BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Business rules management (BRM) standards have been around for years. Enterprise architects need to consider whether these contribute meaningful benefit or are merely checkbox features when evaluating BRM products. New standards for decision management like Decision Modeling Notation (DMN) are important for enterprise architects to consider when evaluating BRM products, said James Taylor, CEO of Decision Management Solutions, a decision management consultancy.
He expects the new standard to take off this year and to be published by the BABOK in April. It will provide a formal way for building a decision requirements model for creating and managing business rules and business logic. This makes it easier for organizations to implement executable business models and recruit business rules experts, said Taylor.
The limits of older standards
A number of business rules standards like Drools, JSR-94, RuleML and Rule Interchange Format have been available, but had limited use for writing business rules applications, said Jacob Feldman, CEO of OpenRules, an open source BRM vendor. "Until recently, there were no standards for rules and decision representation, so everyone created their own proprietary environment," he explained. These older standards only specified the interface for business rules engines. Being JSR-94-compliant ended up meaning little since few enterprises actually implemented the standard in real life.
Jacob FeldmanCEO, OpenRules
The rich functionality of BRM systems could only be leveraged by writing rules within each vendor's proprietary environment. As a result, a BRM application written on one platform could not easily be adopted to another. Also, business experts and IT personnel who had mastered one system had to learn a new tool if they switched organizations or the enterprise decided to use another BRM system.
While DMN does not cover the entire BRM lifecycle, it is a first attempt for allowing interchange of rules. However, the industry is still far from being able to fully implement rules in one system which can be reliably executed on another.
"DMN creates something practical about rules representation," said Feldman. However, DMN does not address rule semantics. RuleML is an attempt to represent the semantics of rules, but in and of itself does not provide practical results for decision management, he said. This is more an evolution of the semantic Web movement focusing on rules, and is not widely accepted by BRM vendors.
Lack of standards hindered implementation
The lack of a standard has been an issue for business rules, said Taylor. A lot of companies were reluctant to broadly implement BRM projects. Having some consensus on how to do the analysis and design of decision management architectures will lead to a community of practice that will make a difference in how broadly companies use business rules.
For example, mortgage industry and insurance industry standards groups are interested in defining rules and decisions that could be shared between companies. The standards will allow them to publish a standard way to do something as a model rather than just describe it in text.
Understand the value of DMN
DMN summarizes real-world best practices in a way that complements business process management (BPM). Business rules and decision models can be plugged into BPM systems. "DMN makes the BPM model simpler and more powerful," said Feldman. "In a way, DMN is an extension of BPM notation (BPMN)."
Taylor said the primary use case of DMN will be used inside of companies, rather than shared across them. As enterprises expand their use of business rules they have to recruit people that know their specific BRMS notation. By using standard techniques and a repository for the core of those rules, they can recruit people with a broader skill set.
"It's easy to find more Java programmers, but finding rules experts is more difficult," said Taylor. The standards will make it possible to draw from a larger pool of experts. He expects to mirror the way that SQL made it easier to recruit database talent, rather than just Oracle or IBM experts. The enterprise will still have a need for specific BRM platform expertise, but more of the work could be done by DMN experts.
Taylor expects DMN to be easier for business users to implement as well. The design process in many BRM systems tends to obscure the rules for many business users. DMN will make it easier for non-technical users to see the shape of the problem and make changes more easily. This could lead to more business-friendly management of the rules over time as well.
BRM standards questions remain
Other experts are still skeptical as to whether new standards will be important in the long term for many enterprises. "Every year, we hear about new standards coming and we have to be ready, but none of them seem get a bunch of traction," said Jim Wray, VP of product management and sales operations at InRule, a .Net oriented BRM vendor. He believes these are only likely to provide value to open source BRM systems that are highly technical and geared towards developers.
"But our customers are not looking for standards they are looking for features or capabilities," Wray said. "When you tie to standards, it ties your hands as to capabilities you can deliver to the customer. Outside of open source, it is not even a consideration in the buying process."
Even with standards in place, using rules across industries or public sectors is not straightforward. For example, Wray sees a lot of government agencies attempting to share rules that are a representation of federal law such as calculating Medicaid payments. But each state has local laws that require the rules to be implemented a little differently. As a result, many agencies might take a base rule set as inspiration but find it easier to just implement their own rules from scratch.
"The work towards widely accepted specification will continue," said Edson Tirelli, principal software engineer at Red Hat, an open source BRM vendor, and also the Drools project lead. "It might be DMN or something else, but the industry needs a standard in order to accelerate its growth.
"We are also already seeing a lot of work towards cloud enablement of decision management towards a common interoperable model," he added. "There is still a lot of room for the development of effective knowledge management tooling, as it is an always-evolving concern that is not easy to address."
Learn why intelligent BPM is headed for widespread adoption
Demand grows for big data management tools