This chapter covers the basics of IP forwarding – what a router does when it gets a packet. How does the router make its fowarding decision?
“Forwarding” and “routing” are synonyms. Routers follow a generic logic flow to determine how to route a packet (assuming inbound on an ethernet interface):
- The router receives an inbound frame and checks FCS. If FCS fails, the packets is chucked in the bit bucket.
- If the frame is good, the ethernet Type field is checked, and the packet liberated from its frame. The layer 2 info (data link header and trailer) is trashed.
- So now we have a packet. If the packet is an IP packet, the router performs a routing table lookup, seeking the most specific route (prefix) it can find matching the destination of the packet.
- So now we found where the packet needs to go. The routing table entry includes the interface and next-hop this packet should use for egress. The router uses this information to throw new layer 2 back on the packet.
- The router updates the time-to-live field (decrementing by one, prevents routing loops, remember?), and consequently recomputes the IP header checksum.
- Finally, the router creates the new frame, specifically, the new Data Link layer information with the destination address, trailer and FCS.