Large Scale Seperation Model Example

[goldcougar]

Many people have asked if anyone has an example of a large scale implementation of the Seperation Model...so, I thought I would post information about our setup, and also ask for some feed back.

We have a solution using the seperatin model with 6 total files. 1 Interface File, and 5 Data Files.

Our Data files total to 162 tables, with the highest number of records for a table over 2 Million.

Our Interface File contains 552 Relationships, 666 Layouts, 353 Value Lists, 30 Custom Functions, and 27 File References.

So, its a big solution desined to run up to 150 users at a time. However, we recently ran a load test, and after 50 users, the server's CPU was maxed out (Dual G5 XServe), and the solution was way too slow to use. We're trying to seperate the solution accross multiple servers, but are having problems. I have a post at FM Forums post

We are also looking at having the Interface File on the Local user's machine to increase speed. Does anyone have any experience with this?

[Fitch]

Look for the usual suspects, e.g. unstored calculated fields. Also, why are there 27 file references in the interface file if there are only 5 data files?

[goldcougar]

The 5 data files exclusively work with the interface file. The other File References are for links to other databases on our server. Our entire system consists of 5 FileMaker Servers totaling to over 250 databases. So, sometimes in our solution of the Interface File we need to lookup data from the other databases. Ex, if we have a database of customer info that we have had for years, with this new system, the interface file needs to get info from there.

We have a lot of unstored calculation fields, but there is no way around it in this solution.