Looking Back 2010: OpenStack Offers Promise

Looking Back 2010 is a series of posts I am planning for this week in which I will highlight significant cloud related events that happened in this year. Please keep in mind that these are my personal opinions based on my industry observations this year. My first pick is OpenStack project (see previous CloudAve coverage) and, in this post, I will lay out my case for the pick. Over the next few days, I will write about other important cloud related events in 2010.

First, a little bit history

OpenStack is an open source IaaS project lead by Rackspace and NASA with support from 40 other companies including Dell, Cloud.com, enStratus, Citrix, etc.. Right now, they have two active projects on Compute and Object Storage. The code for compute came from NASA’s Nebula Cloud project and the code behind their Object Storage came from Rackspace.

Quoting their website

OpenStack compute is the underlying fabric controller for the OpenStack cloud. It is more like an OS for the OpenStack cloud and handles processes like starting and stopping of virtual machines, etc.. They currently support Xen, KVM, QEMU and User Mode Linux. OpenStack Object Storage is a software layer that creates redundant, scalable object storage using clusters of commodity servers to store terabytes or even petabytes of accessible data. The storage supports ACLs for stored objects.

In October, OpenStack released the Austin release of their platform with their Storage code ready for production deployment and Compute code in some sort of an alpha release. They plan to have their second major release, Bexar release, in the first quarter of 2011. With this release, the compute code is expected to mature and, hopefully, 2011 will see further maturation of the codebase and large scale adoption of the platform.

Why I think OpenStack is big

Conventional wisdom predicts a consolidation among the IaaS players with a handful of monopoly players in the market. However, I strongly disagree with CW and I always argue that we will end up with a more open and federated cloud ecosystem. In fact, in July 2009, I laid out an argument, based on political, philosophical and macro-economic considerations, against the “handful of monopoly infrastructure players” idea. I even pointed out to some examples to support my argument against such consolidation. I am pretty convinced that even if there is some level of consolidation in the infrastructure space, we will still have an ecosystem with large number of players competing with each other and differentiating through value added services on top of infrastructure offering. Having explained my vision for the future cloud ecosystem, I see OpenStack as a critical first step to get there. I see this project as an important player with a potential to democratize the IaaS marketplace.

Let me list out some of the reasons why I think OpenStack will be a game changer:

First, and foremost, it is a non-commercial open source project with a flexible Apache license. It could completely disrupt the market and can play a crucial role in the democratization of IaaS field. Even small webhosts without the huge capital infusion needed for setting up infrastructure services can take advantage of OpenStack and repurpose their existing infrastructure as cloud. This will help spring up a cottage industry of IaaS players who can add additional value on top of infrastructure services like geographical location support, local presence (read about regional cloud providers in this regard), managed services, regulatory compliance, etc.
Second, it is truly open source with a good governance model. Even though Rackspace has pumped in large amount of resources, the governance model will ensure that it stays true to the open source spirit. Unlike Eucalyptus, which uses the open core approach and keeps control over the features and roadmap, this is more democratic with competing interests driving the project towards vendor neutrality, thereby, empowering users
Third, one of the biggest issues with the cloud is lock-in. OpenStack is designed to support Open Standards from ground up. So, any ecosystem spawned by OpenStack will be interoperable without any vendor lock-in. The fact that OpenStack supports/plans to support the APIs of other public cloud provider makes it an attractive candidate for any open federated cloud ecosystem
Finally, OpenStack has the support of almost 40 other vendors making it one of the most versatile cloud platforms. The fact that these vendors have invested their time and resources on OpenStack means that their services will be deeply integrated with OpenStack platform.

There are many other reasons to get excited about OpenStack but I have only listed the four I thought are important from the perspective of having an open federated cloud ecosystem future.

One more thing

Even though I am excited about OpenStack and the open federated future it presents, it is too early to claim that OpenStack is going to disrupt the IaaS market upside down and become the dominant cloud platform. Eventually, it all boils down to execution. It is a great vision and I hope it is backed up by proper execution. I am picking the setting up of OpenStack project as one of the most important cloud related story just based on their promise.