There are three classifications for routing protocols: Distance Vector, Link State, and Path Vector.
Distance Vector Routing Protocols (like RIPv2) are also termed “route by rumor” as the devices which participate within their AS do not contain a full map of the network as you’d see with a Link State protocol. They use a Bellman-Ford algorithm based upon distance and direction, with the path preferring the least number of hops (or lowest metric) toward a destination. When deciding on a routing action, each participating device only has the metric and egress interface for its calculations, and each router along the path is like a sign post.
EIGRP also qualifies as a distance vector protocol, though it is often termed as an “advanced” distance vector routing protocol or “hybrid” protocol, because it uses hello packets instead of sending periodic updates, and does not advertise or change routing table information unless something in the topology changes.
Distance Vector = vectors of distance (metric or hop count) and direction (next-hop neighbor)
Link State Routing Protocols (like OSPF) use the Dijkstra algorithm to make best path calculations for a destination route based upon a complete map of the network; instead of sending partial updates to neighbors, every router in the network obtains a copy of the routing table and puts itself as the root.
Link State = based upon SPF algorithm (shortest path first)
IS-IS is also a link-state protocol.
Path Vector Routing Protocols not only account for the existence of remote routing destinations, but the complete path for their reachability. The path information shared easily identifies loops. Advertisements list different path attributes, and there are many path selection tuning options. The only widely used path vector protocol is BGP.
Path Vector = vectors of path (list of path attributes) and direction (next-hop neighbor)
Stuck-in-Active: Journal of an IT-Network Administrator 