The Internet-of-Things paradigm has brought exciting opportunities to increase productivity and efficiency but also to enhance customer experience. Such things are expected to connect and exchange information enabling applications to propose a value added services. Nowadays, we witness a paradigm shift from pure data-centric application where the focus was only on the data, towards asset-centric applications to fully manage a product life cycle. Such asset view is a high level representation of the physical thing leading to the introduction of two separate but very highly connected models, the asset and domain data models. In this paper, we propose to rely on semantic technology to abstract and represent physical things and their asset representation in an integrated knowledge approach.