EIGRP metric calculation and the feasibility condition are among the most misunderstood topics on the CCNA 200-301 exam. They separate candidates who memorize from those who truly understand loop-free path selection. In real networks, EIGRP's Diffusing Update Algorithm (DUAL) provides sub-second convergence, but only when you correctly interpret reported distance, feasible distance, and successor routes. This chapter breaks down the exact mathematics and logic Cisco tests under exam objective 3.5, so you never confuse a feasible successor with a backup route again.
Jump to a section
Imagine UPS runs a continent-wide delivery network. Each regional hub (router) knows the cost to deliver a package to any city based on distance, traffic, and fuel costs (composite metric). The hub's local delivery cost to a city is its 'feasible distance' (FD). When a hub learns a route from a neighbor, that neighbor announces its own delivery cost to the same city — this is the 'reported distance' (RD). UPS has a strict safety rule: a hub will only consider using a neighbor's route as a backup if the neighbor's reported distance is strictly less than the hub's own feasible distance. Why? Because if the neighbor's cost is lower than what the hub already knows, the neighbor cannot possibly have a loop back through the hub. In UPS terms, if the Atlanta hub's FD to deliver to Miami is 100, and the Charlotte hub reports an RD of 80, Charlotte's route is 'feasible' — even if Atlanta's direct route fails, using Charlotte guarantees no package loops back to Atlanta. But if Charlotte reported an RD of 120 (greater than Atlanta's 100), that means Charlotte's path might go through Atlanta, creating a loop. UPS would never use that route as backup. This is EIGRP's feasibility condition in action: RD < FD ensures loop-free alternate paths, called feasible successors. The metric itself is a composite of bandwidth, delay, load, and reliability (defaults use only bandwidth and delay), calculated using the famous formula: metric = [K1*BW + (K2*BW)/(256-load) + K3*delay] * 256. Cisco defaults set K1=1, K3=1, K2=K4=K5=0, so the formula simplifies to metric = (10^7/min_bandwidth + sum_delays/10) * 256. This numerical precision is what makes EIGRP deterministic and exam-relevant.
What is the EIGRP Metric and Why Does It Exist?
EIGRP uses a composite metric to select the best loop-free path to a destination. Unlike RIP's hop count or OSPF's cost based solely on bandwidth, EIGRP's metric can incorporate multiple link characteristics, though by default only bandwidth and delay are used. The metric is a 32-bit value ranging from 0 to 4,294,967,295. The primary purpose is to provide a stable, deterministic path selection that avoids flapping due to transient load changes.
The Metric Formula in Detail
The classic EIGRP metric formula is:
metric = [K1 * bandwidth + (K2 * bandwidth) / (256 - load) + K3 * delay] * 256
If K5 is nonzero, the result is multiplied by K5 / (reliability + K4).
Cisco default K values: K1=1, K2=0, K3=1, K4=0, K5=0. This simplifies to:
metric = (bandwidth + delay) * 256
Where: - bandwidth = (10^7 / minimum_bandwidth_on_path) * 256 — but wait, the formula uses a scaled bandwidth: the actual calculation is (10^7 / min_bandwidth_in_kbps) * 256? No, the classic formula uses bandwidth = (10^7 / min_bandwidth) and then multiplies by 256 after adding delay. Let's be precise:
Cisco's actual calculation in IOS:
metric = [ (10^7 / min_bandwidth) + (sum_of_delays / 10) ] * 256
Where: - min_bandwidth is the smallest bandwidth (in kbps) along the path - sum_of_delays is the sum of all interface delays (in microseconds) along the path - delay values are in tens of microseconds; the formula divides by 10 to convert to microseconds
Example: If a path has a minimum bandwidth of 100000 kbps (100 Mbps) and total delay of 1000 microseconds, then:
metric = [ (10^7 / 100000) + (1000 / 10) ] * 256 = [100 + 100] * 256 = 200 * 256 = 51200
Feasibility Condition and DUAL
EIGRP's Diffusing Update Algorithm (DUAL) guarantees loop-free paths at every instant. The feasibility condition is the core mathematical check:
Reported Distance (RD) < Feasible Distance (FD)
Feasible Distance (FD): The best metric (lowest cost) the local router has ever had for that destination.
Reported Distance (RD): The metric advertised by a neighbor for that destination (i.e., the neighbor's FD to that destination).
If a neighbor's RD is strictly less than the local router's FD, that neighbor is a Feasible Successor (FS) — a loop-free alternate path. If the primary successor (best route) fails, the router can immediately use a feasible successor without sending queries, achieving sub-second convergence.
States and Timers
EIGRP routes can be in one of two states: - Passive: The router has a known successor and is not actively searching for a route. - Active: The router has lost its successor and has no feasible successor; it sends queries to all neighbors and waits for replies. A route stuck-in-active (SIA) occurs if a neighbor does not reply within the active timer (default 3 minutes).
Timers:
Hello interval: 5 seconds on high-speed links (>= 1544 kbps), 60 seconds on low-speed links.
Hold time: 15 seconds (default 3x hello interval).
Active timer: 3 minutes (can be adjusted with timers active-time).
Verification Commands
The following IOS commands are essential:
show ip eigrp topology
show ip eigrp topology all-links
show ip route eigrp
show ip protocolsExample output for show ip eigrp topology:
P 10.1.1.0/24, 1 successors, FD is 51200
via 192.168.1.2 (51200/40960), GigabitEthernet0/0
via 192.168.2.2 (61440/51200), GigabitEthernet0/1The 'P' means passive state.
The FD is 51200.
For the first entry (successor): the first number in parentheses is the composite metric (FD via that neighbor), the second is the RD (neighbor's FD).
The second entry is a feasible successor because its RD (51200) is less than the FD (51200)? Wait — RD must be strictly less than FD. Here RD=51200, FD=51200 — that is NOT less, so it would NOT be a feasible successor. Actually, the condition is RD < FD, not <=. So this second entry would not be used as a feasible successor; it would be an alternate path that requires querying. The output would show 'FS' only if it meets the condition. In this example, the second entry might not be marked as FS, but the topology table shows all paths.
Interaction with Related Protocols
EIGRP can redistribute routes from OSPF, RIP, static, and connected routes. When redistributing, the metric must be set manually or via default-metric. EIGRP also supports route summarization, which can affect the feasibility condition because summary routes have a lower metric (since they are originated locally). EIGRP for IPv6 (named mode) uses the same metric logic.
Understanding the Composite Metric
The EIGRP metric is calculated using the formula: metric = [K1*bandwidth + (K2*bandwidth)/(256-load) + K3*delay] * 256, with optional K4/K5 terms. Cisco defaults: K1=1, K2=0, K3=1, K4=0, K5=0. Thus, only bandwidth and delay matter. Bandwidth is the inverse of the minimum bandwidth along the path (in kbps), scaled: bandwidth = 10^7 / min_bandwidth. Delay is the sum of all interface delays (in microseconds) divided by 10. The result is multiplied by 256. For example, a path with min bandwidth 100000 kbps and total delay 1000 µs yields metric = [(10^7/100000) + (1000/10)] * 256 = (100+100)*256 = 51200. Understanding this calculation is critical for exam questions that ask you to compare two paths or identify the best route.
Identifying Successor and Feasible Distance
The successor is the neighbor offering the best metric (lowest FD) to a destination. The Feasible Distance (FD) is the metric of that best path. The router stores the FD as the route's metric in the routing table. Use `show ip eigrp topology` to see the FD and all neighbors. The first entry under a network is usually the successor. The FD is the first number in parentheses (e.g., 51200). The second number is the Reported Distance (RD) from that neighbor. Only the successor is installed in the routing table. If the successor fails and there is no feasible successor, the route goes active.
Checking the Feasibility Condition
A neighbor is a feasible successor if its Reported Distance (RD) is strictly less than the local router's Feasible Distance (FD). This condition ensures the neighbor's path does not loop back through the local router. For example, if FD=51200 and a neighbor reports RD=40960, then 40960 < 51200, so that neighbor is a feasible successor. If RD=51200, it is not less, so not a feasible successor. This is a common exam trap: candidates think RD <= FD is the condition, but it must be strictly less. Use `show ip eigrp topology all-links` to see all neighbors and their RD, and look for the 'FS' flag next to feasible successors.
Verifying with show Commands
Key commands: `show ip eigrp topology` displays the topology table with FD, RD, and successor/FS status. `show ip route eigrp` shows only the best routes (successors) in the routing table. `show ip protocols` shows EIGRP process details, including K values, metric weights, and active timers. Example: `show ip eigrp topology` output shows 'P' for passive, 'A' for active. The 'via' lines show neighbors; if the line has '(FS)' after the RD, it's a feasible successor. Practice interpreting these outputs to quickly identify which paths are loop-free and which are not.
Troubleshooting Metric Inconsistencies
If a route is not appearing as expected, check that K values match across all routers (mismatched K values prevent neighbor adjacency). Use `show ip eigrp neighbors` to verify adjacencies. If a feasible successor is not being used, verify the RD < FD condition. Sometimes a router may have multiple equal-cost paths; EIGRP can load-balance across up to 4 equal-cost paths by default (maximum-paths). Unequal-cost load balancing requires the `variance` command, which multiplies the FD by a factor; any path with metric <= FD * variance is considered. But variance does not affect the feasibility condition — it only allows multiple successors. Misconfiguring delay (e.g., setting delay on a tunnel interface) can artificially inflate metrics and cause suboptimal routing.
In enterprise networks, EIGRP is often deployed in large hub-and-spoke topologies, such as a corporate headquarters connected to dozens of remote branch offices via WAN links. The feasibility condition is crucial for fast failover. For example, a branch router with two WAN links (primary T1, backup DSL) learns routes from both the HQ router and a neighboring branch. The HQ router's RD might be 1000, while the neighboring branch's RD is 800. If the primary link fails, the branch can immediately switch to the neighbor without waiting for queries, achieving sub-second convergence. This is especially valuable for real-time applications like VoIP.
Another scenario: A data center with multiple aggregation routers running EIGRP. Each aggregation router advertises the same prefixes but with different metrics due to varying path bandwidths. The feasibility condition ensures that if one aggregation router fails, the others can take over instantly without routing loops. Network engineers often tune delay values (not bandwidth) to influence path selection because changing delay does not affect the bandwidth-based metric calculation for other paths. For instance, to prefer a satellite link over a terrestrial link, you might lower the delay on the satellite interface.
Common production pitfalls: forgetting to configure the same K values on all routers in the domain. One router with non-default K values will not form adjacencies. Another issue is using the bandwidth command on interfaces to influence EIGRP metric — this also affects QoS and other protocols. A better practice is to use the delay command, which only impacts EIGRP. Also, when redistributing routes from other protocols, you must assign a metric; otherwise, EIGRP will not advertise them. Misconfiguration can lead to suboptimal routing or black holes. In large networks, the active timer (default 3 minutes) can cause routes to go SIA if queries are not answered quickly, so engineers may increase it or design a hierarchical topology to limit query scope.
The CCNA 200-301 exam tests EIGRP metric and feasibility under objective 3.5 (Determine how a router makes a forwarding decision by default). You will be asked to interpret show ip eigrp topology output and identify successors and feasible successors. The most common wrong answers stem from the following traps:
RD vs FD confusion: Candidates often think the route with the lowest RD is the successor. Actually, the successor is based on the local metric (FD), not the neighbor's RD. The RD is used only to determine feasibility.
Feasibility condition inequality: Many candidates believe RD <= FD is the condition. It is strictly RD < FD. If RD equals FD, the neighbor is NOT a feasible successor. This appears in exam questions where the RD is equal to the FD, and you must choose that the route is not a feasible successor.
Metric calculation errors: The exam may give you bandwidth and delay values and ask for the metric. Remember the formula: metric = [(10^7/min_bw) + (sum_delay/10)] * 256. Watch units: bandwidth in kbps, delay in microseconds. A common trap is using delay in tens of microseconds directly without dividing by 10.
Variance misconception: Variance allows unequal-cost load balancing but does NOT create feasible successors. A route must still satisfy RD < FD to be considered. Variance only affects whether the route is installed in the routing table (if it meets the feasibility condition).
Decision strategy: When given a topology table, first identify the FD (the metric of the best route). Then for each alternate path, compare its RD to the FD. If RD < FD, it is a feasible successor. If RD >= FD, it is not. The route with the lowest metric (FD) is the successor. Only feasible successors can be used immediately upon successor failure. For scenario questions, always check the RD < FD condition before declaring a backup route.
EIGRP composite metric formula: metric = [(10^7/min_bandwidth) + (sum_delay/10)] * 256 (with default K values K1=1, K3=1, others 0).
Feasibility condition: Reported Distance (RD) must be strictly less than Feasible Distance (FD) for a neighbor to be a feasible successor.
Feasible Distance (FD) is the best metric the local router has ever had for a destination; it is the metric of the successor route.
Successor is the neighbor with the lowest metric (FD) to a destination; only successors are installed in the routing table.
Use `show ip eigrp topology` to see FD, RD, and feasible successor status; look for '(FS)' next to feasible successors.
EIGRP default hello/hold timers: 5/15 seconds on high-speed links, 60/180 on low-speed links.
Mismatched K values prevent EIGRP neighbor adjacency; all routers must have identical K values.
Variance allows unequal-cost load balancing but does not bypass the feasibility condition; routes must still satisfy RD < FD.
These come up on the exam all the time. Here's how to tell them apart.
EIGRP Feasible Successor
Uses RD < FD condition to guarantee loop-free alternate path.
Backup route is pre-computed and stored in topology table.
Convergence in sub-second (no query needed if FS exists).
Metric is composite (bandwidth+delay).
Only feasible successors can be used immediately upon failure.
OSFP LSA Type 1 Backup Path
Uses SPF calculation to determine all paths; no separate backup concept.
All routes are computed from link-state database; no pre-computed backup.
Convergence depends on SPF re-run (can be seconds in large networks).
Cost is based on bandwidth only (reference bandwidth / interface bandwidth).
No equivalent of feasible successor; all paths are equal-cost or none.
Mistake
The successor is the neighbor with the lowest reported distance (RD).
Correct
The successor is the neighbor with the lowest feasible distance (FD), which is the metric via that neighbor. RD is the metric the neighbor advertises, not the local metric.
Candidates confuse RD and FD because both appear in the topology table output.
Mistake
A neighbor is a feasible successor if its RD is less than or equal to the FD.
Correct
The condition is strictly RD < FD. If RD equals FD, the neighbor is NOT a feasible successor.
Many study materials incorrectly state 'less than or equal to', but Cisco's DUAL requires strict inequality to guarantee loop-free paths.
Mistake
The EIGRP metric is calculated using bandwidth only by default.
Correct
By default, both bandwidth and delay are used (K1=1, K3=1). Load and reliability are not used unless K2 or K4/K5 are configured.
Some candidates think only bandwidth matters because OSPF uses bandwidth, but EIGRP uses two parameters.
Mistake
If variance is set to 2, all routes with metric less than 2 times the FD are installed in the routing table.
Correct
Variance only affects routes that are already feasible successors (RD < FD). It does not bypass the feasibility condition.
Candidates think variance is a magic multiplier that makes any route usable, but the feasibility condition still must be met.
Reveal each answer, then mark whether you got it right. Score 60%+ to unlock the next chapter.
Feasible Distance (FD) is the metric of the best path (successor) from the local router to a destination. It is the lowest metric the router has ever had for that route. Reported Distance (RD) is the metric that a neighbor advertises for the same destination — essentially the neighbor's FD. In the topology table, the FD is the first number in parentheses when you see 'via neighbor (metric/RD)'. The RD is the second number. The feasibility condition uses RD < FD to determine if a neighbor is a feasible successor. Exam tip: Always compare the second number (RD) to the FD (the first number of the successor entry).
With default K values (K1=1, K2=0, K3=1, K4=0, K5=0), the formula simplifies to: metric = [(10^7 / min_bandwidth) + (sum_of_delays / 10)] * 256. min_bandwidth is in kbps, sum_of_delays is in microseconds. For example, if min bandwidth = 100000 kbps (100 Mbps) and total delay = 1000 µs, then metric = [(10^7/100000) + (1000/10)] * 256 = [100 + 100] * 256 = 51200. Note: delay on Cisco interfaces is in tens of microseconds; the 'show interface' output shows delay in microseconds, but the formula expects that value divided by 10. So if an interface shows delay = 100 µs, you use 10 in the formula (100/10).
Yes, if the current successor fails and there is a feasible successor, that feasible successor immediately becomes the new successor without the route going active. The router simply promotes the feasible successor to successor and updates the FD if needed (the FD may stay the same or change if the new successor has a higher metric). However, if the feasible successor's metric is higher than the old FD, the FD remains the old FD (the best metric ever). This is important: FD is the best metric the router has ever had, not necessarily the current metric. So after promotion, the FD might be lower than the new route's metric, which can affect future feasibility checks.
EIGRP neighbors must have identical K values to form an adjacency. If K values differ, the routers will not become neighbors; they will remain in the Init state and log a 'K-value mismatch' error. This is a common misconfiguration when someone changes K values on one router without updating others. Always verify K values with `show ip protocols`. The default K values are K1=1, K2=0, K3=1, K4=0, K5=0. Changing them is rarely needed and can cause issues.
Variance allows EIGRP to load-balance across multiple unequal-cost paths, but only those paths that are feasible successors (RD < FD). For example, if FD=1000 and variance=2, any feasible successor with metric <= 2000 can be installed in the routing table. However, if a path has RD=1200 and FD=1000, it is NOT a feasible successor (RD > FD), so variance will not make it usable. Variance does not bypass the feasibility condition. Exam tip: Variance only affects which feasible successors are installed, not whether they are feasible.
This command shows all EIGRP-learned routes, including those that do not meet the feasibility condition. The standard `show ip eigrp topology` only shows successors and feasible successors. The 'all-links' version shows every neighbor's route, even if RD >= FD. This is useful for troubleshooting why a particular path is not being used. You can see the RD and FD columns and identify which paths fail the feasibility condition. On the exam, you might be given output from this command and asked to determine which neighbors are feasible successors.
No. By default, EIGRP only load-balances over equal-cost paths (up to 4 by default, configurable with `maximum-paths`). To load-balance over unequal-cost paths, you must configure the `variance` command. However, even with variance, a path must first be a feasible successor (RD < FD). If there are multiple feasible successors with different metrics, variance can include them in the routing table. Without variance, only the best metric path (successor) is installed, even if other feasible successors exist.
You've just covered Exam Trap: EIGRP Metric and Feasibility — now see how well it sticks with free CCNA 200-301 practice questions. Full explanations included, no account needed.
Done with this chapter?