IBM T.J. Watson Research Center, Yorktown Heights, NY, USA
Abstract. A data-centric approach to business process and workﬂow modeling has been emerging over the past several years. This short paper presents a structured framework for a class of data-centric business process models, which are based on “business artifacts”. The paper provides a brief survey of research results on artifact-centric business process, and identiﬁes a broad array of remaining research challenges.
Businesses and other organizations increasingly rely on business process management, and in particular the management of electronic workﬂows underlying business processes. While most workﬂow is still organized around relatively ﬂat process-centric models, over the past several years a data-centric approach to workﬂow has emerged. A key paper in this area is , which introduces the artifact-centric approach to workﬂow modeling. This approach focuses on augmented data records, known as “business artifacts” or simply “artifacts”, that correspond to key business-relevant objects, their lifecycles, and how/when services (a.k.a. tasks) are invoked on them. This approach provides a simple and robust structure for workﬂow, and has been demonstrated in practice to permit efﬁciencies in business transformation. As this approach has been applied, both internal to IBM and with IBM customers [5,6] a family of new requirements has emerged which are not easily addressed by the largely procedural artifact-centric model introduced in . To address these requirements, variations of the original artifactcentric model are now being explored. This short paper presents a framework that can be used to help structure this exploration and more clearly expose the implications of different modeling choices in artifact-centric business process. The paper also highlights a broad array of research challenges raised by the artifact-centric approach. The ﬁeld of artifact-centric business process is still in its infancy; the research results and challenges described here are not intended to be comprehensive, but rather to reﬂect the author’s view on some of the most pressing and useful issues to study. The basic challenge in business process modeling is to ﬁnd mechanisms whereby business executives, analysts, and subject matter experts can specify, in an intuitive yet concise way, the framework and speciﬁcs of how the operations of a business are to be conducted. The speciﬁcation should make it easy to develop IT infrastructures to automate the operations as much as possible. At the same time the framework must support Supported in part by NSF grants IIS-0415195 and CNS-0613998. R. Meersman and Z. Tari (Eds.): OTM 2008, Part II, LNCS 5332, pp. 1152–1163, 2008. c Springer-Verlag Berlin Heidelberg 2008
Artifact-Centric Business Process Models
ﬂexibility at two important levels. The ﬁrst is at the level of individual enactments of the workﬂow – with the increasing intricacy of modern day corporate and personal life, and the emergence of the “mass market of one” , it is essential that workﬂows permit a dramatic improvement in permitting highly varied operation for different customers, products, contexts and regions. The second is ﬂexibility in enabling rich evolution of the workﬂow schema. This is increasingly important as “internet speed” becomes a reality in the deﬁnition and transformation of new narkets, new competitors, and new government regulations. While both forms of ﬂexibility must be supported, monitoring and reporting continue to be critical, and mechanisms need to be found so that the reports are meaningful in spite of highly varied enactments and continually evolving workﬂow schemas. The emerging challenge of generic/specialized is focused on the need to permit the speciﬁcation of a generic workﬂow schema that can have multiple...