Pervasive computing at scale: Challenges and research directions

There is a strong push of public authorities and industries toward the realization of “Smart Cities” (e.g., see the EU Future Internet initiative) and of “Planet-wide Sensing Environments” (e.g., see the EU FuturICT flagship proposal), where a huge number of (mobile) heterogeneous devices should interwork to enable several types of dynamically adaptive and pervasive applications. It starts to be manifest that smart applications for urban-wide (and larger) deployment scenarios require facing the hard technical challenge of scalability. This position paper capitalizes the experiences made in our recent research activities about innovative models, algorithms, and middleware prototypes for scalability in smart cities. Based on that, it tries to identify major technical challenges still open in the field of scalability for pervasive computing and major related solution trends; we think that these solution guidelines should be addressed with relevant investments and research efforts in the years to come. In particular, we claim that a very relevant guideline is to reduce the complexity of the potential space of collaboration opportunities via middleware solutions that proactively determine active context sub-spaces. Preliminary encouraging results in the fields of urban monitoring via vehicular sensor networks (MobEyes), of opportunistic Internet connectivity sharing (RAMP), and within the EU SOFIA project for interoperable smart environments will be rapidly overviewed and discussed in this position paper and in the associated invited speech; the interested reader is referred to other papers in the literature for more in-depth technical descriptions of the already investigated solutions.