Introduction
Performance benchmarks of Business Process Management (BPM) frameworks can be notoriously difficult to measure and accurately predict how they will scale under normal use and load. There is not one “blanket statement” that we can provide, which takes into account each customer’s unique business requirements and specific implementation design. When we’re asked about performance, it becomes one of those critical evaluation questions, in which we utter a frustrating reply of “it depends.”
Estimating performance is increasingly difficult when contemplating a general and flexible BPM framework such as the Red Hat JBoss BPM Suite ecosystem.
Our Approach
We believe that there needed to be a suitable way to alleviate this frustration and ambiguity, while providing a reusable framework that can be used for evaluation. We have developed a set of resources that provides the foundation from which to model and exercise scenarios in simulation. The results can be uniformly evaluated by providing a more concrete picture of how the Red Hat JBoss BPM Suite performs, while accounting for as much of the “it depends” aspects as feasible.
Fill out the form on this page to access the executive overview, technical supplementation, and data explorer page as well as the code for the test framework, and other related resources.
Data Explorer
The Data Explorer page provides interactive exploration of the results of the performance benchmark.
Table of Contents (preview)
- BPM Definitions
- Nature of BPM
- Analyzing BPM Performance
- Runtime Deployment Configurations
- Frameworks - KIE Performance Kit and JMeter
- Test Cases
- Analyzing BPM Performance
- Benchmark Runtime Configurations
- Persistence Adds Overhead
- Human Task Overhead
- Point of Saturation
- Singleton Strategy vs. Per Process
- Core Counts vs. Topology
- Download the Code
- References
- KIE Performance Kit
- JMeter
- Vizuri BPM Test Suite code
- About the Authors