Overview:
Routers have many different routes in their tables, but need to have a fast way (forwarding decision) to find the best path to a destination network.
Study Notes:
- A prefix is a network address and a route mask
- Typically, the forwarding decision in a router is based on three processes - Routing protocols, Routing table and Forwarding Decision
- Routing protocols
- The routing protocol uses its algorithm (its metric) to figure out the best path to a network
- Routes with different prefix lengths get entered into the routing table because they are viewed as being different routes
- Routes with same prefix lengths use the administrative distance as the tie-breaker for the route to enter into the routing table. i.e. Internal EIGRP (AD 90) will win out over OSPF (AD 120) if both protocols are have a route to the same destination network.
- Routing table
- Route with longest prefix length is used for the routing decision
- Routing protocols
#show ip route
172.16.0.0/16 is variably subnetted, 3 subnets, 2 masks
S 172.16.0.0/24 [1/0] via 10.0.0.1
S 172.16.2.0/24 [1/0] is directly connected G0/1
S 172.16.0.128/25 [1/0] via 10.0.0.1
#show ip route cache
Prefix/Length Age Interface Next-hop
172.16.0.0/25 00:02:35 G0/1 10.0.0.2
172.16.0.128/25 00:23:31 G0/1 10.0.0.2
-
- Forwarding decision (switches packets)
- Once the longest prefix has been matched the router sends the packet to the exit interface
- The packet moves down the TCP/IP stack from layer 3 to layer 2 and gets encapsulated
- Depending on which type of switching is used determines how fast the frame gets put on the wire
- Process switching - waits for the CPU to process it.
- Fast switching - first packet is process switched, others use cache
- Cisco Express Forwarding - a router might have tons of routes, but usually only a few neighbors. It makes sense then to preconstruct the L2 frame headers and egress interfaces and save them in the adjacency table. This trades memory for speed.
- Forwarding decision (switches packets)
Links: