Originally Posted by wants
This is looking pretty sweet. Can't wait to try it out.
Just out of curiosity, how are you intending to handle all the database intensive filtering and searching that HEM and PT provide? I assume that you can cache the majority of basic graphs and analysis that people tend to view most often, but for the advanced filtering and analysis that HEM and PT do you will need a crapload of processing power to handle even a minimal simultaneous user load.
One of the reasons we're running this beta is to find out how many server resources we need to commit per-user as we scale. Actually, our initial rationale for wanting to do it web-based was that some of the more advanced statistics we were thinking about would be too resource intensive to run on a desktop and would require a distributed computing solution.