Semantic Federation of Product Information from Structured and Unstructured Sources

Introduction

Product-related information is generated, accessed and manipulated along the product lifecycle in heterogeneous formats. Only part of this information can be accessed using state-of-the-art product information systems as large parts of this information are only available in unstructured sources or distributed along different databases and legacy systems. The challenge to create an all-embracing view on products is huge. Such a comprehensive product information system has to integrate and harmonize data from all phases of the product lifecycle, all different source formats like unstructured documents, sensor information or product databases. Furthermore, it must even cross organization boundaries as different stakeholders may be responsible for the design, production, delivery, and service of a product.

The Aletheia project (Aletheia, 2009) is a unique attempt to bring together industry partners (ABB, BMW, Deutsche Post DHL, Otto, SAP) with five innovative application scenarios from different phases of the product lifecycle and five different landscapes of current state-of-the-art product information management. All these partners have a keen interest in improving the information flow internally as well as with their customers and partners and to open up new sources of product-related information like Web 2.0 pages.

In this paper we try to answer the research question if it is possible to federate structured as well as unstructured sources of product information along the product lifecycle. We use semantic technologies for this purpose and deploy and advance information extraction techniques. The scenarios describe two of the use cases of the Aletheia project clarifying the opportunities of federated product information systems (FPIS). We further discuss requirements derived from these and other scenarios. A discussion of existing architectures for semantic information management and federation shows the need for a new architecture matching the requirements mentioned. The contributions of this paper consist of

• 1.

  A discussion of design decisions for FPIS,

• 2.

  A high-level component architecture for FPIS, including a concept for data sharing between organizations,

• 3.

  A detailed concept of the Aletheia Service Hub, our central component for information federation within organizations,

• 4.

  A reference implementation of a semantic FPIS.

We conclude with a discussion of the results achieved so far.

Scenarios And Requirements

In order to motivate our research, we discuss two scenarios in the industrial sector. They are derived from two case studies conducted in the Aletheia project, focusing on

• 1.

  Product lifecycle management (PLM) at ABB, a large company providing power and automation products, technology, and service¹, and

• 2.

  Knowledge management in automotive engineering at the BMW company.