News

The content has been bookmarked!

There was an error bookmarking this content! Please retry.

StrokeDB, Just Another Distributed Database? Not Really.

Posted by Sebastien Auvray on Apr 22, 2008

Sections: Architecture & Design,; Development,; Operations & Infrastructure
Topics: Data Access ,; Database Design ,; Ruby
Tags: Database

As Distributed Databases get more and more interest, implementations are flourishing. CouchDB showed the way to go and is now incubated as an Apache project. RDDB was one of the first Document-Oriented Distributed Databases implemented in Ruby. StrokeDB is a new actor in the scene. It was written by Yurii Rashkovskii and Oleg Andreev, who did a presentation at Euruko2008 (PDF slides). From the StrokeDB website:

StrokeDB is an embeddable distributed document database written in Ruby. It is schema-free, it scales infinitely, it even tracks revisions and perfectly integrates with Ruby applications.

StrokeDB is only 3 months old and offers already many interesting features, the basic ones for Distributed Databases:

Flat address space of documents identified by UUIDs.
JSON, schemaless document format.
References to other documents with automatic eager loading on access.

And some others that could differentiate itself from its competitors:

Documents revision control with diff/merge facilities built-in.
A flexible object-oriented API.
Simple search index over document slots.
Possibility to write native code for very specific performance issues.

Many more features are in the pipeline.

There's also a promise to port StrokeDB to thin client languages (JavaScript, ActionScript, etc.) to enable offline work with the data.

You can check out the StrokeDB code at GitHub and have a look at Yurii's short introduction to StrokeDB. While the authors are busy writing a clean API and adding new features, it will be interesting to see how the project evolves with maturity and later to have a look at benchmarks to check how it performs.