Program Analysis for Reliable Sensor Network Software

Author(s): Roy Shea; Shane Markstrum; Mani Srivastava; Todd Millstein; Rupak Majumdar | Abstract: This work examines proactive techniques to help developers overcome the difficulty of writing software for the sensor network domain. A framework for static program analysis is used to implement a collection of small checkers to verify that programs follow established safety conventions and conform to system resource models. Initial analysis of the SOS code base has revealed interesting errors and helped to pinpoint regions of overly complex program constructs. From this foundation we hope to develop a suite of staged checks to support program development from compile time through execution time.