The Changing Role of Pervasive Middleware: from Discovery and Orchestration to Recommendation and Planning

Future pervasive computing scenarios will be characterized by an increasing diversity and dynamics of services and of contextual data sources, and by an increasing exploitation of crowdsourcing for social sensing and human computation. Accordingly, the role of middleware should no longer be limited to facilitating interactions and compositions via discovery and orchestration, but should approach that of a recommendation engine capable of dynamically and adaptively planning patterns of service interaction and composition on a best-effort basis. Along these lines, this position paper elaborates on the limitations of traditional middleware infrastructures in meeting the new requirements of the emerging pervasive computing scenarios. Then, it introduces two case study scenarios to motivate and clarify the concepts expressed. Finally, it identifies some key research challenges for future pervasive middleware infrastructures.