EIGRP

EIGRP sends an UPDATE only when a route is added or its metric changes, carries only those prefixes, and the change stops spreading where no router's best path changes; RIP resends its whole table every 30 seconds.

With default K-values, the EIGRP classic metric is 256 × (10^7 / slowest bandwidth in kb/s + summed delay in tens of microseconds). Only the slowest link sets the bandwidth term, so a longer path with a faster bottleneck can win.

A neighbour meets EIGRP's feasibility condition when its reported distance to a prefix is strictly below this router's feasible distance, its best distance since the route last went passive; that neighbour is a feasible successor, a guaranteed loop-free backup.

First check that hellos cross the link: EIGRP enabled, IP protocol 88 and 224.0.0.10 not filtered. Then the three things RFC 7868 makes neighbours agree on: the AS number, the K-values and authentication. Hello and hold timers need not match.

EIGRP sends UPDATE, QUERY and REPLY packets, with their SIA sub-types, reliably: each carries a nonzero sequence number that must be acknowledged. HELLOs go unreliably with sequence 0, and an ACK is a HELLO with no data and a nonzero acknowledgement number.

The successor is the path with the lowest computed distance, which sets the feasible distance; any other path whose reported distance is strictly below that feasible distance is a feasible successor. Only successors, including equal-cost ones, reach the routing table.

EIGRP gives precomputed feasible-successor failover, summarisation on any interface and implementation-provided unequal-cost load balancing, but is essentially single-vendor; OSPF is an open standard whose shared area database and area hierarchy demand more design but give visibility and interoperability.

An EIGRP route is stuck in active when a queried neighbour fails to reply in time; an SIA-QUERY asks whether it is still computing, an SIA-REPLY says yes, and a silent neighbour loses the route or its whole adjacency.

Variance, an implementation feature RFC 7868 does not define, admits a path only if it is a feasible successor and its computed distance lies within variance times the best metric; feasibility is required because only those paths are proven loop-free.

DUAL, the Diffusing Update Algorithm, is how EIGRP picks routes: it keeps backup next hops that pass a loop-freedom test and switches to one at once; when none exists, it freezes the route and queries its neighbours before choosing again.

An EIGRP router multicasts hello packets to 224.0.0.10, by default every 5 seconds, and lists each sender in its neighbour table. Each hello advertises a hold time, 15 seconds by default; a neighbour silent for that long is declared down.

The EIGRP topology table records every path any neighbour advertised for each destination, with its distances and route state; the routing table receives only the successor paths actually used to forward packets.

OSPF is an open IETF Standards Track protocol (RFC 2328 for IPv4, RFC 5340 for IPv6) implemented across router vendors, while EIGRP was one vendor's protocol until Informational RFC 7868 in 2016 and still has few other implementations.

K-values decide how each EIGRP router ranks paths from the vector its neighbours advertise. With different weights, two routers can rank the same paths in opposite orders, and DUAL's distance comparisons assume one shared metric, so RFC 7868 forbids the adjacency.

No. Each EIGRP router advertises its own hold time in its hellos, and neighbours time it out on that value, so timers may differ. But if a router's hello interval exceeds the hold time it advertises, neighbours drop it between hellos.

The router multicasts the UPDATE once with a new sequence number and records it on every neighbour's retransmit list; each neighbour unicasts an acknowledgement of that number, and only neighbours that stay silent get unicast retransmissions.

A new EIGRP neighbour has no topology to apply changes to, so each router sends it the full table by reliable unicast, after an empty INIT-flagged UPDATE proves unicast works both ways. Once synchronised, only changes are multicast.

Load and reliability change with traffic and errors, so an EIGRP metric built on them would change too: traffic would swing between paths, every swing would trigger updates and possible recomputation, and every router would have to change K-values together.

The EIGRP route goes active and unusable; the router queries its neighbours, unaffected ones reply at once, affected ones go active and query onward, and once every reply is in it resets its feasible distance and picks the best answer.

An EIGRP query ends at any router with no entry for the failed prefix, which replies at once with an infinite metric; summarisation hides specific prefixes beyond a boundary, and stub routing, an implementation feature, exempts routers that offer no transit.

EIGRP adjacencies end on local link-down, hold-time expiry, 16 unacknowledged retransmissions or a signalled reset. Behind a switch the link stays up, so a dead neighbour on a quiet link is found when its hold time expires, up to 15 seconds by default.

Healthy hellos prove only that small unacknowledged multicasts arrive. Resets during the table exchange point at reliable delivery failing: unicast UPDATEs, retransmissions or their acknowledgements are lost, so after the retry limit the neighbour is reset and the cycle repeats.

With a feasible successor, an EIGRP router switches locally and the route stays passive; OSPF floods a new LSA and every router in the area reruns SPF. Without one, EIGRP must query its neighbours and can converge more slowly.

EIGRP can summarise on any interface, so hubs send spokes a summary or default and an implementation's stub feature stops hubs querying spokes; OSPF summarises only at area borders, so spokes need their own, ideally stub, areas.

Tag each route where it is redistributed and refuse to redistribute it back: OSPF carries a 32-bit External Route Tag, EIGRP externals an administrative tag. Also seed metrics deliberately, migrate site by site and check route-source preference at the borders.

The classic EIGRP bandwidth term, 10^7 over kb/s, truncates to 0 above 10 Gb/s, and delay counts in 10 µs units. Wide metrics carry unscaled bandwidth and picosecond delay, scale by 65,536 and compare a 64-bit metric.

A neighbour routing through this EIGRP router must report more than this router's feasible distance, having added positive link costs to its advertisement; a lower report proves independence. But a long independent path also reports high, so it is rejected too.

Stop sending and accepting EIGRP hellos on interfaces that face only hosts, and authenticate every adjacency that must exist, preferably with SHA2-256. The AS number and K-values are no defence: every hello carries them in clear text.

RFC 7868 requires EIGRP to prefer internal routes over external ones regardless of metric: while any internal path exists, successors are chosen only among internal routes, which keeps redistributed copies from overriding routes that originate inside the EIGRP domain.