A compilation framework for macroprogramming networked sensors

  • Authors:

  • Animesh Pathak;Luca Mottola;Amol Bakshi;Viktor K. Prasanna;Gian Pietro Picco

  • Affiliations:

  • Ming Hsieh Department of EE-Systems, University of Southern California;Dipartimento di Elettronica ed Informazione, Politecnico di Milano, Italy;Ming Hsieh Department of EE-Systems, University of Southern California;Ming Hsieh Department of EE-Systems, University of Southern California;Department of Information and Communication Technology, University of Trento, Italy

  • Venue:
  • DCOSS'07 Proceedings of the 3rd IEEE international conference on Distributed computing in sensor systems
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Macroprogramming-the technique of specifying the behavior of the system, as opposed to the constituent nodes-provides application developers with high level abstractions that alleviate the programming burden in developing wireless sensor network (WSN) applications. However, as the semantic gap between macroprogramming abstractions and node-level code is considerably wider than in traditional programming, converting the high level specification to running code is a daunting process, and a major hurdle to the acceptance of macroprogramming. In this paper, we propose a general compilation framework for a data-driven macroprogramming language that allows for plugging in different modules implementing various stages of compilation.We also demonstrate an actual instantiation of our framework by showing an end-to-end solution for compiling macroprograms. Our compiler provides the final code to be deployed on real nodes as well as an estimate of the costs the running system will incur, e.g., in terms of messages exchanged. We compared the auto-generated code against a handcoded version for the same application behavior to verify the outcome of our compiler.