Wait, what's that rumble in the storage jungle? Yes, it's Ceph
But can the open-source distributed file system, er, make any money?
Open ... and Shut In the open-source world, there are few enablers to success more potent than being distributed within the mainline Linux kernel.
Every open-source company I know aspires to such broad, built-in distribution. Which is one reason why I'm so bullish on the future of Ceph, an open-source, massively scalable distributed storage file system backed by the company Inktank. In fact, for a variety of reasons Inktank may be one of the few organisations truly suited to directly capitalize on open-source adoption.
If you want to know more about Ceph and how it's becoming a central rallying cry for all Linux distributions, Timothy Prickett Morgan has written a great analysis. It's already a Very Big Deal that Ceph is in the mainline Linux kernel, as noted above, as no other competitor joins it there. But it becomes an even bigger deal when you see Ceph also taking center stage in OpenStack.
So far, so good for Inktank.
It gets better. As Inktank's founders told Sean Kerner back in May, Inktank has been very careful to brand itself separately from Ceph, arguing that "nam[ing] the company after the project gives the company a short-term advantage in terms of visibility…[b]ut it steals from the project's long-term viability as an open source project that is tied to one company's fate."
There are, of course, products and projects that have done very well being tied to a vendor's brand (eg, Red Hat Enterprise Linux, MySQL), but there are far more that never got beyond their commercial community to become de facto industry standards.
Is the name to blame? Perhaps. Perhaps not. But it's still a good sign that the Inktank team is concerned with the question.
But this isn't nearly as important a factor in Inktank's success trajectory as others. For one, Inktank's open-source commoditization strategy seems particularly well-suited to win in a still mostly proprietary storage market. To quote Marc Andreessen, software is beginning to eat this market, which has traditionally been based on hardware appliances, and there is a real chance for a new Red Hat to emerge, as it did in the Linux world. In other words, the storage market is ripe for open-source commoditization.
Beyond this, Amazon S3 has shown that developer-friendly and cheap storage products that are accessible via an API are accelerating innovation. Ceph plays into this nicely with a robust set of easy-to-integrate APIs.
Third, Ceph is well-positioned to capture greenfield distributed storage opportunities through its so-called object storage approach. In my experience, it's this greenfield characteristic that should be a great catalyst for Inktank, as it gives Inktank the chance to grow under the radar of the big, incumbent vendors. Since Ceph will be stealing greenfield sales opportunities that never actually hit the radar of the proprietary vendors' respective sales teams, they won't know they're bleeding until Ceph's momentum is difficult to impossible to stop.
To accelerate this adoption, as mentioned Inktank has built the right partnerships on the Linux side, and is now working with every major open-source cloud platform: Openstack, Cloudstack, Open Nebula.. and from what I hear, Eucalyptus soon, too. If you believe that the private/hybrid cloud is real, and I do, then as this market grows, Ceph will grow as there are no credible alternatives that offer the same cost basis.
Finally, one reason I believe in Ceph and Inktank's ability to monetize it is that Ceph is coming up again and again in my discussions with technologists. If the Redmonk (developers are the new kingmakers) and O'Reilly/OATV (watch the alpha geeks) theses are true, then Ceph will soon be top-of-mind for mainstream enterprise IT.
In other words, as open-source opportunities go, I think this one has real legs. Yes, Inktank still needs to answer how it will effectively monetize Ceph, but I think Ceph adoption is almost a given at this point. That's a good problem - how to monetize a lot of popularity - to have. And I suspect the answer lies in operations-type tools, similar to that used by other open-source companies, including my own, 10gen, the MongoDB company.
But however Inktank opts to make money around Ceph, it strikes me as one of the few open-source companies that can assume significant volume of adoption as the foundation for its business. ®
Matt Asay is vice president of corporate strategy at 10gen, the MongoDB company. Previously he was SVP of business development at Nodeable, which was acquired in October 2012. He was formerly SVP of biz dev at HTML5 start-up Strobe (now part of Facebook) and chief operating officer of Ubuntu commercial operation Canonical. With more than a decade spent in open source, Asay served as Alfresco's general manager for the Americas and vice president of business development, and he helped put Novell on its open source track. Asay is an emeritus board member of the Open Source Initiative (OSI). His column, Open...and Shut, appears three times a week on The Register. You can follow him on Twitter @mjasay.
opinions != knowledge
Matt may have opinions but they always have to be taken with a pinch of salt -when I cover areas I work in they're a mix of underinformed and unrealistic.
FWIW, OpenStack mostly uses Swift, which is less of a distributed filesystem and more of a python service for distributing files across a pool of machines.
As for the "only ceph is in the kernel" story. RedHat's GFS? What about Lustre?
Thanks for the feedback. This isn't straight news as such, it's Matt's regular Open and Shut column, which is his opinion on what's going on in the open-source and startup worlds.
Was that a news article I just read, or a copy/pasted press release?
Hard work this journalism lark.