The Rise of EdgeDB: A New Era in Database Technology

EdgeDB is a graph-relational database that has been under development for over a decade. The journey began in 2008 when Yury and Elvis, the co-founders of MagicStack, started incubating homegrown tools to address the problems they encountered on every project. One of these tools was Caos, a "super ORM" that evolved with each client project and featured a YAML-like, object-oriented schema definition language, support for schema composition, indexes, constraints, and a query language called CaosQL. Caos already had some of EdgeDB's big ideas, such as links, schema mixins, easy deep querying, and introspectability.

In 2014, the co-founders launched BoatBee, an endeavor that proved to be a post-mortem learning experience. It prompted them to look critically at their stack of tools and recognize that while the state of tooling for web development had progressed, the database layer had not. They saw an opportunity to evolve Caos into a new type of database, yet numerous obstacles needed to be overcome.

The journey to the first stable version of EdgeDB took longer than anticipated, but along the way, the co-founders spun out two major open-source projects, helped introduce async/await into the Python language, and grew to a 10-person open-source company. On February 10th, 2023, EdgeDB released its first stable version after 2100 PRs, 4600 integration tests, and 14 pre-releases. Just six months later, EdgeDB 2.0 was released, adding many new features, including a built-in admin UI and object-level security. On November 7, 2022, EdgeDB announced it had raised a $15 million Series A financing round led by Nava Ventures and Accel.

Can we do better than SQL?

The relational model was introduced by Edgar F. Codd in a 1970 paper where he proposed that data in a database could be represented as sets of tuples called relations. He also created a form of first-order predicate logic for describing database queries: tuple relation calculus. In 1974, Donald Chamberlin and Raymond Boyce published a paper introducing SQL as a simplified alternative to the formal relational query languages.

Many experts, including Codd himself, heavily criticized SQL. It lacks compactness, consistency, and cohesion, making it hard to compose, write, and comprehend queries. SQL is not a small language and contains 469 keywords, and just part 2 (out of 14) of the SQL:2016 standard has 1732 pages. This complexity contributes negatively to the ability to write and comprehend SQL queries.

EdgeDB was built from the ground up to solve the foundational problems of SQL and is guided by principles of ergonomics, performance, and correctness. The data model, EdgeQL, and all aspects of EdgeDB are designed to be straightforward to reason about and satisfying to use. EdgeQL features, language bindings, and tooling are also designed for high performance and low latency. While usability and performance are important, correctness is never sacrificed in favor of ergonomics or performance, and nonsensical operations must always generate an error.

In conclusion, EdgeDB represents a groundbreaking approach to databases, addressing the foundational problems of SQL and providing a more usable, higher-performing database. After over a decade of persistent efforts from its co-founders, the vision of EdgeDB has become reality in a stable and feature-rich database. This development, coupled with its recent successful Series A financing round, position EdgeDB to revolutionize the database industry and improve the lives of developers worldwide.