New webMethods Release Draws Business and IT Closer in BPM Work

webMethods unveiled details of its upcoming BPM product suite, webMethods Fabric 7.0, during its customer conference in Washington, D.C. today.

Describing the new set of releases as a "giant leap forward for webMethods in the BPM space," Marc Breissinger, CTO, said the release offers a number of new features, including guided process improvements and enhancements for IT and business collaboration.

Fabric, the name the company uses for its product line, incorporates its core integration technologies -- "historically known as EAI and B2B, as well as SOA-specific technologies and our BPM layer, which includes our business monitoring capabilities," said Breissinger.

Guided process improvements, said Breissinger, is the way the company is describing the tighter integration of its business activity monitoring (BAM) capabilities.

Integration has existed, he said, "at the instance level. You can go and look and see the state of the process instances, data flowing through those processes and deal with exception management at the instance level." But the key performance indicators have been a separate part of the tool. "It was much more of an add-on for activity monitoring." What will be different is the ability for users to look at how well the process is running from a business perspective. For example, new functions let users do event correlation and automatic baselining.

"It learns what normal is at very fine-grained variations for different key performance indicators that you’ve defined for different times of the day," Breissinger said. "So what’s normal at 1 o’clock on Monday for a particular KPI is different from what might be normal for 4 o’clock on a Friday. It will automatically baseline what normal for different points of the day, so it can alert you if something is trending above or below normal."

The functions have been put into a graphical framework that lets users see how well processes are executing and provides "some actionable insight to decide how to fix it," he said. "And if you were to fix this particular portion of that process model, how would that affect your costs?"

Creating better collaboration between IT and the business side of an organization takes the form of "making everybody in the organization a developer," said Breissinger. The approach surfaces in the user interface tool, webMethods Designer. The goal with the new version of this, he explains, is to "enable process designers and implementers to share a common model and allow the implementer to incrementally elaborate on [what the] business designer [defined] to create something that actually executes."

When the implementer (in other words, somebody from the development team) changes something at a semantic level, the software can notify the designer that a core change has happened and that "you might want to get back to the implementer on this." He believes that will cut down on miscommunications between designers and coders.

At the designer level, the new version is "much more about defining the abstract process model -- the requirements, the KPIs." Said Breissinger, users will be able to define costs of different steps and expected throughput at that business level, in order to do business analysis.

"You still see the same pictures at the implementer level," he said. But the implementer would then look in the metadata library (also introduced with version 7.0), to see all of the artifacts or assets -- services or other processes -- that could be reused within the context of the particular process model. If there’s something that can be used, the implementer "can drag the icon of that service from the library right onto the process step to create the executable."

The library allows for search not just on free text but also by concept. As assets are added to the library, the software looks into the services, the service descriptions and into the documents and extracts semantic information as well as relationships between that asset and all the other assets in the library, then makes the metadata searchable. For example, he said, the user can say, "Show me all services related to orders that are used in the context of purchase order processing. It will come back with those types of artifacts for you."

The metadata library will be available in December to collect all artifacts a user is "currently working with." Sometime in January or February, the server-side shared metadata library will be released, said Breissinger, and that version will provide a global view of an organization’s entire library of artifacts.

The new release features a business rules management engine licensed from Fair Isaac. The engine shows up inside the metadata library as another set of services, which can be elaborated as a rule step in the process map. When the user doubleclicks on a rule step, a "codeless" rule editor appears, which presents rules in a decision table or decision tree, exposing the business logic. The user can then edit and deploy from there, independent of the rest of the process model.

Said Breissinger, "If the user needs to tweak rules, it can be done without having to regenerate code or do a regression test, remake or rebuild of the entire process model."

The company said version 7.0 will be released module by module between December 2006 and Summer 2007.

You can view a few screenshots here.