When you are considering building a cloud, there are many important decisions to make. Among the top decisions you'll want to make is how to implement your storage. One open source project which many have turned to as a solution to some of their storage needs is Ceph.
To learn more about Ceph, I attended a Cloud Online Meetup last week to hear from Ceph's community manager Patrick McGarry. I knew some of the basics from when Red Hat acquired Inktank, the commercial entity behind Ceph, last year. But I was eager to learn more about the history of the open source project, what it's currently working on, where it's headed, and of course, how it ties in to OpenStack.
The full slides of Patrick's presentation are below, but I've pulled out some of the highlights which were the most interesting to me.
For those not yet acquainted with it, Ceph is an open source project which offers object, block, and file storage in a distributed cluster. It is designed to be fault-tolerant, and Ceph can run on commodity hardware, but can also be run on a number of more advanced systems with the right setup.
A few things I learned:
- Ceph's creator, Sage Weil, observed when creating the project in graduate school that many people working in the field were going on to work at companies offering proprietary storage solutions. More often than not, their talent would be acquired, and then their projects would fade away.
- At the time it was first being created, in 2006, there were a number of large-scale storage projects, but they tended to either have limited scalability, limited community, limited architecture, or limited enterprise features.
- Part of Ceph's speed comes from its integration with the Linux kernel, which was first merged in 2010.
- One slide which really stood out (below) is that graph of the Ceph community, which shows the exponential growth of the number of authors contributing to Ceph over time. Clearly the interest in cloud and storage is on the rise.
Interested in watching the event? The full video clocks in at just under an hour, and is embedded below.
The Cloud Online Meetup series is a great place to learn more about what's going on with OpenStack and the related open source projects in the broader cloud and OpenStack ecosystem. In this past week, I've attended sessions on autoscaling MySQL database services, the MidoNet network virtualization project, OpenStack's Trove Database-as-a-Service project, and there are lots of other upcoming events; if you're interested in what's going on with cloud, I highly recommend you check out their calendar.