Is Cloud Foundry something we need?
VMware's new platform-as-a-service offering is not open for open's sake
Following the technology offerings in open source cloudspace is a bit challenging of late, what with all the PR and hype about "we're open!" flying about.
(Of course, over in mobile space "we're open" actually is code for we're not open source anymore, eh, Nokia?)
What, then, are we to make of VMware's recent Cloud Foundry announcement?
If you missed it, and I think that unlikely given the buzz the announcement got this week, Cloud Foundry is VMware's open source platform as a service (PaaS). And, unlike last week's Facebook hardware announcement, this baby is actually open source, released under the Apache license.
The main idea of Cloud Foundry is enabling the extension platforms across multiple cloud providers. The service enables developers to pick and choose the framework, cloud, and application services for their applications. So, write once and off you go. The code and other goodies for Open Foundry are at CloudFoundry.org. You can try out the service over at CloudFoundry.com.
But let's not kid ourselves here. VMware's offering to the community, while interesting and useful, is still just another attempt to circumvent developer and vendor interest in Red Hat's cloud offerings. The competition between VMWare and Red Hat has become a fierce rivalry in this tech sector. Industry watchers know this, and every once in a while someone at either company slips their smiles about cooperation and lets their true feelings show.
Part of the heat from this rivalry comes from VMware's resentment at Red Hat's playing in their sandbox. I'm reasonably sure VMware would have preferred that virtual datacenters and VDI would have stayed useful in the enterprise longer, rather than the cloud leapfrogging such technologies and making them almost obsolete. And then to have an operating system company roll in and provide an end-to-end cloud stack? Unthinkable.
Yet, here we are, with VMware playing catch-up to Red Hat's open cloud offering, Cloud Formation. Cloud Formation is not a single project, per se, but rather an umbrella term for the sixty-plus cloud-related projects with which Red Hat is involved. That VMWare emphasized its own vFabric as a compatible application service in the Cloud Foundry announcement is no accident, either: all the better to detract from Apache's Deltacloud, which is shooting to be the open standard for such services. (Deltacloud, you may recall, was started by Red Hat in 2009, and then later contributed to the Apache Software Foundation.)
While I am not naïve enough to think that there won't be competition in any business and technology sector, it is a bit aggravating to see yet-another open cloud service start up when there are other perfectly good services out there. So, as usual, I am taking VMware's commitment to open with a big grain of salt. Open is all right, it seems, as long as it's our open.
This is where we get into the diversity vs. uniformity argument that is a central point of debate in the open source community. Diversity is choice, and choice is good. Versus uniformity breeds standards, and standards are good. I tend to lean towards the standards, if only because I know that there are a finite number of IT professionals to go around, and too much choice will take away resources from one project or another.
On it's own, Cloud Foundry is certainly an interesting new offering. But is it something we really need?