R1 and R2, share the link 10.0.12.0/24, and R1 also has a LAN 192.168.1.0/24. The classic configuration is short:
R1(config)# router eigrp 100
R1(config-router)# eigrp router-id 1.1.1.1
R1(config-router)# network 10.0.12.0 0.0.0.255
R1(config-router)# network 192.168.1.0 0.0.0.255
R1(config-router)# passive-interface GigabitEthernet0/1
R1(config-router)# no auto-summary
R2(config)# router eigrp 100
R2(config-router)# eigrp router-id 2.2.2.2
R2(config-router)# network 10.0.12.0 0.0.0.255
R2(config-router)# network 172.16.0.0 0.0.255.255Points that trip people up. The network command takes a wildcard mask and enables EIGRP on interfaces whose IP addresses fall in that range; it also advertises the connected subnet of those interfaces. The AS number (100) must match on both routers. passive-interface advertises a LAN into EIGRP without sending hellos out of it, which both saves resources and prevents unauthorized neighbors. Auto-summary was on by default in very old IOS releases and is off by default in current releases, but the explicit no auto-summary is harmless and still appears in labs.
Named mode
Named mode puts all the settings under one hierarchy and enables wide metrics. The same network on R1:
R1(config)# router eigrp CERTSHARP
R1(config-router)# address-family ipv4 unicast autonomous-system 100
R1(config-router-af)# eigrp router-id 1.1.1.1
R1(config-router-af)# network 10.0.12.0 0.0.0.255
R1(config-router-af)# network 192.168.1.0 0.0.0.255
R1(config-router-af)# af-interface GigabitEthernet0/1
R1(config-router-af-interface)# passive-interface
R1(config-router-af-interface)# exit-af-interface
R1(config-router-af)# exit-address-familyIn named mode the process name (CERTSHARP) is only a label; the AS number is configured inside the address family and is what must match between neighbors. A router in classic mode and one in named mode can still become neighbors if the AS numbers match.
How do I verify EIGRP with show commands?
R1# show ip eigrp neighbors
EIGRP-IPv4 Neighbors for AS(100)
H Address Interface Hold Uptime SRTT RTO Q Seq
(sec) (ms) Cnt Num
0 10.0.12.2 Gi0/0 13 00:12:41 1 100 0 5
R1# show ip eigrp topology
EIGRP-IPv4 Topology Table for AS(100)/ID(1.1.1.1)
P 172.16.0.0/16, 1 successors, FD is 28416
via 10.0.12.2 (28416/28160), GigabitEthernet0/0
R1# show ip route eigrp
D 172.16.0.0/16 [90/28416] via 10.0.12.2, 00:12:30, GigabitEthernet0/0The output is illustrative but matches the earlier worked example. In the topology table, P means Passive (stable), and the pair in parentheses is (feasible distance / reported distance). In the routing table, D marks an EIGRP route and the bracket shows [administrative distance / metric]: 90 for an internal route. Also useful: show ip protocols (AS, router ID, K-values, networks, timers), show ip eigrp interfaces and show ip eigrp topology all-links (to see routes that are not feasible successors).
How does EIGRP unequal-cost load balancing work?
Most protocols only balance across equal-cost paths. EIGRP can also use unequal ones. By default it installs up to 4 equal-cost paths (maximum-paths can raise this). With variance, any feasible successor whose feasible distance is less than or equal to the best FD multiplied by the variance is installed too.
R1(config)# router eigrp 100
R1(config-router)# variance 2
R1(config-router)# maximum-paths 6Using the earlier example with FD 3000 via R2, R3 at 4500 and R4 at 5000: with variance 2 the limit is 6000. R3 is a feasible successor and 4500 is within 6000, so it is installed. R4 is also within 6000, but it fails the feasibility condition, so it is not used. Traffic is shared in inverse proportion to the metrics, so the cheaper path carries more. The classic exam trap is believing a high variance can include any path.
EIGRP vs OSPF: what is the difference?
| Feature | EIGRP | OSPF |
|---|---|---|
| Type | Advanced distance vector | Link state |
| Algorithm | DUAL | Dijkstra SPF |
| Standard | Cisco-originated; basic protocol in RFC 7868 (informational) | Open standard (RFC 2328 for OSPFv2) |
| Administrative distance | 90 internal / 170 external | 110 |
| Metric | Bandwidth and delay (default), composite | Cost based on interface bandwidth |
| Transport | IP protocol 88, 224.0.0.10 | IP protocol 89, 224.0.0.5 and 224.0.0.6 |
| Hierarchy | No areas; summarize anywhere | Areas, with Area 0 as backbone |
| Backup route | Feasible successor, instant failover | SPF recalculation on topology change |
| Load balancing | Equal and unequal cost | Equal cost only |
| Multi-vendor support | Limited in practice | Universal |
| Timers must match? | No | Yes (hello and dead) |
Practical guidance: EIGRP is simple and fast in Cisco-only networks, and OSPF is the standard when you need multi-vendor interoperability. For the exam, the safest rule is to know the table above and be able to say which protocol a scenario describes from its clues (areas and LSAs mean OSPF; DUAL, successors and variance mean EIGRP).
What are the common EIGRP problems?
- No neighbors. Check the AS number, K-values, authentication,
passive-interface, network statements and that the interface primary addresses share a subnet. Useshow ip eigrp interfacesto confirm the interface is enabled. - Missing routes. The
networkstatement does not cover the interface, auto-summary is hiding discontiguous subnets in an older IOS, a distribute list or route filter blocks the route, or the route is feasible only and not a successor. - Neighbor flapping. Link errors, congestion that drops hellos, or aggressive timers. Look at the hold time column and logs.
- Stuck in Active. A neighbor never answers queries. Summarization and stub routers limit how far queries spread and reduce this risk.
- Suboptimal routing. Interface bandwidth values are used by the formula and may not match the real speed. Changing
bandwidthon an interface changes EIGRP metrics (and only the metric, not actual speed). - Duplicate router IDs. Cause external routes to be ignored. Set
eigrp router-idmanually when in doubt.
Is EIGRP on the CCNA exam?
EIGRP is not a named configuration topic on the current CCNA 200-301, but routing concepts that include it, such as administrative distance and metrics, can still appear.
Be careful with outdated advice. Older CCNA versions required full EIGRP configuration. In Cisco's published CCNA 200-301 blueprint, single-area OSPFv2 is the named dynamic routing protocol to configure and verify, while routing concepts such as comparing routing protocol types, administrative distance, metrics and the routing table remain in scope, which is where EIGRP can still appear in a question. Cisco updates this blueprint, and its wording changes between versions, so please confirm what is currently listed on the official Cisco CCNA exam topics page before deciding how deeply to study EIGRP. Our CCNA exam guide follows the blueprint in more detail.
A sensible study position: learn the concepts and the administrative distance table well (they are cheap points), practice OSPF configuration thoroughly, and treat EIGRP configuration as useful real-world knowledge and as a bonus rather than a core exam skill unless your blueprint version says otherwise. It is certainly worth knowing if you will work in Cisco-heavy environments, which is part of why people ask whether CCNA or Network+ suits them better. Related reading: the CCNA cheat sheet, the subnetting guide (wildcard masks in network commands come from it), and the CCNA exam cost breakdown. For first hop redundancy on the host side of the same routers see HSRP, and for the Layer 2 topology beneath them read spanning tree protocol.
Check yourself: three quick EIGRP questions
Check 1. A route has a feasible distance of 3000. A neighbor advertises the same prefix with a reported distance of 2500. Is that neighbor a feasible successor?
Yes. 2500 is lower than 3000, so the feasibility condition is met and the neighbor is a loop-free backup.
Check 2. Which default K-values are used in the classic EIGRP metric, and which link attributes do they select?
K1 = 1 and K3 = 1, with K2, K4 and K5 = 0. K1 weights bandwidth and K3 weights delay, so by default only the minimum bandwidth and the cumulative delay count.
Check 3. Do the hello and hold timers have to match for an EIGRP adjacency?
No. The AS number, K-values, subnet and authentication must match, but timers can differ. Mismatched timers would break an OSPF adjacency, not an EIGRP one.
EIGRP practice questions with answers
More routing practice: CCNA practice questions or the CCNA practice exam.
1. What is the administrative distance of an internal EIGRP route, and of an external one?
90 for internal and 170 for external. Summary routes generated by EIGRP use 5. OSPF is 110 and RIP is 120, so EIGRP wins against them for the same prefix.
2. Which algorithm does EIGRP use to calculate loop-free paths?
DUAL, the Diffusing Update Algorithm. Dijkstra’s SPF is what OSPF and IS-IS use. Bellman-Ford is the basis of classic distance vector protocols like RIP.
3. What is the difference between the feasible distance and the reported distance?
The reported (advertised) distance is the metric a neighbor says it has to the destination. The feasible distance is the lowest total metric from the local router to the destination, which is the neighbor’s reported distance plus the cost of the link to that neighbor.
4. A route to 10.9.0.0/24 has FD 3000. Neighbor X reports 1000, neighbor Y reports 2500 and neighbor Z reports 3500 (all via different next hops). Which are feasible successors?
Y is a feasible successor, because 2500 < 3000. Z is not, because 3500 is not lower than 3000. X is the successor (the best route); with link cost 2000 its FD is 3000. Z stays in the topology table only as a possible route, and EIGRP would have to send queries if X failed and no feasible successor existed.
5. Two directly connected routers do not form an EIGRP neighbor relationship. Which mismatch is NOT a possible cause?
A hello timer mismatch is not a cause. A different AS number, different K-values, different authentication, an ACL or passive interface blocking hellos, or addresses on different subnets can all prevent the adjacency.
6. What does variance 3 do?
It allows feasible successors with an FD up to 3 times the best FD to be installed as additional routes, so traffic is balanced across unequal-cost paths. Routes that fail the feasibility condition are never used, no matter how high the variance is.
7. Which multicast address and IP protocol number does EIGRP use?
224.0.0.10 for IPv4 hellos and updates, and IP protocol 88. EIGRP does not use TCP or UDP; it uses its own Reliable Transport Protocol (RTP) on top of IP.
8. A router shows a route in the Active (A) state in show ip eigrp topology for a long time. What does this mean?
The router lost its successor and had no feasible successor, so it sent queries to its neighbors and is waiting for replies. A route stuck in Active (stuck in active, SIA) usually points to a neighbor that is not answering, such as a failing link or an overloaded router. A healthy route is in the Passive (P) state.
Frequently asked questions
What is EIGRP?
EIGRP (Enhanced Interior Gateway Routing Protocol) is a routing protocol for exchanging routes inside one organization, developed by Cisco and published as informational RFC 7868. It is often called an advanced distance vector or hybrid protocol: routers share routes with neighbors rather than a full topology map, but they use the DUAL algorithm to find loop-free paths and keep backup routes ready.
What is the EIGRP metric formula?
With default K-values (K1 = 1, K3 = 1, others 0), the classic metric is 256 x (10,000,000 / lowest bandwidth in kb/s + sum of delays in tens of microseconds). Bandwidth is the slowest link on the path and delay is the cumulative delay of all outgoing interfaces. Newer "wide" metrics, used with named mode, are 64-bit values so that very fast links can be told apart.
What is the administrative distance of EIGRP?
Internal EIGRP routes have an administrative distance of 90, external routes redistributed into EIGRP have 170, and EIGRP summary routes have 5. Compared with OSPF at 110 and RIP at 120, a router prefers EIGRP when it learns the same prefix from several protocols.
What is a feasible successor in EIGRP?
A feasible successor is a backup next hop that is guaranteed to be loop-free. It qualifies when its reported distance (the metric the neighbor itself advertises) is lower than the feasible distance of the current best route. Because it satisfies this feasibility condition, EIGRP can switch to it instantly without querying other routers.
What must match for two routers to become EIGRP neighbors?
The autonomous system number must match, the K-values must match, the routers must be on a common subnet and be able to receive each other’s hellos, and any configured authentication must match. Hello and hold timers do not need to match, which is a difference from OSPF.
Is EIGRP on the CCNA exam?
It depends on the exam version. The current Cisco CCNA 200-301 blueprint emphasizes OSPFv2 for configuration, and EIGRP is not, to our knowledge, a named configuration topic, though routing protocol comparison and administrative distance questions can still mention it. Cisco revises the blueprint, so check the official exam topics page before building your study plan around or away from EIGRP.
How do I do unequal-cost load balancing in EIGRP?
Use the variance command under the EIGRP process. A feasible successor is installed in the routing table alongside the successor if its feasible distance is less than or equal to the successor’s feasible distance multiplied by the variance. Variance 1 (the default) means equal-cost paths only. Traffic is then shared in inverse proportion to the route metrics, so lower-metric paths carry more.
How we wrote and checked this
CertSharp is a practice-exam provider, and we sell the CCNA practice questions linked in this guide. We wrote this article from RFC 7868 and Cisco's EIGRP documentation, linked in Sources. The metric calculations were recomputed with a script, and the feasible successor example follows the feasibility condition exactly. The router output is illustrative, not captured from production equipment. Last reviewed October 2026. Cisco changes the CCNA blueprint, so confirm what it lists on the official page.
Sources and further reading
- RFC 7868: Cisco's Enhanced Interior Gateway Routing Protocol (EIGRP)
- Cisco: Understand and Use the Enhanced Interior Gateway Routing Protocol
- Cisco Learning Network: CCNA 200-301 exam topics
Defaults and command syntax can vary by IOS release and platform. Outputs on this page are illustrative; confirm on your device before changing production routing.
Practice routing and the rest of the CCNA
30 free questions per session in practice mode, no signup and no credit card. The full 500-question CCNA bank with exam mode is $11.99 lifetime, or $9.99/month Pro unlocks every CertSharp certification.