Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Watch a 14-minute conference presentation from OOPSLA 2025 that introduces a novel client-centric technique for verifying the correctness of database systems through observational analysis. Learn about a groundbreaking approach that checks both semantic correctness and isolation-level guarantees for black-box database systems by analyzing observations collected from transaction executions. Discover how the research team developed symbolic encodings for semantic correctness of database transactions under weak isolation and isolation-level guarantees, which are processed by a Satisfiability Modulo Theories solver to verify observational correctness. Explore the practical application of their tool Troubladour, which successfully verified observational correctness in several database systems including PostgreSQL and an industrial system under development, leading to the detection of two new bugs. Understand how this methodology can identify known semantic correctness bugs and detect isolation-related anomalies, providing database developers and researchers with a powerful verification framework. The presentation covers the theoretical foundations of observational correctness, the technical implementation details of the symbolic encoding approach, and real-world case studies demonstrating the tool's effectiveness in finding critical database system vulnerabilities.
