Lead Investigators

Mani Srivastava (UCLA), Mark Hansen(UCLA), Jeff Burke (UCLA), Deborah Estrin (UCLA)
Mark Allman (ICIR/ICSI), Vern Paxson (ICIR/ICSI)

Overview

Application context inevitably drives the architecture design choices and the definition of services needed in a network. The creation over the past decade of unanticipated applications of the Internet, such as web services, peer-to-peer file sharing, networked gaming, podcasting, and voice telephony, is resulting in a recent rethinking of the core Internet infrastructure and the original architecture choices. In this project however we propose to go beyond reacting to these applications that have already emerged, and proactively consider the network architecture implications of a new class of applications involving embedded sensing technology as it moves from scientific, engineering, defense, and industrial contexts to the wider personal, social and urban contexts. Today, applications are emerging which draw on sensed information about people, objects, and physical spaces. These applications enable new kinds of social exchange: By collecting, processing, sharing, and visualizing this information, they can offer us new and unexpected views of our communities. They require new algorithms and software mechanisms because unlike scientific applications of distributed sensing, a single system is widely distributed, intermittently connected, and privately administered; and unlike traditional Internet applications the physical inputs are critical to the behavior.

Intellectual Merit: Our research seeks to develop the principles and abstractions that are vital for FIND to incorporate such applications, to identify the network fabric architecture options, and to develop the key components and services needed to realize them. We will distill the concrete new requirements imposed by personal, social and urban sensing applications on the network core, and explore and evaluate solutions to what we have identified as the two critical core network services: (1) context verification and resolution control: To provide increased utility and assurance to the subscriber while preserving its own privacy, a publisher might want the network to add to the published data stream attested contextual information of specified resolution, such as location, time, and comparisons with measurements at nearby sensors. Likewise, a subscriber may seek from the network an audit trail of contextual information to establish data validity and authenticity; (2) application services for naming, dissemination, and aggregation: The combination of physical and personal coupling enabled by personal wireless sensing devices introduces new naming, dissemination, and aggregation requirements and constraints; in particular, personal physical sensing demands fine grain articulation of selective sharing. Moreover, urban and social sensing applications will flourish, provided a flexible sensor information fabric exists with support for naming, dissemination, and a wide range of data remixing, correlation and fusion functions. We will conduct this research by constructing, implementing and evaluating solutions based on peer to peer overlay techniques, as well as more radical native network services support. Experimental validation of the ideas will be done via concrete driver systems with three complementary characteristics: one involving sensors carried on body to sense physical state and context of mobile individuals for personal health applications, the second involving mobile sensors contributing to a soundscape characterization in support of an urban application, and the a social data sharing application based on location-tagged multimedia data. Consequently a second area of intellectual contributions will be development of concrete instances of personal, social and urban embedded networked sensing applications to act as design drivers for the broader community of researchers architecting the new Internet core. The primary intellectual merits of this research will thus be to yield a fundamental understanding of software architectures, networking models and data processing techniques that are needed to support a citizenry actively participating in collection, sharing, and interpretation of physical sensor data in the public sphere at multiple scales: an individual, a group, or a place.

Broader Impact: The technology of personal, social and urban sensing will cut across many segments of society, from child and elder care, to entertainment and social discourse, and lead to new applications. This research involves faculty and students from engineering to statistics, and to the arts. This breadth of disciplinary participation brings with it an opportunity for far greater diversity than is usually possible in technology-only projects. Through our undergraduate research internship activities and development of an interdepartmental program, we will not leave this broader impact and opportunity for diversity to chance.

This material is based upon work supported by the NSF under award CNS-0627084. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF.