I'm rolling out my solution to clients with the main FM database stored on FM13. A network with 10GB switches will provide the link between the two. The server is a virtualised machine and set up to deliver more than is needed. It's not used for anything else besides FM.
I need to ensure that the user experience is snappy, particularly when moving from one layout of data to another layout of data, but for the same customer. I have a main data table which has around 70,000 rows, requires lots of summaries and case calculations. Much of the data is warehoused overnight to a number-only table, to make page loads quicker and many reports are based from here. However, I still need to load some summary data from the main live table and this is what slows things down are where I need some help.
In one example, of the 70,000 rows, I only actually need to load around 18. I limit my searching first by customer (to get around 400 rows) then by row type to get to the 18 or so I need. This is done on a script with an initial find, followed by a constrain to further find the type.
Is this the most efficient way to get to the 18 rows I need? It feels right, ie minimising the found set before sorting, but I may have missed something fundamental as I've not got too much experience with the FM server environment.
Thanks in advance guru's and wise ones.