Essentially, it is made up of various forms you must fill out in order to populate a database that would then spitback the information at you, in a slightly more organized or graphical way.
If the server farm is doing a lot of work to take in cell tower distance data and spitback car locations, it can certainly do each calculation faster than any endpoint, particularly a small one, but what happens when 1 million endpoints are hitting the cloud at the same time?