Red Hat Inc. Friday updated its Red Hat Process Automation Manager with machine learning-based predictive modeling capabilities and a variety of user interface enhancements.

The software suite, which was previously called JBoss Business Process Management Suite, combines BPM with business rules management business resource optimization and complex event processing into a platform for developing applications based on software containers and microservices.

They represent an approach to building applications that combine loosely coupled components in a way to improve resiliency and speed development. The cloud-based suite generates cloud-native applications in container images that can be deployed on Red Hat’s OpenShift container orchestration.

In the new version, the microservices architecture has been expanded to user interface development in an approach the company calls “micro-frontend architectures.” That means developers can “construct the user experience in the same way we assemble and application with microservices,” said Phil Simpson, Red Hat’s JBoss product marketing manager. “It’s a way of composing a user interface from individual pieces that don’t necessarily know about each other.”

Although Red Hat is best known for its infrastructure software, it also has a large application development business, into which Process Automation Manager fits. “It’s a relatively small component of our overall business but one of our faster growing areas,” Simpson said. That latest release is intended to make the suite more accessible to business analysts and nonprofessional developers.

Process Automation Manager is based primarily on Drools, an open source project developed by Red Hat that combines a business rules engine, web authoring tools and a rules management application. With the addition of machine learning capabilities, rules-based applications can be loaded with training data that enables them to make better decisions based upon past actions.

For example, in a credit-approval scenario, “you can funnel in years of credit decisions and create a predictive model that determines how those decisions should be made in the future,” Simpson said. “Basic rules can be dramatically improved and made available to a wider range of applications” using microservices.

The new release also enables users to import and execute predictive models expressed in Predictive Model Markup Language, an industry standard for integrating and exchanging information between machine learning platforms. Its also compliant with Business Process Model and Notation 2.0 and Decision Model and Notation 1.2, which are open standards for describing and documenting business models that are useful in model analysis and regulatory reporting.

Other enhancements in this release include automated lifecycle management support via OpenShift Operators, better process visibility using visualizations, support for continuous operation through multi-node deployment, which protects against node failure and customizable templates for business resource optimization. The software is available now.

Photo: Red Hat