Longest Match Concept

Tommy2727

Greetings all I think I have this concept down but when it's referred to L3 when referencing the FIB table on a multilayer switch I think I'm confused..... I.E.. A packet is received on a interface, the layer 3 brain of the switch will check the header for the destination subnet, then the layer 2 brain will look at the FIB to see if it already knows about the destination. If not, consult the routing table and look at the longest match to forward to that subnet I.E /29 over a /16... Is that correct ???



Thanks...

Forsaken_GA

Yes, when making routing decisions, the longest prefix wins. For example, if you have an aggregate route for an entire /16, but also routes for /24's that fell under that /16, then when a destination comes in that can match on both, a /16 and a /24 prefix, the /24 prefix wins.

The other way to think of it is most specific triumphs. He who has better information, wins. If I know the guy I want to kill is in redmond, WA, it's better to know exactly which building to lob the tactical nuke at instead of just frying the entire city.

Tommy2727

Nice thanks....