The PPM front-end communicates exclusively with the master server. The master server passes on queries to the connected sub-servers. The calculation of analysis results takes place simultaneously on the sub-servers. The master server aggregates the results from the sub-servers and sends the overall result to the front-end. To achieve an optimum parallelization result, all sub-servers should, as far as possible, be evenly utilized. Overall system performance is determined by the slowest sub-server.
System configurations are imported using the master server. The master server passes them on to the sub-servers. All PPM servers have the same configuration upon completion of the configuration process.
Example: Data flow between front-end and scaled server system