RIB (Routing Information Base) update is the last post of OSPF fast convergence series whereas earlier posts have already covered OSPF Event Detection, OSPF Event Propagation and OSPF Event Processing.
Whenever there is change of interface state arises, OSPF detect that change and according to that Dijkstra runs the algorithm and update the RIB table according to that. This process is time consuming and requires lot of CPU and memory cycles.
Reducing convergence time depends how fast the OSPF detects the state change and notifies the RIB. In OSPF we normally perform the summarization of the prefixes on ABR and ASBR, this is done only to save the CPU cycles so that every state change information is not flooded across the OSPF routing domain.
Secondly, interface purge command can be used to reduce convergence time. This optimization allows interface link state changes to directly impact RIB rather than waiting for routing protocol updates to RIB to take place.
(Read More About OSPF High Availability)