A reconfigurable component model with semantic type system for dynamic WSN applications

Klaas Thoelen, Danny Hughes*, Nelson Matthys, Lei Fang, Simon Dobson, Yizhou Qiang, Wei Bai, Ka Lok Man, Sheng Uei Guan, Davy Preuveneers, Sam Michiels, Christophe Huygens, Wouter Joosen

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)


Runtime reconfigurable component models provide several attractions with regard to the management of wireless sensor network (WSN) applications operating in dynamic environments and under evolving application requirements. One such attraction is the runtime discovery of suitable components for reuse in changing application compositions. Syntactic interface typing, provided by contemporary component models, however only supports exactinterface matching. This causeslimited reuse of components and complicates management of WSN applications. We argue that more flexibility is required to efficiently manage the complex, large-scale and dynamic WSN deployments of the future. In this paper, we describe the addition of semantic service descriptions to component interfaces to support compatibility and subtype testing. This allows rich discovery and reuse of third-party functionality and reasoning at the level of equivalent service types. We report on the incorporation of these semantic interface definitions in the Loosely Coupled Component Infrastructure (LooCI). Evaluation thereof shows that the scheme imposes minimal computational and memory overhead, while significantly reducing the complexity and cost of reconfiguration.

Original languageEnglish
Pages (from-to)277-290
Number of pages14
JournalJournal of Internet Services and Applications
Issue number3
Publication statusPublished - Dec 2012


  • Component models
  • Reconfiguration
  • Semantic type system
  • Wireless sensor networks


Dive into the research topics of 'A reconfigurable component model with semantic type system for dynamic WSN applications'. Together they form a unique fingerprint.

Cite this