The increasing complexity and volume of simulation and experimental data in engineering and scientific domains demand robust strategies for data structuring, publication, and long-term accessibility. This contribution proposes a framework for encapsulating structured scientific data in FAIR (Findable, Accessible, Interoperable, Reusable) digital objects, registered in publicly accessible databases, and hosted in institutional or domain-specific repositories with persistent identifiers and long-term availability. At the data level, the approach leverages JSON Schema to formally describe the structure of scientific data in a standardized, human-readable and machine-actionable format. Where applicable, these schemas can be extended to JSON-LD using domain ontologies, enabling semantic interoperability. At the metadata level, a knowledge graph is constructed using established standards such schema.org , enriched with schema-specific descriptors to support fine-grained discovery and filtering of data objects. The paper discusses the conceptual and technical foundations of this approach and compares it to existing curated global databases, highlighting trade-offs in scalability, maintenance, and openness. Potential applications include the publication of data from round-robin tests, interlaboratory studies, and individual experiments, with the goal of fostering reproducibility, transparency, and community-driven validation. By initiating a discussion on decentralized yet structured scientific data publication, this contribution aims to explore future pathways for sustainable data sharing and reuse in simulation and test-driven research.