Profile-based data diffusion in mobile environments

Abstract

People on the move often are interested in information about their close proximity. This information can either be retrieved from static nodes or other mobile nodes spreading information via wireless technologies. This approach to data diffusion potentially overloads the mobile device if no filtering is applied. Filtering mechanisms, however, cannot rely on the existence of a common information model among the peers. We present an approach for profile-based filtering which does not require shared knowledge among the nodes. Instead, our system can translate between competing taxonomies. It uses synonyms and a tolerant search algorithm to cope with morphological and lexical variations and structural differences of the taxonomies. A prototype of our approach was implemented and successfully tested.