DV routing protocols calculate a route’s metric based on the number of hops, this has the disadvantage that a route that is slower may be chosen because it has less hops:
If a distance vector routing protocol such as RIP was running on the above topology it would choose the one with the least hops. In this case the route IOU1=>IOU2=>IOU3 is actually slower than the higher hop count route IOU1=>IOU4=>IOU5=>IOU3.
In order to prevent routing loops distance vector protocols have a split horizon. This means they only share routes on one side of the horizon to the other.
Here is an excerp from the Official CCNA 200-125 Exam guide:
Distance vector protocols poison dead routes by using an infinite metric.
Link-State Routing Protocols
LS routing protocols create a virtual map of all of the links in a network and choose routes based on the highest bandwidth. The most popular link-state routing protocol is OSPF which implements the Dijkstra algorithm to calculate be best path in a network for each route.
If a link-state routing protocol was running in the above topology it would have chose the best route based on total speed not the amount of hops.
In order to create a topology OSPF creates a LSDB that contains all the information about all the links in that network, OSPF will then use a LSA (link state advertisement) to advertise this with it’s neighbors. LSU (Link state updates) are used to exchange information with neighbors until each neighbor has a full map of the entire topology.
This process is CPU intensive and means the LS routing protocols have a higher overhead than DV protocols.