Need opinions on the feasability of huge aggregrate graphs

[sheol]

Hello all,
I work for a University and we're looking at the efficiency of our lab usage, basically we know how often they're booked but not how often nor how many users are in each class. Total number of machines I'm concerned with at this stage is around 800.

I was going to query each machine for just logged in user as per the standard Windows template but was wondering how possible and how usable Cacti would be to not only group these machines by labs of 26 for average log in times but also one monster graph to get an average logged in time for all 800 machines. Also, being able to print the hard data out in some sort of tabled form would be beneficial. There's no point gathering all of this information if it's not usable. I've looked at nagios, zabbix, mrtg, rrdtool and cacti so far and would welcome any suggestions.

[rony]

You can create aggregate graphs of the hosts users to consolidate the total number of users in a lab.

As for outputted data, you can use an rrdtool export or rrdtool fetch to get the data you need out of the rrd files.