Towards a Business Network Management

Abstract

Abstract. Enterprises are part of value chains consisting of crossenterprise business processes forming large business networks of customers, vendors, partners and competitors. These business process networks run on (technical) integration networks, which are semantically interlinked with the processes, while both are correlated to organizational (social) networks, consisting of technical and business domain experts. The insight into these networks promise a competitive advantage through the visibility into the linked enterprise data. However, this information is hidden in complex, dynamic and heterogeneous enterprise domains.In this position paper, we introduce the new domain of Business Network Management (BNM), which strives to make business networks visible within network views and sets them into context with each other. For that, we extend the Network Mining (NM) domain by a virtualized network approach towards a BNM. That means automatic discovery, mining and inference capabilities are combined with expert knowledge from different domains to compute the networks and a semantic correlation between entities of different perspectives. We comprehensively define the domain and discuss a system implementation, which we evaluated in a real-world customer case.

Advanced Data Processing in the Business Network System

“…3. REQ-3 The user shall be able to enrich (e.g., labeling, grouping) and to enhance the BN data (e.g., adding/ removing nodes and edges) through the client API, while delta-changes from existing data sources and initial loads from new data sources are merged into the existing BN (from [15]). 4.…”

“…The center of Figure 1 shows the core elements of a NM system, theoretically discussed in [13] and [15], which computes the perspectives of the BN for client access. After the loaded data has been checked for conformance to the inference model it is stored as raw data for the continuously running network reconstruction programs using logic programming (i.e., our approach uses Datalog due to the rationale in [16]).…”

“… Enhancements from the user are valued higher than the information from the automatic discovery and replace the discovered information within the surrogate (refer to [15] for the theory of coming from the computed "as-is" to a "to-be" network through expert knowledge, i.e., enhancements)  The enhancements can be re-deployed to the sources if sufficient lineage tracing exists (i.e., for the creation of new instances, e.g., a participant, the source has to be specified and enough domain specific information has to be provided to create and activate the instance in the source system. These sources can be middleware system, for which the new instance should be runnable as any other created locally.…”

“…Section II guides from the theoretical work conducted in the area of NM [13], Business Network Management (BNM) [15], and its reference architecture [16] to the real-world query and data processing and provenance requirements and capabilities of a BNS (refers to (1)). Section III provides an overview of BNS's data transformations, provenance and data processing, including query and update processing (refers to (2)) and sketches a high-level view on the system's architecture (refers to (3)).…”

A Network-Centric BPMN Model for Business Network Management

“…Even though the BN is very important for enterprises, there are few -if any -people in the organisation who understand this network as the relevant data is hidden in heterogeneous enterprise system landscapes. To change that, Business Network Management (BNM) [9] allows enterprises to get insight into their technical, social and business relations. It identifies relevant data hidden within heterogeneous and distributed systems in complex enterprise landscapes to computationally link it into business process and (technical) integration networks [10].…”