A service-oriented architecture for dynamic macroprogramming of sensor networks

In the late 1990s, advances in sensing and computer technology have enabled the development of tiny, inexpensive, low-power wireless sensor platforms. By integrating sensing, communication, and computational capabilities, these smart sensors were poised to revolutionize our view of the environment we inhabit by linking the physical world with the digital realm of traditional computing. Smart sensors have been available to researchers for more than a decade; however, few large-scale applications have emerged outside the laboratory setting, and the commercial potential of this technology has been limited. The principal reason for this outcome is the difficulty inherent in programming wireless sensor networks (WSNs) consisting of more than a handful of sensors: built from inexpensive components, individual nodes in this distributed system are prone to failures; interaction with the physical world imposes real-time constraints on computation and communication; and the limited energy of battery-powered sensor nodes leads to stringent energy efficiency requirements. Combined, these challenges have caused WSN software development to lag behind the capabilities offered by the hardware. The goal of this research is to enable robust, large-scale application development for wireless sensor networks, allowing the full potential of WSN technology to be realized. To this end, we leverage two powerful techniques, service-oriented architecture (SOA) and macroprogramming. Adapting SOA, which is typically seen in Internet-scale web applications, to WSNs enables application components to cooperate and share limited resources in an intelligent manner, while providing useful high-level programming abstractions to the application developer. Macroprogramming -- specifying the aggregate behavior of a distributed system rather than each node individually -- builds on SOA to create lightweight, mobile applications that can combine and control the services resident in the network to take advantage the capabilities of the network as a whole. This approach has proven successful, enabling a long-term deployment of a dense array of structural health monitoring (SHM) sensors on a cable-stayed bridge in Jindo, South Korea. The software resulting from this work, which integrates the service-oriented application development framework with a suite of domain services and comprehensive applications for SHM, has been released as the open-source Illinois SHM Services Toolsuite. It is currently in use by over 70 research groups worldwide.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2012

Author dc:creator

• Mechitov, Kirill A.

Contributors dc:contributor

• Agha, Gul A.
• Spencer, Billie F., Jr.
• Abdelzaher, Tarek F.
• Gupta, Indranil

• wireless sensor networks
• service-oriented architecture
• macroprogramming
• structural health monitoring

Statement dc:rights

• Copyright 2011 Kirill Alexandrovich Mechitov

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/29664

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/29664