Tuning Financial Management Application Servers

When tuning Oracle Hyperion Financial Management application servers, you should start with baseline tests to measure key user activities with representative user concurrency. When using multiple Financial Management clusters, often to separate reporting and Oracle Smart View for Office user activity from consolidation activity, it is likely to see different tuning changes affect each server differently based on the user task being measured. For example, an Financial Management application server primarily used for reporting will see no benefit by increasing NumConsolidationThreads, while a server running many consolidations should see improvements in consolidation times. Likewise, an application server primarily used for reporting would likely see better response times for repeated reports when MaxNumDataRecordsInRAM is set high enough to keep all records in memory, while a server running many consolidations is unlikely to see consolidation times improve. Another point to consider when deciding what role an application server will have is CPU speed, CPU core count, RAM amount and RAM speed. An application server primarily dedicated to running consolidations, running one Financial Management application that has intensive rules, will typically see the best performance with faster CPU clock speeds, with at least 8 cores, rather than using more CPUs/cores, but slower clock speed.