12.2 How are routing decisions made?

In order to demonstrate how the router makes its routing decision, we will use the same demonstration network we have used in the previous demonstrations.

This time we will use a ping from PC0 to PC1. We will ignore what happens in the internet router at first, and just look at the decisions made by Router0. As the ICMP packet comes in from the internet router it will have the destination IP address of 192.168.0.10 in its Internet layer header. The destination IP address in the packet is 192.168.0.10. The routing table contains entries for the connected networks of 172.16.0.0/24, 192.168.0.0/24 and also the 1.1.1.0/30 network in which the packet came in from. Note that each connected network shows up twice, one with the code of C which is the network it is on, and one with the code of L – which is the exact IP address of the router’s port. Note also that there is an entry marked S* at the end of the list. We will come back to this in a moment.

Now, back to the routing decision. The router will look at the destination address of the packet, in this case 192.168.0.10, and compare this address with the addresses in the routing table. The router will try and match the address with the entries in the routing table. If it finds multiple matches it will choose the best match – the one with the most bits matched will be selected. In this case there is a match with the network 192.168.0.0/24, which is on GigabitEthernet0/0. The packet will now be sent out of this interface and on to the switch, which will then send it on to PC1. PC1 will then respond with an echo-reply, which will have the source address of itself, and the destination of 10.0.0.10, which was the source of the echo request in the first place. When Router0 receives this packet it will look for 10.0.0.10 in its routing table. As you can see, there is no entry for any 10.0.0.0 network in this routing table. If the router does not find any match, it should drop the packet completely. This would cause the ping to fail and is not a very desirable outcome. We could either fix this by putting static routes into the routing table for all networks we wish to reach, or we can put a default route into it. This default route would then direct any traffic for networks not in the routing table to a valid destination. This default route has already been put into the routing table and is the entry with the code of S* we mentioned earlier in this presentation. This S* entry will also be listed above the routing table as the gateway of last resort when you do a show ip route.

So, back to our ping packet. There is no match for 10.0.0.0 in the routing table so the router will use the default route, which will send the packet out of GigabitEthernet0/2, which points towards the internet router. The ping will now continue on its journey to PC0. Now let’s have a quick look at the command prompt output which has happened while we were watching our ping packets going from the source to the destination and back. The first three ping packets have in fact failed, and the fourth one has succeeded. You should be able to guess why this happened. Think further back in this session to the ARP problem. The reason for the first failures is because the ARP process needed time to obtain all the MAC addresses for each section of the network – so PC0 needs to find the internet router’s MAC address, the internet router needs to find R0’s mac address, R0 needs to find PC1’s MAC address. This all takes time and means that some pings may be dropped. Remember we told you that these MAC addresses are cached for a period of time. This can be proved when you execute the ping again soon after. This time all four pings pass.