Posts

Open Source and Mozy: The Debut of Mozy Code

This article was originally posted on Mozy’s blog.

The popularity of Mozy online backup confronts our developers with intense engineering challenges. Consider the scale that Mozy handles:

• More than 1 million total customers
• 60,000+ business customers
• 50+ petabytes (or 50+ million gigabytes, if you prefer) of data
• Customer data stored in multiple data centers around the globe

These numbers increase every day, relentlessly pressing us to keep refining Mozy’s storage infrastructure for faster throughput and more efficient storage.

Five years ago, Mozy was an aspiring startup with a vision for an affordable way to make data loss a thing of the past. We could not have achieved this goal without the availability of high-quality, freely available, open source software. We have relied upon, and contributed to, numerous free software projects to build the Mozy business.

Today, we step beyond simply contributing to existing projects with the public debut of Mozy Code. Mozy Code’s roots began with an internal program for fostering innovation among Mozy engineers called Mozy Labs, started in 2009. It soon became apparent that some of these internal projects could be of value to developers outside of Mozy for use in their own endeavors. For that reason, we have begun to identify some of our most promising and useful innovations, to make them available as Mozy-lead free software projects.

With Mozy Labs’ debut, we introduce two projects focused on high-speed I/O operations:

• Mordor - Mordor is an I/O library based on fibers that allows you to focus on performing a logical task, instead of how to make that task conform to a specific threading/event model. It provides lightweight, easy-to-use abstractions for complex concepts, while still providing superior flexibility. It can be used to build high performance network servers or simple data processing applications.
• Ruby Protocol Buffers - Originally created by Google, Ruby Protocol Buffers provides a way to encode structured data in an efficient yet extensible format. This library is optimized for rapid encoding and decoding, enabling high-performance communication between Ruby and C++ systems.

You can find out more about these first two projects on the Mozy Code website. Keep an eye out for future projects as we grow our list of projects.