Hello Binoj,
short answer is to increase robustness of the protocol.
OSPF is a link state routing protocol.
All SPF routers that are members of a given area must have the same contents of LSDB ( at least for that area).
This is required to ensure loop free routing choices when executing SPF on one or multiple nodes.
Each node executes the SPF indipendently from other nodes using the local version of LSDB.
Protocol designers feelings were that it would be an always on process to keep all routers in synch.
As a result of this they introduced the age for each LSA: an LSA is a data structure with a built-in expiration time that can be checked by all nodes.
To be noted an LSA is identified by three information elements:
- the OSPF Router ID of the node that generated the data structure
- a Sequence number ( a 32 bit integer starting with 0x8000000)
- the age of LSA expressed in seconds that must be less then Max Age
LSA are propagated by flooding with a scope that depends on LSA type.
An OSPF node can only remove from the database the LSAs that it owns, it cannot remove LSAs received by other nodes (if LSA age has not expired). This is a key point.
Each node is responsible of providing a fresher version of each of self originated LSAs before the expiration of Max-age timer.
The new LSA will have the same OSPF RID, a new sequence number, a little age.
New LSA is flooded and all other nodes replace the older version with the newer one.
To be noted that the amount of bandwidth used to renew the LSAs is quite modest and it is not an issue at all even with thousands of routes/LSAs.
LSAs can be packed in a single OSPF packet up to MTU for transmission efficiency
see
http://www.ietf.org/rfc/rfc2328.txt
section 14
14. Aging The Link State Database
Hope to help
Giuseppe
