Redistribution is when a routing protocol is used to advertise routes that are learned by some other means, such as static routes, directly connected routes, or by another routing protocol.
While it is desirable to run a single routing protocol throughout your entire IP internetwork, multi-protocol routing is common for a number of reasons, including company mergers, multiple departments managed by multiple network administrators, and multi-vendor environments. Often, running different routing protocols is part of a network design. A multiple protocol environment makes redistribution a necessity.
Differences in routing protocol characteristics (such as metrics, administrative distance classful, and classless capabilities) can affect redistribution. For redistribution to be successful, these differences must be considered.
Note: When redistributing between a classful routing protocol (such as Routing Information Protocol Version 1 [RIP V1], Interior Gateway Routing Protocol [IGRP] ) and a classless routing protocol (such as Open Shortest Path First [OSPF]), RIP V2, Enhanced Interior Gateway Routing Protocol [EIGRP] ) will not advertise routes out an interface if those routes are on the same major network, but have a different mask than that particular interface.
Note: When redistributing routes into OSPF, a common mistake is to omit the subnets keyword from the redistribution statement. This prevents redistribution from taking place.
When there is mutual redistribution between routing protocols, there is the possibility of creating routing loops. A routing loop is when one routing protocol announces a route learned by another routing protocol through redistribution. To avoid this, never announce the information originally received from routing process x back into routing process x. This can be done with the use of distribution lists or route maps.