Self-Organizing Services for Browsing the World: Challenges and Directions

Summary form only given. The imminent mass deployment of pervasive computing technologies such as sensor networks and RFID tags, together with the increasing participation of the Web community in feeding geo-located information within tools such as Google Earth, will soon make available an incredible amount of information about the physical and social worlds and their processes. Overall, both the above trends contribute to the forming of a decentralized infrastructure that, by accumulating and making available in a ubiquitous way detailed digital information about the surrounding world, opens up the possibility of conceiving innovative context-aware services for "browsing the world" around us, as well as of enhancing the quality and effectiveness of current ICT services via context-awareness and dynamic personalization. However, this will also introduce a dramatic complexity increase in data and service management, making impossible for humans to stay in the control loop and directly manage the configuration and functioning of these systems, and compulsory calling for unsupervised self-organizing approaches at both the data and the service level. On the one hand, proper self-organizing approaches must be defined for processing and aggregating large amounts of data into a meaningful and manageable "world model" to be made available to services. On the other hand, proper distributed computing models and infrastructures must be defined to enable services to exploit such information to self-organize their distributed and context-aware activities at the best. In this talk, after having introduced my general vision of future pervasive computing scenarios, I try to identify the key research challenges that should be faced towards the effective provisioning of self-organizing and context-aware services in such scenarios, and will sketch some possibly promising research directions currently being investigated within the "Agents and Pervasive Computing Group" of the Universit- a di Modena e Reggio Emilia