cid 07b1dc334a530d0376a9608709956fdf4d1c80b0ea0edc953be6d931a6ec466a · original · open access · via openalex
Title: The NoSQL Paradox: Why MongoDB is Still Chasing SQL Authors: Daniel Rosehill; Gemini 3.1 (Flash); Chatterbox TTS Venue: Zenodo (CERN European Organization for Nuclear Research) Year: 2026 Canonical: doi:10.5281/zenodo.19034742 URL: https://doi.org/10.5281/zenodo.19034742 Collector: openalex OA: true ABSTRACT Episode summary: Fifteen years after the NoSQL revolution promised to kill the relational database, SQL remains the undisputed industry standard. This episode explores the technical and business reasons why "schema-on-read" flexibility often led to operational debt, and how PostgreSQL eventually neutralized the NoSQL threat by adopting its best features. We also dive into the modern database landscape, discussing the impact of MongoDB's licensing shifts, the rise of open-source alternatives like FerretDB, and why document stores have become a vital "utility player" for developers building AI-driven applications and vector search pipelines. Show Notes In the late 2000s, the tech industry was convinced that the relational database was a relic of the past. The NoSQL revolution promised a future where developers could dump JSON directly into document stores, bypassing the "straitjacket" of rigid schemas and complex JOIN statements. Yet, over a decade later, PostgreSQL usage dwarfs that of MongoDB, and the industry has largely returned to the safety of structured data. ### The Reality of Schema-on-Read The initial appeal of document databases like MongoDB was speed. Developers could ship code without worrying about foreign keys or migrations. However, this "schema-on-read" approach shifted the burden of data integrity from the database engine to the application code. Over time, this often resulted in "operational nightmares" where inconsistent data formats caused application crashes and reporting errors. The industry eventually realized that schemas are not restrictions, but safety harnesses that prevent technical debt from compounding. ### The Convergence of SQL and NoSQL One of the biggest shifts in the database market was the "multi-model" evolution of PostgreSQL. By introducing JSONB support, Postgres allowed users to store and index unstructured data with performance that rivaled native document stores. This neutralized the primary reason many teams left SQL in the first place. Today, the "just use Postgres" philosophy has become dominant because it offers the reliability of a relational engine with the flexibility of a document store when needed. ### Licensing and the Rise of Alternatives The business landscape of document databases changed significantly in 2018 when MongoDB introduced the Server Side Public License (SSPL). This move aimed to prevent cloud giants from profiting off their open-source code but effectively turned MongoDB into a proprietary product. In response, the ecosystem has seen the rise of compatibility layers like FerretDB, which allow developers to use the MongoDB API while storing data in a standard PostgreSQL backend. This "full circle" moment highlights a trend where the document model is treated as a convenient interface rather than a fundamental storage requirement. ### A Second Life in AI and RAG Despite the dominance of SQL, document databases have found a critical niche in the era of Artificial Intelligence. In Retrieval-Augmented Generation (RAG) pipelines, data is often semi-structured—consisting of raw text, metadata, and JSON outputs from Large Language Models. Document stores are uniquely suited for the rapid iteration required in AI development, where prompt changes might result in new data fields every day. While specialized vector databases like Pinecone or Weaviate are preferred for high-scale AI applications, document stores like MongoDB Atlas offer an integrated "good enough" solution for teams looking to avoid the complexity of managing multiple databases. As the industry moves forward, document databases have transitioned from being "the SQL killer" to a versatile utility player in the modern developer's toolkit. Listen online: https://myweirdprompts.com/episode/mongodb-sql-ai-evolution