The Distant Vector algorithm is based on information exchange between adjacent routers. Two routers are adjacent if there is a direct link between them, i.e. if they have an attachment on the same local network. This algorithm is very simple to implement, but its performance is limited and its scaling factor resistance is small. An analogy can be made with hotel doormen who share gossip with doormen in other hotels. The protocol can be very chatty and false rumors can persist for a very long time.
RIP mainly enables the automatization of learning of prefixes present on a site. It prevents errors that could occur during the manual configuration of routers.
Each router initially only knows the cost of its own links. Routers broadcast their rudimentary routing table consisting of their accessible neighbors and the link cost towards adjacent nodes. When a router receives a new table, it carries out the Bellman-Ford algorithm for each input of the received table:
— If the input is not in the table, it adds it;
— If the cost of the proposed route by the table plus the cost of the traveled route is smaller than the stored route, the table is modified to take this new route into account;
— Otherwise there is no change.