Business Process Management: Past, Present, Future

Wil M.P. van der Aalst

Information technology has changed business processes within and between enterprises. More and more work processes are being conducted under the supervision of information systems that are driven by process models. Examples are workflow management systems such as Staffware, enterprise resource planning systems such as SAP and Baan, but also include many domain specific systems. It is hard to imagine enterprise information systems that are unaware of the processes taking place. Although the topic of business process management using information technology has been addressed by consultants and software developers in depth, a more fundamental approach has been missing. Only since the nineties, researchers started to work on the foundations of business process management systems. As a result, many questions still wait to be answered. Moreover, new developments like Web Services pose new ones.

Trends

To put workflow management in the proper context, it is interesting to consider ongoing trends. In the sixties information systems were built on top of a small operating system with limited functionality. Since no generic nor domain specific software was available, these systems mainly consisted of tailor-made applications. Since then each year new types of product software are absorbing new functionality. Today's operating systems offer much more functionality than in the sixties. Database management systems offer functionality, which used to be in tailor-made applications. As a result of this trend, the emphasis shifted from programming to assembling of complex software systems. The challenge no longer is the coding of individual modules but orchestrating and gluing together pieces of software from each of the four layers. Another trend is the shift from data to processes. The seventies and eighties were dominated by data-driven approaches. The focus of information technology was on storing and retrieving information and as a result data modelling was the starting point for building an information system. The modelling of business processes was often neglected and processes had to adapt to information technology. Management trends such as business process reengineering illustrate the increased emphasis on processes. As a result, system engineers are resorting to a more process driven approach. The last trend to be mentioned is the shift from carefully planned designs to redesign and organic growth. Due to the omnipresence of the Internet and its standards, information systems change on-the-fly. Few systems are built from scratch. In most cases existing applications are partly used in the new system. As a result, software development is much more dynamic.

Business process management systems are information systems that are “process aware”, i.e., systems that go beyond the automation of individual tasks. Notable examples of business process management systems are workflow management systems such as Staffware, MQSeries, and COSA, and case handling systems such as FLOWer. Note that leading enterprise resource planning systems also offer a workflow management module. The workflow engines of SAP, Baan, PeopleSoft, Oracle, and JD Edwards can be considered as integrated business process management systems. The idea to isolate the management of business processes in a separate component is consistent with the three trends identified. Business process management systems can be used to avoid hard-coding the work processes into tailor-made applications and thus support the shift from programming to assembling. Moreover, process orientation, redesign, and organic growth are supported. For example, today's workflow management systems can be used to integrate existing applications and support process change by merely changing the workflow diagram.

Workflow Management Systems

Despite the efforts of the Workflow Management Coalition (WfMC), workflow management systems use a large variety of languages and concepts based on different paradigms. Most of the products available use a proprietary language rather than a tool-independent language. Some workflow management systems are based on Petri nets but typically add both product specific extensions and restrictions. Other systems use a completely different mechanism. For example, IBM's MQSeries workflow uses both active and passive threads rather than token passing. The differences between the various tools are striking. One of the reasons attributed to the lack of consensus of what constitutes a workflow specification is the variety of ways in which business processes are otherwise described. The absence of a universal organizational “theory”, and standard business process modelling concepts, it is contended, explains and ultimately justifies the major differences in workflow languages - fostering up a “horses for courses” diversity in workflow languages. What is more, the comparison of different workflow products winds up being more of a dissemination of products and less of a critique of workflow language capabilities. In a joint effort of Eindhoven University of Technology and Queensland University of Technology, a new framework for comparing and evaluating workflow systems has been developed. This framework is based on a comprehensive set of patterns and allows for a more objective evaluation of workflow management systems.

We have collected a set of about 30 workflow patterns and have used 20 of these patterns to compare the functionality of 15 workflow management systems (COSA, Visual Workflow, Forte Conductor, Lotus Domino Workflow, Meteor, Mobile, MQSeries/Workflow, Staffware, Verve Workflow, I-Flow, InConcert, Changengine, SAP R/3 Workflow, Eastman, and FLOWer). For more information, we refer to http://www.tm.tue.nl/it/research/patterns.

Web Services Composition Languages

There are two trends coming together in the world of E-business that are creating both opportunities and pressures to automate business processes across organizational boundaries. One is the technology push created by enabling technologies taking XML-based standards and the Internet as a starting point. The other trend is the need to improve the efficiency of processes from a business perspective. After the dotcom crash there is a pressing need to truly utilize the potential of Internet technology by automating business processes across enterprise boundaries. The goal of web services is to exploit XML technology and the Internet to integrate applications than can be published, located, and invoked over the Web. A typical example of a web services application is the Galileo system that connects more that 42,000 travel agency locations to 37 car rental companies, 47,000 hotels, and 350 tour operators. To truly integrate business processes across enterprise boundaries it is not sufficient to merely support simple interaction using standard messages and protocols. Business interactions require long-running interactions that are driven by an explicit process model. This raises the need for web services composition languages such as BPEL4WS, WSFL, XLANG, WSCI, and BPML.

Development with respect to web services composition languages have been mainly driven by software vendors like IBM, Microsoft, Sun, BEA, SAP, and Intalio. This has resulted in an abundance of standards having overlapping functionality. When looking at the standards in more detail, it is clear these are often based on existing products. A good example is WSFL, which is almost a copy of IBM’s Flowmark/MQ Series Workflow language. Standards, which involve multiple software vendors, are often a compromise between competing viewpoints. As a result such standards tend to be imprecise or unnecessarily complex. WfMC’s XPDL is an example of a standard which is imprecise thereby allowing vendors to have their own interpretation of the standard (thus making the standard useless). BPEL4WS joins viewpoints from both WSFL and XLANG thus making the language very complex.

Given these observations it is useful to look for objective measures for comparing web services composition languages. For the control-flow aspect of such languages, one can use some of the results from workflow research. One way to compare standards like BPEL4WS, BPML, WSCI, XLANG and WSFL is to use the set of workflow patterns available from http://www.tm.tue.nl/it/research/patterns. Recently, this site has been extended to include an evaluation of BPEL4WS, BPML, WSCI, XLANG and WSFL based on the 20 basic workflow patterns.

To Conclude

The site of BPMI.org, one of the organizations proposing a web services composition standard, states that “BPMI.org defines open specifications such as the Business Process Modeling Language (BPML) and the Business Process Query Language (BPQL) that will enable the standards-based management of e-Business processes with forthcoming Business Process Management Systems (BPMS), in much the same way SQL enabled the standards-based management of business data with off-the-shelf Database Management Systems (DBMS).” The goal to obtain standards similar to SQL for web services is ambitious. As history shows such standards do not originate from vendors pushing their own products. Recall that the Entity-Relationship model by Chen and the Relational Model by Codd enabled languages like SQL. Although there are well-established process modeling techniques combining expressiveness, simplicity and formal semantics (cf. Petri nets and process algebras), the software industry has chosen to ignore these techniques. As a result, the world is confronted with too many standards which are mainly driven by concrete products and/or commercial interests. The only way to stop this is to ignore standardization proposals that are not using well-established process modeling techniques. This will force vendors to address the real problems rather than creating new ones.

References

1. W.M.P. van der Aalst, K.M. van Hee. Workflow Management: Models, Methods, and Systems. MIT press, Cambridge, MA, 2002.

2. W.M.P. van der Aalst, A.H.M. ter Hofstede, B.Kiepuszewski, and A.P. Barros. Workflow Patterns. QUT Technical report, FIT-TR-2002-02, Queensland University of Technology, Brisbane, 2002. (To appear in Distributed and Parallel Databases, also see http://www.tm.tue.nl/it/research/patterns.)

W.M.P. van der Aalst, M. Dumas, A.H.M. ter Hofstede, and P. Wohed. Pattern-Based Analysis of BPML (and WSCI). QUT Technical report, FIT-TR-2002-05, Queensland University of Technology, Brisbane, 2002.

4. F. Curbera, Y. Goland, J. Klein, F. Leymann, D. Roller, S. Thatte, S. Weerawarana, S. Thatte. Business Process Execution language for Web Services (Version 1.0). IBM, July 2002.

5. F. Leymann. Web Services Flow Language (WSFL 1.0). IBM, May 2001.

6. S. Thatte. XLANG: Web Services for Business Process Design. Microsoft, 2001.

7. Workflow Management Coalition. Workflow Process Definition Interface - XML Process Definition Language (XPDL), WFMC-TC-1025, Version 1.0 Beta, 2002.

8. P. Wohed, W.M.P. van der Aalst, M. Dumas, and A.H.M. ter Hofstede. Pattern-Based Analysis of BPEL4WS. QUT Technical report, FIT-TR-2002-04, Queensland University of Technology, Brisbane, 2002.

-----

Prof.dr.ir.Wil van der Aalst is a full professor of Information Systems and head of the section of Information and Technology of the department of Technology Management at the Technische Universiteit Eindhoven. He is also a part-time full professor at the Computing Science faculty at the department of Mathematics and Computer Science at the same university. He can be reached at P.O.Box 513, NL-5600 MB Eindhoven, The Netherlands. His email is:

Wil van der Aalst is the lead-author of Workflow Management: Models, Methods, and Systems (Cooperative Information Systems) (MIT Press, 2002)

He is a leading researcher in workflow patterns and has applied workflow patterns to BPEL4WS and other XML BP languages. For more information on this, check the Workflow Patterns site: http://www.tm.tue.nl/it/research/patterns.