What Is Open Cloud?
I've read a bit of angst about cloud lock-in, a lot of weed pulling in the form of interoperability standards for the cloud, and a manifesto or two about 'Open Cloud'. And in between, I've seen lots of interesting new tools for cloud computing, and lots of narratives about how the tools, combined with the formalization of use cases, pave the way for open clouds.
But what, exactly, does "Open Cloud" mean? And what role does open source play? Jim Zemlin, Executive Director of the Linux Foundation, likes to say that open source and the cloud go together like peanut butter and chocolate. But does open source necessarily mean open cloud, and vice-versa?
I'm not at all convinced that simply using open source software means that cloud-based services will be open, too. Nor do I think that *not* using open source software necessarily means a 'closed' cloud. Of course, what Jim means is that without open source, cloud computing would not be possible. This is true; I'm just not sure how relevant it is. After all, automobiles are impossible to use without roads, but nobody seriously questions whether we'll continue to pave over dirt.
Jim also believes that cloud vendors will be incentivized, just like their earthbound ISV brethren, to participate in open source communities.One can see this already happening. Facebook, Yahoo! and Google all participate in massive open source projects, often to great fanfare. These are without a doubt good things, and yet there are plenty of developers and users alike who question the motives of larger players and whether they will 'play nice' to avoid screwing over users after they get more industry leverage. Clearly, something is missing if we're to create the same sustainable model as we did with open source / free software.
When I think of "Open Cloud" I think of open source's friendly cousin - someone who acts similarly but has grown up in a different household and with slightly different values. 'Open Cloud' doesn't refuse open source code, but it doesn't require it, either. Rather, think of 'Open Cloud' as a definition of explicit rights or freedoms, much like those in the Open Source Definition
I think we've reached the point where some of these rights are beginning to take shape:
- The freedom to access data in both raw and packaged formats at any time
- The freedom to build any application or service on top of data, regardless of technology, platform or commercial provider
- The freedom to mimic services and "hotswap" data stores to competing services
What I'm trying to get at, above, is that any discussion about 'open cloud' is really a discussion about data. And not just the kilobytes of storage, but the data vectors - the context, velocity and ultimate destination of the data. And what about ownership? Who owns the data? Do we have the right to dictate how the data is owned? Should we have the right to dictate the vectors of that data? In a world where data is currency, all of these things matter.
See, Tim O'Reilly tried to warn us. He tried to tell us that the rules of open source that we were accustomed to were increasingly irrelevant in a post-web world. And by golly, he was right.
So tomorrow, April 14, i have the privilege of moderating a panel consisting of Sam Ramji (Codeplex Foundation), Doug Tidwell (IBM), David Lutterkort (Red Hat) and James Urquhart (Cisco) as we discuss "Does Open Source Mean Open Cloud?". I have no doubt that we will resolve nothing, but hopefully we'll steer the conversation in the right direction.
The Linux Foundation does a great job of putting on the Collaboration Summit, so you should participate even if you're not interested in this conversation.