Introduction

IBM® FileNet® P8 provides a comprehensive, enterprise-class content lifecycle, document management, and workflow platform, for Enterprise Content Management (ECM) and Business Process Management (BPM).

IBM FileNet P8 provides application APIs and tools to enable customers and business partners to create custom applications, as well as customizable out-of-the-box clients.

The P8 Content Engine (CE) server is designed to manage enterprise-wide documents and custom objects. It manages the classes, properties, storage, and metadata that form the foundation of an Enterprise Content Management system. It is installed on an application server with a database back-end on a separate guest. Document content can be stored either in the back-end database or, more commonly, in a file store.

The P8 Process Engine (PE) server manages all aspects of business processes (also called workflows), such as process execution, process routing, rules management, process simulation and modelling, and workflow analysis. For the FileNet P8 5.1 release used in these tests, PE is a standalone Java™ application with a database back-end on a separate guest (same database as for CE in our environment). Note that starting with FileNet P8 version 5.2, CE and PE are now combined into the Content Platform Engine (CPE), running under control of the application server, and with a single combined CPE database.

This white paper presents the performance and scalability characteristics of the IBM FileNet P8 5.1 Content Engine (CE) and Process Engine (PE), in a virtualized environment running as Linux™ guests on z/VM® on IBM System z®. FileNet P8 is an ideal candidate for deployment in a virtualized environment, with all its flexibility for resource configurations and virtual network features where the communication between the components is implemented inside the local memory without hardware requirements. Performance tests driven using the CE and PE Java APIs demonstrate how FileNet P8 5.1 on Linux for System z can scale to effectively exploit an increasing number of Linux processors with good response times and throughput rates, with a workload that represents a realistic production transaction mix.

The performance results reported represent test data models and workloads run in an isolated laboratory on specific operating environments and system configurations. Actual performance in real customer environments with production workloads may vary significantly, and will depend on the unique circumstances of each customer's configuration and workload, and many factors including other applications running on the systems and configuration of the storage or network.