It looks to me like they are using SNMP to query their switch for traffic data on certain switchports, but they made their own script that runs check_snmp and queries input and output kbytes/sec at the same time, and then changes the outgoing traffic value to a negative value. That way when your nagios to RRD script gets a hold of it, you can define the negative value as outgoing traffic and the positive value as incoming traffic and make a pretty graph. Lots of people just overlap incoming and outgoing traffic, these guys had a good idea. With their way you get the outgoing going downwards (negative) and the incoming going upwards when it is graphed.
To get the check data into an RRD file, they are using n2rrd which is a perl script that grabs nagios check data and puts it into an rrd file. To display that RRD with pretty colours and legend and stuff, they are using rrd2graph.
Another way to grab traffic information would be to use something like iptraf on the server itself you want bandwidth info from, and then parsing output from there.