MongoDB Case Study: Wordnik

At six times the size of the Oxford English Dictionary, Wordnik helps bring words to life by showing the conversations happening around them. Because Wordnik relies on data from real-time web posts, they needed a reliable and performance-tested database solution.

In 2009, Wordnik engineers realized that their existing MySQL data store couldn't keep pace with the user-generated content which constantly expanded the site's dictionary. The engineers created a prototype to test MongoDB, migrating five billion records to the non-relational database in a single day. Over the course of the next month, the entire database was migrated to MongoDB under the watchful eyes of a single developer.

MongoDB now handles every site request sent to Wordnik—often over 20 million API calls per day, from millions of unique users each month. Under their previous MySQL system, Wordnik's IT engineers frequently dealt with locked tables and outages when too much data was added at once; using MongoDB as a non-relational database solution has eliminated the problem, and the system handles bursts of as many as 50,000 words per second during busy periods without breaking a sweat.

For Wordnik's engineers, life with MongoDB has meant faster data retrieval and 75% less code. To read more about how Wordnik has deployed MongoDB (and how it changed day-to-day life for developers), visit the full-length case study, or check out Wordnik VP of Engineering Tony Tam's presentation at MongoSV 2011.

Tagged with: mongodb, 10gen, wordnik, nosql, database, data store