Service Registration, Indexing, Discovery & Selection; An Architectural Survey Toward a GenAI-Driven Future

The emergence of sixth-generation (6G) networks marks a paradigm shift: by unifying an edge-to-cloud computing continuum with ultra-high-performance networking, 6G will enable capabilities far beyond today's boundaries. As use-case diversity grows exponentially and user adoption drives traffic to unprecedented and highly dynamic levels, novel service orchestration mechanisms are indispensable. In this paper, we adopt an architectural viewpoint, examining Service Registration, Indexing, Discovery, and Selection (SRIDS) as fundamental elements of 6G service provision. We first establish the theoretical foundations of SRIDS in 6G by defining its core concepts, detailing its end-to-end workflow, reviewing current standardization efforts, and projecting its future design objectives, including reliability, scalability, automaticity and adaptability, determinism, efficiency, sustainability, semantic-awareness, security, privacy, and trust. We then perform a comprehensive literature review and gap analysis encompassing both existing surveys and recent research efforts, identifying conceptual and methodological gaps that hinder unified SRIDS in 6G. Next, we introduce a taxonomy that classifies SRIDS mechanisms into centralized, distributed, decentralized, and hybrid architectures, and systematically examine the relevant studies within each category. Each work is evaluated against the extracted design objectives. Building on these findings, we propose a hybrid architectural framework, combining centralized data management to ensure consistency and agility with distributed coordination to enhance scalability in emerging 6G use cases. The framework incorporates innovative technologies, such as Generative Artificial Intelligence (GenAI). We conclude by highlighting open challenges and suggesting directions for future research.