Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Watch this 14-minute conference presentation from OOPSLA 2025 that introduces a groundbreaking methodology for detecting and identifying heisenbugs—elusive software bugs that change behavior when observed—using diversified execution strategies. Learn about the challenges these notorious bugs pose in cyber-physical edge systems, where they evade static detection tools and exploit complex dynamics with unpredictable physical environment interactions. Discover how researchers from Carnegie Mellon University developed Beanstalk, an experimental WebAssembly-backed dynamic data-race detection framework that leverages execution diversity across instrumentation density and hardware platforms to achieve superior bug detection compared to homogeneous strategies. Explore the critical insight that execution diversity benefits outweigh efficiency reductions from limited instrumentation or weaker devices, making the approach scalable for low-overhead deployments across numerous compute nodes. Understand the introduction of the "heisen factor," a novel statistical metric for categorizing heisenbugs and measuring detector effectiveness, along with practical insights for debugging strategies in both development and deployment environments. Gain knowledge about how this methodology addresses the fundamental challenge of determining a bug's "difficulty" or "heisen-ness" while maintaining performance requirements for real-world adoption.
