OSCON Week: Openstack.org, A Game Changer?

Yesterday, we saw the announcement by Rackspace that they have joined hands with NASA and 25 other companies to create Openstack.org, an open source cloud community. Rackspace also announced that they are open sourcing their entire cloud stack and give it to the project. As expected, it has created both positive and negative reactions from the pundits, community, proprietary vendors, AWS fanboys, etc.. Even though many in tech blogosphere jumped on the news immediately after the embargo was removed, I wanted to talk to Rackspace, other companies supporting Openstack.org and other open source folks I respect. After a delay of one day, here is my take on the announcement in the typical Redmonk Q&A model.

Rackspace’s Email division is a client of Diversity Analysis but this is about their Cloud Division. More importantly, it is about Open Source and not Rackspace per se. I also want to state my inherent bias towards both open source and cloud computing. I am fairly convinced that these two are going to change the way we do computing in the future. Once again, I also want to state that this analysis is based on the talks I had with not just Rackspace but some of the vendors supporting Openstack and independent open source pundits and analysts. 

Openstack.org is a common platform for a collection of open source projects that can be used to build clouds. It is the brainchild of Rackspace Cloud but there are more than 25 companies on board now including NASA. The story goes like this. Rackspace, for whatever reason, wanted to open source their cloud stack. They talked to NASA folks working on Nebula Project to gauge their interest. NASA folks got interested and then Rackspace started talking to other companies big and small (including one man open source projects) and asked them to come to Texas for brainstorming, explained their design and called on them to jump on board. 25 companies came on board immediately and many more are trying to understand what is on table for them. 

With Openstack, anyone can build a cloud using open source components. Right now, Openstack has two active projects and more are expected to be available in the future. The first one is Openstack Object Storage which is nothing but the open source version of the software that is driving Rackspace’s Cloud files. The other one is Openstack Compute, the computing side of the cloud. Right now, it is in the Developer Preview mode and they are working hard to put together Rackspace’s compute code with code from NASA’s Nebula Cloud code.

I had the same doubt and asked the Openstack guys about it. They said that they are talking to all the virtualization players and when Openstack comes out later this year, it will support many hypervisors. They said, even if others are not on board by then, Openstack will come out with support for both Xen and KVM. This is where it gets interesting. Rackspace Cloud is a Xen shop and Nebula has strong support for KVM. It is not clear what will be the impact of “Hypervisor politics” on Openstack. For those who keep their toes immersed in the open source politics, we all know that KVM is strongly gaining at the expense of Xen. In fact, when I spoke to The Planet on their initial stage cloud offering, they told the same about KVM and its momentum. In fact, Planet was even more categorical in saying KVM is the future. I would love to see how this thing plays out. However, as I mentioned above, the folks at Openstack.org tell me categorically that they will support all hypervisors including VMware and Hyper-V (if they are interested).

Quite possible. There is a very good chance that Rackspace might push the project to match their interests. When I asked the Openstack folks about this, they categorically told me that it will not be the case. They pointed out to me that there are 25+ companies participating in the project and they wouldn’t have come on board if they had suspicions about Rackspace’s intentions. They also pointed out that every company or interested parties can push their patch/features and it will be discussed among everyone participating in these projects and get committed based on merits. 

In fact, I am not all too worried about Rackspace forcing their way into the project. Such concerns comes due to the lack of understanding of open source. As long as the license is open source, I have the right to fork any project if I feel that people in charge of committing to open source projects are driving the project to help their own interests, I can fork the project and drive in a different direction. This freedom is what differentiates open source from the cathedral proprietary development model or what I like to call communist model. Few years back, when Nick Carr did a little FUD on Open Source, I wrote a blog post highlighting this point. Relax, even if Rackspace wants to drive the project in a direction that is not to the liking of other participants or open source community in general, anyone can fork it and drive it in the right direction. 

They are working on a Governance model that will soon address these concerns and more.

Possible. I did ask them the same thing and they told me that Rackspace is committed to open source for a long time and pointed me out to the other open source projects they support and the number of open source developers on their payroll. I do agree with them. They do support many open source developers and projects for a long time. Even if we take all this as marketing fluff, I don’t really care what Rackspace’s intentions are. It may be a Hail Mary pass or it may be a desperate attempt to establish their legitimacy in the cloud world. All I care about it is freedom. A freedom to set up my own cloud in any place I want in any way I want. The freedom to see what is under the code driving the cloud I am using. This freedom matters. Rackspace’s intention may be noble or a pure business decision but open source wins. That is the bottom line.

Yes. Let me explain why it goes beyond my bias for open source. Even though folks like Tim O’ Reilly (and at the Clouderati level, Sam Johnston) emphasize that architecture and formats triumphs licensing, I have been pushing strongly about the importance of open source in the cloud based era. I have pointed out many times how open source plays an important role in the cloud world from IaaS all the way to SaaS. More than anything else, open source is the only savior against the predictions by pundits that the infrastructure and platform space will be dominated hy a handful of monopoly players. I have argued against this idea for long and I have been advocating open federated clouds as an antidote against this possibility. Open architecture and formats may be good enough to have an open cloud ecosystem but we need open source to have a more federated cloud ecosystem. If the underlying technology to build clouds are proprietary and expensive, we are not going to see proliferation of cloud providers emerging in the market. Such expensive proprietary technologies can only lead to consolidation and, eventual, monopoly. If we really want an open federated cloud ecosystem, open source is equally important as open architecture and open formats. Openstack gives us the necessary technology under the open source license. And, this is the reason I am excited about Openstack. 

Not at all. Many people who dismiss Openstack.org as a non-starter fail to understand that it is just a start. Openstack.org doesn’t claim it has the miracle pill for all the ills in the cloud. It is a new initiative and it is just starting off. In fact, the compute code is not even ready. In my opinion, it is a start and any outright dismissal or claims of victory is not only premature but also amateurish. The only thing it offers is a potential to disrupt the cloud marketplace in the future. There are many factors in play before we can even see the results of this move. If they execute it right, it can stir up the marketplace. It may not destroy the leading players in the market. The best I expect from this initiative is that it will level the marketplace by making it easy for smaller players to jump in and try their chance. I could compare it to what open source did to traditional IT. It didn’t dislodge the existing powerful players. Rather, it created a competitive market conditions and forced the market leaders to listen to their customers. This is very important for the cloud market, especially its users.

• Check out the video podcast by Barton George, Cloud Computing evangelist at Dell, on Openstack.org. He talks to Rich Clark, Will Reese, Lew Moorman and Joshua McKenty.