Well, that's almost correct, but let me provide a more succinct definition.
We make a clear distinction between *forwarding* and *routing*, and anyone
who works with routers needs to understand the distinction. Routing is the
process of building path information about the network, whether from dynamic
routing protocols or from static routes. A routing table within a distance
vector (e.g. RIP, IGRP) or link-state (e.g. OSPF, EIGRP, Dual IS-IS) is an
example of routing. The routing tables maintained by a router are generally
referred to as RIB's, or routing information bases.
Forwarding is the act of actually forwarding packets. Traditionally, the
FIB (forwarding information base) is either pre-populated (topology-driven)
or dynamically populated (traffic-driven on-demand cache) from information
gleaned from the RIB.
It is important to make this distinction since there can be many instances
where routing processes can do extraordinary things (topology recalculation
using SPF, feasible successor recalculation, et al.) which may have no
effect on the forwarding mechanisms. Likewise, there are also instances
where excessive route instability can cause massive amounts of pain (e.g.
At 12:35 PM 2/24/97 -0800, Kenneth kron wrote:
>IP forwarding is not routing. A system becomes a "router" when it starts
>IP forwarding but everytime a system sends an IP packet it *routes*, it
>decides which of it's interfaces to use to transmit the packet and whether
>to forward it to a router or send it directly to it's destination.
Paul Ferguson || ||
Consulting Engineering || ||
Herndon, Virginia USA |||| ||||
tel: +1.703.397.5938 ..:||||||:..:||||||:..
e-mail: pferguso @
com c i s c o S y s t e m s