10gen

Shutterfly, a leading Internet-based social expression and personal publishing service, safeguards more than six billion images for millions of customers. As the only photo sharing site that does not down-sample, compress, or force delete photos, Shutterfly faced massive data growth that pushed the performance limits of its existing Oracle database. After an extensive analysis of open source relational and non-relational alternatives, Shutterfly chose MongoDB as its persistent data store.

Before implementation of MongoDB, Shutterfly stored its more than 20 TB of photo metadata in an Oracle RDBMS, vertically partitioning the data by function. The complex infrastructure became hard to manage after a decade of development and made development of new features difficult to execute quickly or correctly. The situation was further exacerbated by the the high costs associated with licensing and hardware. These challenges compelled the Shutterfly team to look at open source options that would also provide improved performance and a simpler API at a reduced cost.

Shutterfly chose MongoDB as its new storage solution primarily because the data model matched common use cases. The rich JSON-based data structure was easy for the development team to use, reduced time to market for new features, and could be leveraged for many different projects. In addition, Shutterfly saw significant performance improvements with the more natural data model. Finally, the MongoDB solution was cost-efficient, providing an open source solution that would enable them to scale horizontally across commodity hardware.

Shutterfly’s migration to MongoDB resulted in a 900% performance improvement when compared to their previous Oracle implementation. In terms of cost, Shutterfly realized a 500% reduction in moving from Oracle to MongoDB. Overall, MongoDB provided Shutterfly with a high performance solution at a significantly reduced cost.

To learn more, read the full case study, or for more resources, check out Shutterfly’s Kenny Gorman’s presentation at MongoSV 2010 “Sharing Life’s Joy using MongoDB: A Shutterfly Case Study” or the follow-up Q&A with Gorman hosted by NoSQLDatabases.com. More recently, at MongoSF 2011 Gorman presented on “MongoDB Profiling and Tuning.”

10gen CEO and Co-Founder Dwight Merriman visited the London MongoDB User Group on January 31st, presenting Crazy Stuff: hacks, internals, and sneaky tricks.

Early this morning, the 10gen team watched as the last tickets to MongoDB Boulder were claimed. Tomorrow, the sellout crowd in Colorado will enjoy presentations from 10gen engineers Kyle Banker, Dan Crosta, Robert Stam, and Ryan Nitz, as well as from guest speakers James Ward, Brian Fromme, Krishnanad Khambadkone, Nathan Wells, David Blado, and Chris Merz. Attendees will also hear from 10gen’s Presiden, Max Schireson.

Following Boulder, the 10gen team has a busy spring ahead. A free meetup on February 6 in Brussels will give the city’s MongoDB fans an opportunity to learn and ask any nagging questions. On February 15th, developers and other industry professionals from across the Lone Star State will meet for MongoDB Austin to hear from engineers at 10gen, X.commerce, and RedHat. Beginning February 28th, the 10gen team will also bring MongoDB to Strata, a conference dedicated to the business and practice of data innovation, hosted in Santa Clara, CA.

Not to be forgotten, our webinar series continues with:

-Monitoring Your MongoDB Deployment (Feb 9)
-MongoDB Schema Design: How to think non-relational (Feb 17)
-MongoDB Schema Design: Principles and Practice (Feb 23rd)
-Building Your First MongoDB Application (Mar 15)

If you’re already using MongoDB in production, tell us about your application, or contact us to learn more about training and other solutions from 10gen.