SSO dan Single Logout Berbasis JWT untuk Sistem Informasi Universitas Nurul Jadid

Abstract

Digital transformation in Indonesian universities requires reliable identity and access management as more campus systems are introduced. At Universitas Nurul Jadid (UNUJA), this study identifies three main problems: (1) low user efficiency because users must log in repeatedly (average 6.8 minutes to access four systems with a 14% login error rate), (2) higher security risk due to scattered credentials and uncoordinated logout across systems, and (3) heavy administrative workload for creating and maintaining accounts in multiple applications. This research aims to design, build, and evaluate an integrated Single Sign-On (SSO) and Single Logout (SLO) solution that meets both functional and non-functional requirements in a pesantren-based university environment. A Waterfall method is applied in five stages: analysis (literature review and field study with 35 participants), design (modular architecture and UML), implementation (Laravel 12, PHP 8.3, MySQL 8.4, Redis 7.0, JWT-based authentication), testing (functional, security, performance, and extreme-scenario tests), and maintenance (monitoring with Laravel Telescope and Redis Monitor). The proposed system uses a centralized Identity Provider (IdP) that supports OpenID Connect for modern applications and SAML 2.0 for legacy systems. SLO is implemented through both front-channel and back-channel methods, supported by Redis-based token blacklisting. Test results show the system meets all functional requirements, withstands the tested security scenarios, and remains stable under load (average latency 1.2–2.1 s; P95 1.8–3.2 s; success rate 98.7–99.9%; throughput up to 850 req/s). This study contributes: (1) a “session cycle” framework that combines SSO and SLO for Indonesian higher education, (2) a mixed evaluation rubric covering technical, security, usability, and administrative metrics, and (3) practical guidance on protocol choices, legacy integration, MFA options, and SLO patterns with recovery mechanisms. The results can serve as a blueprint for other institutions facing multi-system authentication challenges.