i have made this kind of software sometime back.There are 2 approaches to it based on the whether you want to have client server configuration ( which needs to have client running on all n/w pcs.) or you only required to haev server pull data from n/w pcs (this will not require any installation on n/w machines)
So former approach is push mechanism. Clients installed on n/w machines will run at predetermined time and gather all the information and push it to centralised db. Later is pull mechanism in which server will put batch file or c or
java code (based on n/w machine compatibility) in remote machines and schedule them to run after sometime which will generate files on the local machines and in second pass server will collect all files from different machines and remove them from remote machines. so no footprints of inventory will be there.
I have tried later one.