Effective Fuzzing with Constraint-Aware Oracle to Detect Logical Bugs in Database Management Systems

MASTER Assignment

Effective Fuzzing with Constraint-Aware Oracle to Detect Logical Bugs in Database Management Systems

Type : Master M-CS

Period: April, 2024 - September, 2024

Student : Parlanti, N. (Niccoló, Student M-CS)

Date Final project: September 30, 2024

Thesis

Supervisors:

dr.ir. A. Continella (Andrea) Associate Professor

+31534892752

 a.continella@utwente.nl

Building: Zilverling 2023

Personal page

dr. P. van den Bos (Petra) Assistant Professor

+31534892677

 p.vandenbos@utwente.nl

Building: Zilverling 3120

Personal page

dr. L. Mariot (Luca) Assistant Professor

+31534894544

 l.mariot@utwente.nl

Building: Zilverling 2035

Personal page

Abstract:

This thesis introduces a novel method for detecting logical bugs in Database Management Systems (DBMS), focusing on constraints such as CHECK and UNIQUE. It combines a custom oracle that simulates correct DBMS behavior with an evolutionary algorithm that generates diverse SQL queries. The approach aims to thoroughly test DBMS constraint enforcement, addressing areas that traditional testing methods may overlook. The system is designed to efficiently generate and validate queries to detect bugs related to constraint violations in DBMS environments.