Mention the words "open source" to IT pros interested in adopting cloud computing, and their ears likely will perk up. Open source software offers a solution to the vendor lock-in concerns many enterprises have with committing to a cloud platform. And cloud platforms like the OpenStack Foundation, which fosters 'coopitition' among seeming competitors in the hot cloud computing market, give companies the option to build interoperable open source clouds. But what options do enterprises have when seeking open source PaaS?
SearchCloudComputing spoke with Cédric Thomas, chief executive officer of OW2, a Paris-based, independent community focused on creating a code base of open source software for cloud computing. Similar to Apache and the OpenStack Foundation, OW2 concentrates on infrastructure software as well as the tools to develop, deploy and manage applications for approximately 100 projects over a 60-member community that's 22,000 developers strong.
How does OW2 differ from Apache, OpenStack and other open source cloud platforms?
Cédric Thomas: We were founded out of Europe; our members are mainly corporations, organizations and universities rather than individuals, which is the case for Apache. Other than that, we don't differentiate in terms of technology.
We recently had meetings with OpenStack and realized there was no competition at all between OW2 and OpenStack, because everything OW2 does is actually a layer above that. OpenStack concentrates on the Infrastructure as a Service (IaaS) layer with the compute, network and storage functionalities. Everything OW2 does targets the Platform as a Service (PaaS) layer -- the applications or the application platforms for the cloud. All our cloud projects now have OpenStack as a reference for the development platform.
Cedric ThomasCEO, OW2
We are implementing on our own infrastructure because we host services for our members. We are implementing an OpenStack cloud platform so that our project and test development will [run] on OpenStack. You can think of OW2 as bringing value to the OpenStack ecosystem. We didn't want to find ourselves in competition with them because, in the open source world, there is no reason to compete. We want to reuse or expand in areas that are not covered by our open source friends.
OpenStack talks about the idea of 'coopitition,' which joins competitors together to foster cloud computing projects. What are the challenges with this coopitition approach?
Thomas: Up until 2010, open source was in competition against proprietary software. Every category that has been developed, invested, marketed, explored, sold, supported or maintained was done by proprietary vendors who eventually sold in the open source world. The way it was … open source software just copied proprietary software. You can even say it started with GNU; it's not unique and it was defined, by comparison, with something that already existed.
Something changed in 2010. We realized cloud computing was not just a marketing fad and decided to position OW2 to cloud computing. We realized open source was, with cloud computing, racing -- not following -- proprietary software. That open source innovation was driving innovation in cloud computing. So, open source was the best way to collaboratively innovate with cloud technology. With open source, the basic relationship is defined by the open source license, which makes open source a very easy way to cooperate, collaborate and innovate -- even if vendors are competitors.
We support a handful of cloud projects and our vendors are also potential competitors. The whole open source movement will deliver technologies that will not be products. That's important. I think open source will help companies reach a level of technology or state of the art with technology, and then they can define the offering. The real challenge is that open source will deliver a half-baked product or a product that's only 80% or 90% finished, then vendors will finish them on a commercial basis. That's what we're doing with some of our projects, and that's what happening with OpenStack.
On the PaaS front, are you competing with proprietary solutions?
Thomas: At the Infrastructure as a Service (IaaS) level, all the technologies are pretty much identified. And it's all a matter of just productizing and getting it to industry-grade offerings. And that helps companies deploy workloads to cloud computing platforms.
Now what's at stake and the real challenge is to develop the cloud-oriented information systems for companies and certain services. These cloud-oriented information systems will leverage social applications, data, resources and services -- some of these are inside the enterprise and some of them are outside the enterprise. So, cloud-oriented information systems will be totally hybrid. And that's what PaaS is about: To help companies and CIOs develop the next generation of information systems. What we usually call PaaS, at its very simplest form, is where this innovation will take place. We develop the middleware that uses resources provided by different technologies of different companies.
Some criticisms of OpenStack are that it hasn't been ready for full production or critical updates happen a bit slower than the community would like. Is that part of the problem with having so many competing vendors working together?
Thomas: You're looking at the glass half empty. In 2010, there was nothing at all. There were three people; now [OpenStack] organizes events with 3,000 people in less than three years. You can't go any faster than that. That is bringing and converting a lot of expertise, a lot of feedback. That's what happens in open source collaboration projects; vendors all contribute, but at some point, they all look for their own interests. It's an industry in the making.
Since you feel huge changes can happen in two years with open source, what do you see happening in the next two years?
Thomas: There will be many more OpenStack distributions out there. IaaS will somehow be more commoditized. It will use community off-the-shelf [COTS] platforms. I think the hosting industry will be consolidated; there will be a small number of large hosters leading the market. In North America, it will be Rackspace, Amazon Web Services and Google; in Europe, each country will develop national hosters to deal with data compliance issues. Each country will have a handful of national hosting companies quite often controlled by the large telcos.
[These vendors] will offer services and applications and added value. OpenCloudware, for example, develops the technologies to manage lifecycle applications over the cloud. We're not talking about infrastructure. It addresses the needs of the application and looks upward in the information stack. The focus is on the applications and all the tooling you need to manage them.
The new generation of applications will have a cloud-oriented philosophy. It's not an application that will leverage databases or processing cycles; these new apps will be service-oriented and will call on Web services provided by third-party computing companies. Everything will have to be standardized. These applications will also be deployed across servers and mobile devices. Servers will need to handle the processing requirements of mobile devices. This is a new architecture and new methodologies in software engineering as well.
Michelle Boisvert is executive site editor for SearchCloudComputing, SearchWindowsServer and SearchDataCenter. Contact her at firstname.lastname@example.org.