Wordnik's case study from MySQL to MongoDB

About Wordnik

Wordnik is based on the principle that people learn words best by seeing them in context.

It’s an online dictionary that uses words and definitions that are in “traditional” dictionaries and ones that have not made it there yet. The Wordnik site wants to show you all the conversations, pictures, and other talk about your word.

The case study

One year ago today they started the investigation to find an alternative to MySQL to store, find, and retrieve our corpus data. After months of experimentation in the non-relational landscape (and running a scary number of nightly builds), they settled on MongoDB. The primary driver for migrating to MongoDB was for performance, one year later they posted an article about their experience on moving from MySql to MongoDB and provide some interesting statistics:

MongoDB servers an average of 500k requests/per hour. With 4x that number during peak periods
• More than 12 billion documents stored by MongoDB
• Around 3 TB storage per node
• Easily sustained insert speeds of 8k documents/second, often with bursts to 50k documents/second
• Every retrieval type is faster than MySql
• example fetch time reduced from 400ms to 60ms
• dictionary entries from 20ms to 1ms
• document metadata from 30ms to .1ms
• spelling suggestions from 10ms to 1.2ms

Leave a Reply

You must be logged in to post a comment.