Database Reference
In-Depth Information
3.6 Other Architectural Challenges
Proper design of the above system components gives rise to other
important challenges that must be solved in order to enable development
and deployment of successful mobile sensing applications that adequately
meet user needs. The following relates challenges described in a recent
NSF-sponsored report
1
on social sensing.
From the application perspective, mobile sensing applications depend
significantly on social factors (user adoption, peer pressure, social norms,
social networks, etc) as well as the nature of physical phenomena being
monitored or controlled. Exciting interdisciplinary research challenges
exist in describing the properties of distributed socio-physical applica-
tions. For example, what are the dynamics of information propaga-
tion in such systems? What are closed-loop properties of interaction
involving social and physical phenomena? What are some fundamental
bounds on capacity, delivery speed, and evolution of socio-sensing sys-
tems? Answering such questions is fundamental to informed design and
performance analysis of sensing applications involving crowd-sourcing.
From the underlying physical network perspective, mobile sensing ap-
plications herald an era where many network clients are embedded de-
vices. This motivates the investigation of a network architecture, where
the main goal from networking shifts from offering a mere communica-
tion medium to offering
information distillation services
. These services
bridge the gap between myriads of heterogeneous data feeds and the
high-level human decision needs. In a network posed as an information
service (as opposed to a communication medium), challenges include
division of responsibilities between the end-device (e.g., phone) and net-
work; paradigms for data collection on mobile devices, architectural sup-
port for data management, search, and mining; scalability to large-scale
real-time information input and retrieval; improved context-awareness;
support for predictability; and investigation of network and end-system
support for reduction of cognitive overload of the information consumer.
Other challenges in the design of network protocols for mobile sensing in-
clude energy management, integration of network storage, personalized
search and retrieval, support for collaborative sensing, and exploitation
of a rich realm of options in information transfer modalities and timing,
including deferred information sharing and delay-tolerant communica-
tion.
1
National Science Foundation Workshop Report on Future Directions in Networked Sens-
ing Systems: Fundamentals and Applications, The Westin Arlington Gateway, Arlington, VA,
November 12-13, 2009.
