EIGRP metric calculation is a core topic for the CCNA 200-301 exam and a fundamental skill for real network engineering. Understanding how EIGRP computes its composite metric—using bandwidth, delay, load, and reliability—enables you to predict and troubleshoot routing decisions in enterprise networks. This chapter covers the exact formula Cisco uses, the default K values, and how to verify and influence metric calculation on IOS routers. Mastering this will help you ace exam questions on metric manipulation and EIGRP path selection.
Jump to a section
Imagine you are a delivery driver who must choose the best route from a warehouse to a customer. You have a GPS app that calculates a route score based on several factors: road width (bandwidth), travel time (delay), traffic congestion (load), and road surface quality (reliability). The app uses a formula: Score = (width factor * 10^7 / road width) + (delay factor * sum of delays). By default, the app ignores traffic and surface quality, focusing only on width and travel time. You can adjust the importance of each factor by changing sliders (K values). For example, if you set the traffic slider to 1, the app will also consider congestion. The road width is measured in Mbps (like EIGRP's bandwidth in kbps), and travel time is measured in microseconds (like EIGRP's delay in tens of microseconds). The app sums the delays of all road segments and uses the minimum road width along the path. This mirrors EIGRP's composite metric: Metric = (K1 * BW + K2 * BW/(256-Load) + K3 * Delay) * K5/(Reliability+K4), but with default K values (K1=1, K3=1, others=0), it simplifies to Metric = BW + Delay, where BW = (10^7 / minimum bandwidth in kbps) * 256 and Delay = (sum of delays in tens of microseconds) * 256. Just as you can tweak the app's sliders to prioritize different factors, a network engineer can modify K values on EIGRP routers to influence path selection.
What is EIGRP Metric Calculation?
EIGRP (Enhanced Interior Gateway Routing Protocol) is a Cisco-proprietary advanced distance vector routing protocol that uses a composite metric to determine the best path to a destination. Unlike RIP's simple hop count or OSPF's cost based on bandwidth, EIGRP's metric is flexible and can incorporate multiple link characteristics: bandwidth, delay, load, and reliability. By default, EIGRP uses only bandwidth and delay, which makes it stable and predictable. The metric calculation is crucial because it directly influences which routes are installed in the routing table and which paths are used for load balancing.
The Composite Metric Formula
The EIGRP composite metric is calculated using the following formula:
Metric = [K1 * BW + (K2 * BW) / (256 - Load) + K3 * Delay] * [K5 / (Reliability + K4)]
Where:
BW = (10^7 / minimum bandwidth along the path in kbps) * 256
Delay = (sum of delays along the path in tens of microseconds) * 256
Load = load value (1-255, where 255 is 100% loaded)
Reliability = reliability value (1-255, where 255 is 100% reliable)
K1, K2, K3, K4, K5 are constants (K values) that can be configured.
Default K values on Cisco IOS are: K1=1, K2=0, K3=1, K4=0, K5=0. With these defaults, the formula simplifies to:
Metric = BW + Delay = (10^7 / min_bandwidth * 256) + (sum_delays * 256)
Note: When K5 = 0, the term [K5 / (Reliability + K4)] is treated as 1 (i.e., omitted).
Step-by-Step Calculation
Consider a route with two hops:
R1 to R2: Bandwidth = 100 Mbps (100,000 kbps), Delay = 100 microseconds (10 tens of microseconds)
R2 to R3: Bandwidth = 10 Mbps (10,000 kbps), Delay = 1000 microseconds (100 tens of microseconds)
Determine minimum bandwidth: min(100000, 10000) = 10000 kbps
Calculate BW component: (10^7 / 10000) * 256 = 1000 * 256 = 256,000
Sum delays: 10 + 100 = 110 tens of microseconds
Calculate Delay component: 110 * 256 = 28,160
Total metric: 256,000 + 28,160 = 284,160
K Values and Metric Tuning
K values are configured globally with the metric weights command under router eigrp. For example:
router eigrp 100
metric weights 0 1 0 1 0 0This sets K1=1, K2=0, K3=1, K4=0, K5=0 (default). To include load, set K2=1. To include reliability, set K4=1 and K5=1. All routers in an EIGRP domain must have identical K values; otherwise, neighbor relationships will not form (a common exam trap).
Verification Commands
To see the metric components for a route:
show ip eigrp topology 10.1.1.0/24Example output:
P 10.1.1.0/24, 1 successors, FD is 284160
via 192.168.1.2 (284160/281600), GigabitEthernet0/0The FD (Feasible Distance) is the metric from this router to the destination. The reported distance (RD) is the metric from the neighbor to the destination (281600).
To see the interface bandwidth and delay values used:
show interfaces GigabitEthernet0/0Look for "BW 100000 Kbit/sec" and "DLY 100 usec".
To see the K values:
show ip protocolsExample snippet:
EIGRP-IPv4 Protocol for AS(100)
Metric weights K1=1, K2=0, K3=1, K4=0, K5=0Interaction with Other Features
EIGRP metric calculation interacts with offset-lists (used to artificially increase metrics), variance (for unequal-cost load balancing), and route summarization (which can affect metric when summarizing). It also works with EIGRP stub and named mode configurations. Understanding the metric is essential for troubleshooting suboptimal routing and for designing networks with predictable path selection.
Identify interface bandwidth and delay
Use `show interfaces` to find the bandwidth (in kbps) and delay (in microseconds) for each interface along the path. For example, `show interfaces GigabitEthernet0/0` might show "BW 100000 Kbit/sec" and "DLY 100 usec". Note that EIGRP uses the configured bandwidth and delay, not the actual link speed. By default, Cisco interfaces have default bandwidth and delay values based on interface type (e.g., GigabitEthernet has BW=1000000 kbps? Actually default for GigabitEthernet is 1000000 kbps? No, default for GigabitEthernet is 1000000 kbps? Wait, default for GigabitEthernet is 1000000 kbps? Actually, the default bandwidth for GigabitEthernet is 1000000 kbps (1 Gbps) and delay is 10 microseconds. But on many IOS versions, it's 1000000 kbps and 10 usec. For FastEthernet, BW=100000 kbps, delay=100 usec. Verify with `show interfaces`.
Convert delay to tens of microseconds
EIGRP expects delay in tens of microseconds. If the interface delay is shown in microseconds (usec), divide by 10. For example, if delay is 100 usec, then tens of microseconds = 10. In the metric formula, the sum of delays is multiplied by 256. So if you have two hops with delays 100 usec and 1000 usec, the sum in tens of microseconds is 10 + 100 = 110. Then Delay component = 110 * 256 = 28160.
Find minimum bandwidth along path
Identify the smallest bandwidth among all outgoing interfaces on the path to the destination. For example, if one interface has BW=100000 kbps and another has BW=10000 kbps, the minimum is 10000 kbps. This minimum bandwidth is used in the BW component: BW = (10^7 / min_bandwidth_kbps) * 256. So (10^7 / 10000) * 256 = 1000 * 256 = 256000.
Calculate composite metric
With default K values (K1=1, K3=1, others=0), the metric = BW + Delay. Using the values from previous steps: BW = 256000, Delay = 28160, so metric = 284160. This is the Feasible Distance (FD) from the local router to the destination. If you need to include load or reliability, adjust K values accordingly and use the full formula.
Verify with show commands
Use `show ip eigrp topology` to see the FD and RD for each route. For example: `show ip eigrp topology 10.1.1.0/24` might output "FD is 284160, via 192.168.1.2 (284160/281600)". The first number is the FD, the second is the RD (metric from the neighbor). To see the metric components per interface, use `show ip eigrp interfaces detail` or `show interfaces`.
Tune metric using offset-list
To influence path selection without changing K values, use an offset-list. For example, to add 1000 to the metric for route 10.1.1.0/24 coming from interface Gig0/0: `access-list 10 permit 10.1.1.0 0.0.0.255` then `router eigrp 100` then `offset-list 10 in 1000 GigabitEthernet0/0`. This increases the metric by 1000, potentially making an alternate path more preferred.
In a large enterprise network with multiple redundant WAN links, EIGRP metric calculation is critical for ensuring that traffic uses the most appropriate path. For example, consider a company with two sites connected via a primary T1 (1.544 Mbps) and a backup DSL (10 Mbps but higher delay). The default metric (bandwidth and delay) might favor the DSL because of its higher bandwidth, even though the T1 has lower delay. A network engineer might adjust the delay on the DSL interface to make the T1 preferred for voice traffic, while still using the DSL for bulk data. This is done by setting the delay on the DSL interface to a higher value using the delay command under the interface.
Another common scenario is unequal-cost load balancing using the variance command. By default, EIGRP only load-balances across equal-cost paths (same metric). With variance, you can include paths whose metric is within a multiplier of the best metric. For example, if the best path metric is 100 and variance is set to 2, paths with metric up to 200 can be used. The traffic is distributed proportionally to the metric (lower metric gets more traffic). This is useful when you have a fast link and a slower backup link that you still want to use partially.
Misconfiguration can lead to suboptimal routing or even routing loops. For instance, if K values are changed on one router but not on its neighbors, EIGRP neighbor relationships will fail (because K values must match). Also, if an engineer mistakenly sets a very high delay on an interface, that path may be avoided entirely, causing traffic to congest another link. In production, it's common to use the default K values and only tweak delay or bandwidth to influence metrics, as changing K values can cause instability. Always document any metric changes and verify with show ip eigrp topology and show ip route.
The CCNA 200-301 exam tests EIGRP metric calculation under exam objective 3.5 (Configure and verify EIGRP). You must be able to calculate the composite metric given bandwidth and delay values, understand the role of K values, and interpret show ip eigrp topology output. Common wrong answers include:
Assuming load and reliability are used by default: Many candidates think EIGRP uses all four metrics by default. In reality, only bandwidth and delay are used (K1=1, K3=1, others=0). The exam may present a scenario with load values and ask for the metric; you must know to ignore load unless K2 is set to 1.
Using the wrong bandwidth unit: The formula uses bandwidth in kbps, but interface bandwidth is often shown in kbps (e.g., 100000 for FastEthernet). However, delay is shown in microseconds, and you must convert to tens of microseconds (divide by 10). A common trap is to use delay in microseconds directly, resulting in a metric that is off by a factor of 10.
Forgetting the multiplication by 256: Both the BW and Delay components are multiplied by 256. Candidates sometimes forget this and compute BW as (10^7 / bandwidth) without the *256, leading to an incorrect metric.
Confusing FD and RD: The Feasible Distance (FD) is the metric from the local router to the destination. The Reported Distance (RD) is the metric advertised by the neighbor. The exam may ask which value is used for feasibility condition (RD must be less than FD). Many candidates mix them up.
Decision rule: When given a scenario with multiple paths, calculate the metric for each path using the formula. The path with the lowest metric is the successor (best path). If metrics are equal, EIGRP will load-balance (up to 4 equal-cost paths by default). If you see a question about unequal-cost load balancing, remember the variance command and that traffic is distributed proportionally.
Specific values to memorize: Default K values (1,0,1,0,0). The constant 10^7. The multiplication factor 256. Default maximum paths = 4. Variance default = 1 (equal-cost only).
EIGRP composite metric formula: Metric = [K1*BW + K2*BW/(256-Load) + K3*Delay] * [K5/(Reliability+K4)], with default K1=1, K2=0, K3=1, K4=0, K5=0, simplifying to BW + Delay.
BW = (10^7 / minimum bandwidth in kbps) * 256; Delay = (sum of delays in tens of microseconds) * 256.
Default K values must match on all EIGRP routers in the same AS; mismatch prevents neighbor adjacency.
Use `show ip eigrp topology` to see FD and RD; `show interfaces` for bandwidth and delay values.
Offset-list can adjust metric without changing K values: `offset-list {acl} {in|out} {offset} [interface]`.
Variance command enables unequal-cost load balancing: `variance multiplier` (default 1).
Maximum equal-cost paths default is 4; can be changed with `maximum-paths`.
These come up on the exam all the time. Here's how to tell them apart.
EIGRP Metric
Composite metric based on bandwidth, delay, load, reliability (default BW+Delay).
Formula: Metric = (10^7/minBW + sumDelay) * 256 (with defaults).
Metric is 32-bit, up to 4.2 billion.
Unequal-cost load balancing supported via variance.
K values must match across all routers in AS.
OSPF Cost
Cost based solely on interface bandwidth (cost = reference BW / interface BW).
Formula: Cost = 100 Mbps / interface BW (default reference).
Cost is 16-bit, range 1-65535.
Only equal-cost load balancing by default (no variance).
Cost is locally significant; can be manually set per interface.
Mistake
EIGRP uses load and reliability in the metric by default.
Correct
By default, K1=1, K2=0, K3=1, K4=0, K5=0, so only bandwidth and delay are used. Load and reliability are ignored unless K2 and K4/K5 are set to non-zero.
Cisco documentation often shows the full formula, leading candidates to assume all variables are active.
Mistake
The delay value used in the formula is in microseconds.
Correct
The formula uses delay in tens of microseconds. Interface delay is shown in microseconds, so you must divide by 10 before using it in the sum.
The `show interfaces` output displays delay in microseconds, which is easy to misinterpret.
Mistake
The metric is computed using the bandwidth of the first hop only.
Correct
EIGRP uses the minimum bandwidth along the entire path (all outgoing interfaces) to the destination.
Some distance vector protocols use only the first hop metric, so candidates may carry that assumption.
Mistake
Changing K values on one router is sufficient to affect path selection.
Correct
All routers in the EIGRP domain must have identical K values; otherwise, neighbor relationships fail. K values are exchanged in hello packets, and mismatch prevents adjacency.
Candidates may think EIGRP is like OSPF where you can change cost locally without affecting neighbors.
Reveal each answer, then mark whether you got it right. Score 60%+ to unlock the next chapter.
First, find the minimum bandwidth (in kbps) among all outgoing interfaces on the path. Then compute BW = (10^7 / min_bandwidth) * 256. Next, sum the delays of all outgoing interfaces in tens of microseconds (divide interface delay in microseconds by 10). Compute Delay = sum_delays * 256. With default K values, metric = BW + Delay. For example, if min bandwidth = 100000 kbps and sum delays = 100 tens of microseconds, BW = (10^7/100000)*256 = 100*256 = 25600, Delay = 100*256 = 25600, metric = 51200. Always verify with `show ip eigrp topology`.
The default K values are K1=1, K2=0, K3=1, K4=0, K5=0. This means only bandwidth and delay are used in the metric calculation. Load and reliability are ignored. You can verify with `show ip protocols`. If you change K values, all routers in the EIGRP domain must use the same values; otherwise, neighbor adjacencies will not form.
EIGRP supports unequal-cost load balancing using the `variance` command. The variance is a multiplier applied to the best metric (successor's FD). Any feasible successor whose metric is less than or equal to the best metric times the variance is eligible for load balancing. Traffic is distributed proportionally to the metrics (lower metric gets more traffic). By default, variance is 1 (equal-cost only). The command is `router eigrp [AS]` then `variance [multiplier]`.
Yes, you can use an offset-list to add a value to the metric for routes received or sent on a specific interface. For example, `offset-list 10 in 1000 GigabitEthernet0/0` adds 1000 to the metric for routes matching access-list 10. You can also manually set the bandwidth or delay on an interface using the `bandwidth` or `delay` commands, which will affect the metric calculation for all routes using that interface.
One common reason is mismatched K values. Check with `show ip protocols` on each router. Other causes include mismatched autonomous system numbers, passive interfaces, or authentication issues. Also ensure that the primary IP addresses of the interfaces are in the same subnet. Use `debug eigrp packets` and `show ip eigrp neighbors` to diagnose.
FD (Feasible Distance) is the metric from the local router to the destination network. RD (Reported Distance) is the metric advertised by the neighbor for that destination. In `show ip eigrp topology` output, the format is (FD/RD). For a route to be considered a feasible successor, the RD must be less than the FD of the successor. This ensures loop-free paths.
When EIGRP summarizes routes, the metric of the summary route is calculated based on the best component route (lowest metric) within the summary. The summary route inherits the metric of that best route. This can cause suboptimal routing if the summary hides more specific routes with different metrics. Use `summary-metric` command to manually set the metric of the summary route.
You've just covered EIGRP Metric Calculation — now see how well it sticks with free CCNA 200-301 practice questions. Full explanations included, no account needed.
Done with this chapter?