EIGRP (Enhanced Interior Gateway Routing Protocol)

Enhanced Interior Gateway Routing Protocol (EIGRP) represents Cisco's sophisticated answer to the limitations of traditional distance-vector protocols like RIP, combining the best characteristics of distance-vector and link-state protocols into what Cisco originally called an 'advanced distance-vector' or 'hybrid' protocol. Initially proprietary to Cisco, EIGRP was published as an open standard in 2013 (Informational RFC 7868), though practical deployment remains predominantly in Cisco-dominated networks. Understanding EIGRP's operational characteristics, advantages, and appropriate use cases is essential for CCNA certification and for designing efficient routing solutions in Cisco infrastructures. EIGRP's Administrative Distance of 90 makes it the most trusted dynamic routing protocol in Cisco's default hierarchy—more trusted than OSPF (110), RIP (120), or external routes. This reflects Cisco's confidence in EIGRP's routing decisions and its sophisticated metric calculation. When a router runs multiple routing protocols simultaneously (common during migrations or in complex redistributed environments), EIGRP routes are automatically preferred over OSPF or RIP routes to the same destination, assuming all have default AD values. This preferential treatment can be modified using administrative distance manipulation, but the default AD=90 indicates EIGRP's intended role as a primary IGP in Cisco networks. The protocol's most compelling advantage is rapid convergence through the Diffusing Update Algorithm (DUAL), which we'll explore in depth later. Unlike distance-vector protocols that must wait through hold-down timers or link-state protocols that must run complex SPF calculations across the entire topology, EIGRP with proper feasible successors can converge almost instantaneously—often subsecond—when primary paths fail. DUAL guarantees that backup routes (feasible successors) are absolutely loop-free at all times, eliminating the counting-to-infinity problem that plagued early distance-vector protocols. This loop-freedom assurance allows immediate failover to pre-calculated backup paths without triggering queries or recalculation. EIGRP's bandwidth efficiency represents another significant advantage. Unlike link-state protocols that flood complete topology information periodically, EIGRP sends partial updates only when topology changes occur, and only to affected routers. Initial neighbor establishment involves full topology exchange, but subsequent updates contain only changes—a new route, a metric change, a withdrawn route. This bounded update approach dramatically reduces protocol overhead compared to periodic full updates. Updates are sent reliably using EIGRP's own reliability mechanism (RTP - Reliable Transport Protocol), ensuring critical information isn't lost while avoiding TCP overhead. Classless routing with full VLSM and CIDR support allows EIGRP to operate efficiently in modern networks with variable-length subnetting. EIGRP carries subnet mask information in all updates, enabling precise route advertisement regardless of classful boundaries. The 'no auto-summary' command (now default in modern IOS but previously required explicit configuration) disables legacy classful behavior, ensuring EIGRP advertises routes with their actual masks rather than summarizing to classful boundaries. This is critical for proper operation in networks using VLSM. Unequal-cost load balancing, unique among the routing protocols covered in CCNA, allows EIGRP to utilize multiple paths to a destination even when they have different metrics. While OSPF and RIP only load-balance across equal-cost paths, EIGRP's variance command enables proportional load distribution across paths with different costs, maximizing available bandwidth. For example, traffic might split 75/25 between a high-bandwidth primary link and a lower-bandwidth backup, rather than leaving the backup completely idle. EIGRP uses multicast address 224.0.0.10 for updates and hello messages, reducing impact on devices not running EIGRP. This differs from RIP's 224.0.0.9 and OSPF's 224.0.0.5/224.0.0.6. Neighbors are discovered automatically through hello messages, and adjacencies form when neighbors agree on common parameters (AS number, K values, authentication). Understanding EIGRP fundamentals provides the foundation for effective deployment in enterprise networks requiring fast convergence and efficient routing.

EIGRP's composite metric represents one of the protocol's most sophisticated yet frequently misunderstood features. Unlike OSPF's simple bandwidth-based cost or RIP's primitive hop count, EIGRP can consider multiple path characteristics including bandwidth, delay, reliability, load, and MTU, creating a nuanced metric that reflects true path quality. However, understanding what EIGRP actually uses by default versus what it theoretically can use is critical for both CCNA exam success and practical troubleshooting. The EIGRP metric formula is: Metric = 256 × [(K1 × Bandwidth) + (K2 × Bandwidth)/(256 - Load) + K3 × Delay] × [K5/(Reliability + K4)], where K1 through K5 are configurable constants called K values. This intimidating formula simplifies dramatically with default K values: K1=1, K2=0, K3=1, K4=0, K5=0. With these defaults, the formula reduces to: Metric = 256 × (Bandwidth + Delay). This means EIGRP uses only bandwidth and delay by default, completely ignoring reliability, load, and MTU despite these values being advertised in EIGRP packets. Changing K values is strongly discouraged except in very specific circumstances with expert guidance, as it can destabilize routing. The bandwidth component is calculated as: Bandwidth = 10,000,000 / minimum bandwidth in kbps along the path. The constant 10,000,000 (or 10^7) represents 10 Gbps expressed in kbps. For a Gigabit Ethernet link (1,000,000 kbps), bandwidth = 10,000,000 / 1,000,000 = 10. For Fast Ethernet (100,000 kbps), bandwidth = 10,000,000 / 100,000 = 100. For a T1 line (1,544 kbps), bandwidth = 10,000,000 / 1,544 ≈ 6,476. Critically, EIGRP uses the minimum (slowest) bandwidth along the entire path, not the sum. A path traversing Gigabit Ethernet, then T1, then Gigabit Ethernet uses the T1's bandwidth value (6,476) since it's the bottleneck. The delay component represents cumulative delay along the path, measured in tens of microseconds (not milliseconds). EIGRP sums the delay values of all outbound interfaces along the path. Default interface delays are: Gigabit Ethernet = 10 (meaning 100 microseconds), Fast Ethernet = 100 (1,000 microseconds or 1 ms), Serial T1 = 20,000 (200,000 microseconds or 200 ms). These delays are interface-type dependent and configured using the 'delay' command, though modifying delay for metric manipulation is generally considered a hack—use bandwidth statements instead for clarity. To calculate a complete EIGRP metric, determine the minimum bandwidth and sum of delays, then apply the formula. Example: A path crosses three interfaces: GigE (BW=10, Delay=10), Fast Ethernet (BW=100, Delay=100), and another GigE (BW=10, Delay=10). Minimum bandwidth = 100 (from the FastE). Sum of delays = 10 + 100 + 10 = 120. Metric = 256 × (100 + 120) = 256 × 220 = 56,320. This metric is purely mathematical—lower values indicate better paths. When comparing routes, EIGRP installs the route with the lowest metric as the successor. The bandwidth value used in EIGRP calculation comes from the interface's configured bandwidth (not necessarily the actual physical bandwidth). Use 'bandwidth [kbps]' in interface configuration to set the value EIGRP uses for metric calculation. This doesn't change actual interface speed—it only affects routing protocol metric calculations and QoS mechanisms. On serial interfaces, always configure bandwidth to match the actual line speed since they default to T1 speed (1544 kbps) regardless of actual provisioning. Misconfigured bandwidth causes suboptimal routing—EIGRP might choose a slower path if bandwidth values don't reflect reality. Understanding EIGRP metric calculation enables effective path engineering and troubleshooting unexpected route selection.

The Diffusing Update Algorithm (DUAL) represents EIGRP's most sophisticated and mathematically rigorous feature, providing guaranteed loop-free routing at every instant—even during convergence. Unlike distance-vector protocols that can form transient loops during topology changes or link-state protocols that must complete SPF calculation before trusting new routes, DUAL ensures that backup routes (if they exist) are provably loop-free before any failure occurs, enabling instant failover. Understanding DUAL's key concepts and how it guarantees loop freedom is essential for CCNA certification and for appreciating EIGRP's convergence capabilities. The Successor is DUAL's term for the best current route to a destination—the route with the lowest metric. This is the route EIGRP installs in the routing table and uses for forwarding packets. If multiple routes have identical lowest metrics, EIGRP can install all of them (up to the maximum paths configured, default 4) for equal-cost load balancing. The successor's metric becomes the Feasible Distance (FD) for the destination—essentially the best metric your router has ever seen to reach this network. The FD serves as the reference value against which potential backup routes are evaluated. The Feasible Successor (FS) is a backup route that meets the Feasibility Condition, guaranteeing it's loop-free. Not all alternate routes qualify as feasible successors—only those that can mathematically prove they don't route through your router to reach the destination. This is where DUAL's genius appears. Each neighbor advertising a route includes their metric to the destination, called the Reported Distance (RD) or Advertised Distance (AD). The Feasibility Condition states: For a route to qualify as a feasible successor, its Reported Distance must be less than your current Feasible Distance (RD < FD). This simple mathematical test guarantees loop freedom. Why does RD < FD guarantee loop freedom? Consider the logic: If a neighbor's metric to the destination (RD) is less than your best metric (FD), that neighbor cannot possibly be using you as part of their path to the destination—they have a better path than any path through you. Therefore, forwarding to this neighbor cannot create a loop. If RD >= FD, you cannot prove the neighbor isn't routing through you, so the route doesn't qualify as feasible successor despite being an alternate path. It might still be loop-free, but DUAL cannot prove it without running a query process. When the successor route fails (link down, neighbor disappears), DUAL checks if a feasible successor exists. If yes, DUAL immediately promotes the FS to successor and installs it in the routing table—convergence is instantaneous, typically subsecond. No queries sent, no recalculation needed, no waiting periods. This is DUAL's primary advantage: pre-calculated, guaranteed loop-free backup routes enable immediate failover. If no feasible successor exists, DUAL must go active for the destination, sending queries to all neighbors asking if they have a path. This query process takes longer (seconds) but is still relatively fast compared to full routing protocol reconvergence. The topology table ('show ip eigrp topology') displays all routes learned from all neighbors, showing which routes are successors, which are feasible successors, and which are alternate routes that don't meet the feasibility condition. Understanding how to read the topology table is crucial for troubleshooting. Successors are marked with 'P' (passive state), feasible successors show with multiple entries and lower RD than FD. Routes in Active state indicate DUAL is querying for a path—prolonged Active state (Stuck-in-Active or SIA) indicates network instability or reachability issues. Optimizing EIGRP topology for maximum feasible successors improves network resilience. Networks designed with symmetric paths and multiple redundant links typically have more feasible successors. Asymmetric routing or hub-and-spoke topologies often lack feasible successors because spoke routers' metrics to hub routes exceed the hub's direct metrics. Understanding DUAL enables both effective EIGRP deployment and rapid troubleshooting of convergence issues in production networks.

Configuring EIGRP requires understanding several key commands and concepts that work together to establish neighbor relationships, advertise networks, and control routing behavior. While basic EIGRP configuration is relatively straightforward, production deployments demand attention to autonomous system numbering, network statements, summarization control, and passive interfaces to ensure efficient and secure routing operation. Mastering EIGRP configuration is fundamental for both CCNA certification and practical network engineering. EIGRP configuration begins with enabling the EIGRP routing process using 'router eigrp [AS-number]' in global configuration mode. The AS number (Autonomous System number) is a value from 1 to 65535 that identifies this EIGRP routing process. Critically, the AS number must match between routers for them to become EIGRP neighbors and exchange routes—routers in AS 100 ignore EIGRP packets from routers in AS 200. The AS number is locally significant (unlike BGP AS numbers, which are globally assigned), meaning you choose it freely for your network. Common practice uses values like 1, 100, or values derived from organizational identifiers. You can run multiple EIGRP processes on one router with different AS numbers, enabling route redistribution between processes or separate routing domains, though this is uncommon. The network statement activates EIGRP on interfaces and determines which networks are advertised. The syntax is 'network [network-address] [wildcard-mask]' in EIGRP router configuration mode. Unlike OSPF's network statement which also assigns interfaces to areas, EIGRP's network statement simply enables EIGRP on matching interfaces and advertises their directly connected networks. For example, 'network 192.168.1.0 0.0.0.255' enables EIGRP on all interfaces with IPs in the 192.168.1.0/24 subnet. The wildcard mask uses the same inverse-of-subnet-mask concept as ACLs: 0 bits must match, 1 bits don't matter. You can use broad network statements like 'network 10.0.0.0 0.255.255.255' to enable EIGRP on all 10.x.x.x addresses, or very specific statements like 'network 192.168.1.1 0.0.0.0' to enable it on only one IP address. Classful auto-summarization, enabled by default in older IOS versions, causes EIGRP to automatically summarize routes at classful network boundaries (Class A, B, C). This legacy behavior creates serious problems in modern VLSM networks. For example, if you have subnets 172.16.1.0/24 and 172.16.2.0/24, auto-summary causes EIGRP to advertise them as a single 172.16.0.0/16 route, potentially causing routing loops or blackholes. The command 'no auto-summary' in EIGRP router configuration mode disables this classful behavior, allowing proper VLSM operation. In modern IOS (15.0+), 'no auto-summary' is default, but explicitly configuring it ensures correct behavior and documents intent. Always use 'no auto-summary' in CCNA exam scenarios unless specifically instructed otherwise. Passive interfaces prevent EIGRP from sending hello packets and forming adjacencies on specified interfaces while still advertising those interfaces' networks into EIGRP. This is crucial for security and efficiency on interfaces connecting to end-user networks, DMZs, or external networks. Use 'passive-interface [interface-id]' in EIGRP router configuration mode. For example, 'passive-interface GigabitEthernet0/0' prevents EIGRP neighbor formation on that interface but continues advertising the GigabitEthernet0/0 network to neighbors on other interfaces. Alternatively, 'passive-interface default' makes all interfaces passive (secure default), and you selectively enable EIGRP with 'no passive-interface [interface-id]' on trunk interfaces connecting to other routers. This approach prevents accidental EIGRP neighbor formation on incorrectly connected ports. The router-id command explicitly sets EIGRP's router ID: 'eigrp router-id [IP-address]' in EIGRP router configuration mode. While EIGRP automatically selects a router ID (highest loopback IP, or highest physical interface IP if no loopback), explicitly configuring it ensures predictability and stability. Use a unique, easily identifiable value for each router. Verification commands include 'show ip protocols' (displays EIGRP configuration, AS number, networks), 'show ip eigrp neighbors' (neighbor adjacencies), and 'show ip route eigrp' (EIGRP routes in routing table). Understanding these configuration elements enables effective EIGRP deployment in production networks.

Effective EIGRP verification and troubleshooting requires systematic examination of neighbor relationships, topology tables, routing tables, and protocol parameters. Network engineers must quickly identify configuration mismatches, detect convergence problems, and validate proper EIGRP operation to maintain stable routing. Understanding the essential show commands, common failure scenarios, and diagnostic methodologies is critical for both CCNA certification and production network support. The primary EIGRP verification command is 'show ip eigrp neighbors', which displays all EIGRP neighbor relationships and their status. The output shows each neighbor's IP address, the interface through which the neighbor is reached, hold time (seconds remaining before declaring neighbor dead, resets to maximum with each hello), uptime (how long the adjacency has existed), Smooth Round Trip Timer (SRTT, average time for reliable packet acknowledgment), Retransmission Timeout (RTO, timeout before retransmitting), Queue Count (packets waiting to be sent), and Sequence Number (last packet received). High or increasing queue counts indicate congestion or the neighbor being overwhelmed. Hold time counting down to zero then disappearing means neighbor loss—investigate physical layer, hello/hold timer mismatches, or authentication failures. Neighbor states are simpler than OSPF's complex state machine: Pending (neighbor discovered but not yet acknowledged bidirectional communication) or Up (fully formed adjacency). Most neighbors rapidly transition from Pending to Up. Neighbors stuck in Pending indicate one-way communication—verify physical layer, check for ACLs blocking EIGRP packets (protocol 88), confirm network statements enable EIGRP on both sides. Neighbors flapping (repeatedly down/up) suggest marginal physical layer issues, timer mismatches, or routing loops causing instability. The 'show ip eigrp topology' command displays the complete EIGRP topology table, showing all routes learned from all neighbors regardless of whether they're installed in the routing table. This is EIGRP's equivalent to OSPF's link-state database but simpler—just routes and metrics, not complete topology. For each destination, the topology table shows the Feasible Distance (FD, best metric), Successor route(s) with their metrics and next-hops, and any Feasible Successor routes that meet the loop-free criteria. Routes marked 'P' are Passive (stable), while 'A' indicates Active (DUAL is querying for a path—this should be brief, prolonged Active indicates problems). The topology table is invaluable for understanding why EIGRP chose specific paths and whether backup routes exist. Stuck-in-Active (SIA) represents EIGRP's most notorious problem. When a route goes Active, DUAL sends queries to all neighbors. Each neighbor must reply with either a route or confirmation they have no route. If replies don't arrive within the Active timer (default 3 minutes), the route becomes SIA and EIGRP tears down the neighbor relationship that failed to respond. SIA typically indicates network design problems: too many routers in the query scope (solve with route summarization creating query boundaries), unreliable links causing packet loss (fix physical layer), routing loops (check redistribution configurations), or underpowered routers unable to process queries (upgrade hardware). The 'show ip eigrp topology active' command shows routes currently Active—seeing many routes stuck Active for extended periods indicates serious network issues requiring investigation. The 'show ip route eigrp' command displays only EIGRP routes installed in the routing table (successors). Routes marked with 'D' indicate internal EIGRP routes, 'D EX' indicates external routes (redistributed from other protocols). The command 'show ip protocols' displays comprehensive EIGRP configuration: AS number, router ID, K values, networks being advertised, passive interfaces, administrative distance, and maximum paths for load balancing. This is essential for verifying configuration matches across routers. Common troubleshooting issues include neighbor adjacency failures due to mismatched AS numbers (check 'show ip protocols' on both routers), mismatched K values (extremely rare since defaults work, but if modified must match exactly), ACLs blocking EIGRP multicast (224.0.0.10) or unicast (during query processes), mismatched authentication (if configured), and network statements not covering interface IPs (interfaces won't participate in EIGRP). Routes not appearing in the routing table despite being in topology table might indicate administrative distance issues, better routes from other sources, or maximum-paths limitations. Understanding EIGRP verification commands and common failure patterns enables rapid troubleshooting and reliable routing operation in production networks.