Interaction-Driven Definition of e-Business Processes

Business-to-business interaction (B2Bi) is the next step for corporate IT [1]. Business relationships become increasingly dynamic, and new requirements emerge for data and process management. Standardisation initiatives are successfully targeting business ontology [4]. Still, business agility mainly depends on the flexibility of the business processes of a company. In the B2B space, traditional approaches to process modelling and management are inadequate. Today more than ever, traditional workflow management is crucial for the internal effectiveness of a company. Internal efficiency is a prerequisite for external agility. From both a technical and a business perspective, internal workflow management relies on specific assumptions in terms of resources involved in the process, as well as the process itself [2]. Level of control, availability, reliability, and cost stability are parameters that traditional process models and technology can almost take for granted. A single authority ruling on the process definition and the total control over process execution are also basic concepts for internal workflows. From a business perspective, a big upfront investment is put in the complete definition of process specifications. A different conceptual framework is required for the definition and management of e-business processes [3, 5]. The intrinsic capability to adapt to rapidly changing business requirements becomes crucial. The line of research explored in this paper derives from an approach to process modelling and management that explicitly targets the peculiarities and dynamics of B2Bi. In the model we propose, the upfront specification of the interaction logic of a company can be limited to partially specified processes and basic interaction rules. Specific information is then gathered from the observation of actual instances of business interaction, and used to refine and extend the initial model. In addition to the enforcement of explicit business requirement, the goal is to capture and leverage implicit operational knowledge. In the following sections, we present an overview of the methodology we are currently experimenting with for the inference of complex processes from business interaction flows. For our initial experiments, we focus on business messages compliant with the RosettaNet standard [4].