A research paper by Professor Wu Rongxin's team, titled "ValScope: Value-Semantics-Aware Metamorphic Testing for Detecting Logical Bugs in DBMSs", has been accepted to the 20th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2026), the first paper accepted by the OSDI with Xiamen University (XMU) as the first-affiliated institution.
OSDI is one of the flagship conferences in computer systems research, covering operating systems, database systems, distributed systems, and system software. It is classified as a Class A conference by the China Computer Federation.
The paper introduces ValScope, a testing approach designed to detect logical bugs in database management systems (DBMSs), which underpin applications in finance, e-commerce, and cloud services. Unlike system crashes, logical bugs do not interrupt operations but can silently produce incorrect query results, posing risks to data reliability.
To address limitations in existing testing methods, which primarily rely on result equivalence or set inclusion, the research proposes a unified semantic modeling approach that integrates set semantics with value semantics. This enables the detection of errors where result sets appear unchanged but underlying computations are incorrect.
Based on this framework, the team developed an automated tool, VALSCOPE, which generates, mutates, and validates SQL queries to identify inconsistencies. Experiments across six widely used DBMSs — MySQL, MariaDB, OceanBase, Percona, PolarDB, and TiDB — uncovered 67 previously unknown logical bugs. Of 57 confirmed bugs, 48 dated back to before 2020, with some persisting for up to 20 years.
