Analisis Perbandingan Metode White Box dan Black Box Testing pada Pengujian Modul Autentikasi Sistem Web

Abstract

The security of authentication modules in web systems is a critical issue directly impacting user data integrity and system resilience against cyber threats. Inadequate testing can lead to serious security vulnerabilities including SQL injection, authentication bypass, and token exploitation. This study analyzes, compares, and integrates the effectiveness of white box testing and black box testing in testing a PHP-based web system authentication module using JSON Web Token (JWT). A quantitative experimental approach was adopted using seeded defect technique, deliberately injecting 32 defects to objectively measure detection effectiveness of each method. White box testing was implemented using McCabe's basis path testing with PHPUnit 10 and Xdebug-based code coverage analysis, while black box testing used equivalence partitioning and boundary value analysis through Postman API Client. White box testing detected 28 of 32 defects (87.5%), excelling in identifying logic errors and JWT cryptographic algorithm weaknesses. Black box testing detected 29 defects (91.3%), with significant advantages in uncovering input validation errors and HTTP parameter-exploitable security vulnerabilities. Combined, both methods synergistically achieved 93.75% defect coverage (30 of 32 defects). The study concludes that integrating white box testing at unit testing stage and black box testing at integration testing stage provides the most comprehensive quality assurance for web system authentication modules, supported by a strategic matrix for optimal testing method selection.