Optimize Analytic Engine

The Analytic Engine needs 1GB virtual swap space and a disk subsystem with 10GB available free space. In a production environment, the disk subsystem must have a redundant array of independent disks (RAID 5), an UltraWide SCSI, and 10K to 15K RPM drives.

Also in a production environment, the Analytic Engine needs additional hard drive space for log files. The recommended amount is 100MB; the engine needs 50MB of space for each log file.

Optimize Web Service Data Collector

Each Web Service Data Collector needs 128MB of virtual swap space. In a production environment, each Web Service Data Collector needs additional hard drive space for log files. The recommended amount is 75MB; each Web Service Data Collector needs 5MB of hard drive space for each log file. More hard drive space might be needed if you use debug level or higher logging.

You might need additional RAM and hard drive space, depending on the number of process instances you import and store in Process Performance Manager. The more process instances, the more RAM and hard drive space you will need.

If you use persistent topics or queues, or persistent messages, Universal Messaging needs additional hard drive space to persist the published data. The amount of space required would be loosely based on this equation: messages per second x message size x message time to live. If you store messages in memory only, Universal Messaging needs additional heap within the JVM to hold references to these messages. The heap size required is based on the same equation.

Universal Messaging realm servers support high-performance spin locks. If you enable spin locking for a realm server, and then add the realm server to a cluster, the realm server needs two additional CPUs to handle high-performance cluster event processing.

If you enable a Universal Messaging realm server to use shared memory, the realm server needs two additional CPUs for each client that connects to it with shared memory.