12.3 How are switching decisions made?

In this part we are going to look at how a switch works. Switches work at the Network Access layer of the TCP/IP network model, which means that they use MAC addresses to forward frames of data. Although there are more powerful and advanced switches which use IP addresses and operate at the Internet layer as well, most switches know nothing about the Internet layer and only use MAC addresses to make their switching decisions. This means that normal switches do not generate ARP requests or have anything to do with IP addresses.

When a switch is first powered on, its MAC address table will be empty and will contain no addresses. This can also be simulated by issuing the command clear mac-address-table, which we have done in the Packet Tracer network. The MAC address table is currently empty. To populate the MAC table we must generate some traffic between the PCs. In this case we are going to start with a ping from PC2 to Router0. We have set up a table to the left of the diagram which will show you the current entries in the MAC table of Switch1 at the relevant time. PC2 pings Router0. The packet leaves the PC in an Ethernet frame and enters Switch1 on port FastEthernet0/1. When the switch receives the frame, it looks in the Ethernet frame at the source address in that frame. The address is then added to the MAC table and referenced to the port number the frame came in on. The switch looks in its table to see if the destination MAC address is there. Currently it is not, so the switch forwards the frame out of all ports except the one it came in on. This means that both Router0 and PC3 will receive the frame. Only Router0 will get a match to the destination MAC address; PC3 will ignore it completely. When Router0 replies back, the reply frame will contain its own MAC address as the source address and the destination MAC address of PC2. At this point the switch will look at the incoming frame on Gigabit0/2 and read the source MAC address and add it into the MAC table. It will then look at the destination MAC address and see if there is a match. There will now be a match as the MAC address of PC2 was already added to the table previously. The frame will now be sent out only on the port Fa0/1 and on to PC2.

This is the content of Switch1’s MAC address table, just to show you that the table we created to the left of the screen is genuine. In order to complete the network, we need to generate some traffic from PC3 so that its MAC address can be added to the table of the switch too. PC3 pings the router on 172.16.0.1. The frame arrives at the switch and its source address is added to the MAC table. The destination MAC address is checked against the entries in the table. There is a match for port Gigabit0/2, so the switch will send the frame back out on that port to the router. The ping reply frame will leave the router and enter the switch on Gigabit0/2. The switch will record the source MAC address in the table. Because it’s already there, the table will simply be updated with the time-out counter reset; in other words, the entry is refreshed. The switch will then look up the destination MAC address and find that it matches that of PC3 on port Fa0/2. The frame is then sent out of that port where it will be received by PC3, competing the ping. Here we can see the fully populated MAC address table of the switch.