IP Routing

How EIGRP Works

Unlike traditional DV protocols such as RIP and IGRP, EIGRP does not rely on periodic updates: routing updates are sent only when there is a change. Remember that RIP and IGRP reset the invalid and flush timers upon receiving a route update. When a route is lost, the updates stop; the invalid and flush timers grow and grow (the timers are not reset), and, ultimately, the route is flushed from the routing table. This process of convergence assumes periodic updates. EIGRP’s approach has the advantage that network resources are not consumed by periodic updates. However, if a router dies, taking away all its downstream routes, how would EIGRP detect the loss of these routes? EIGRP relies on small hello packets to establish neighbor relationships and to detect the loss of a neighbor. Neighbor relationships are discussed in detail in the next section.

RIP and IGRP suffer from a major flaw: routing loops . Routing loops happen when information about the loss of a route does not reach all routers in the network because an update packet gets dropped or corrupted. These routers (that have not received the information about the loss of the route) inject bad routing information back into the network by telling their neighbors about the route they know. EIGRP uses reliable transmission for all updates between neighbors. Neighbors acknowledge the receipt of updates, and if an acknowledgment is not received, EIGRP retransmits the update.

RIP and IGRP employ a battery of techniques ...