The Presence of Open Source in Cloud Computing
August 5, 2008 by MatthewSacks
The popularity and emergence of cloud computing, the question of the role of open source comes into question when navigating the clouds. Some of the questions that are raised about this topic are will the major players stick to an open platform for cloud computing, or will it be an exclusive network of key technology players with proprietary systems and large budgets?
I suspect that the answer to that question will be both. It seems that with any new major development, theology or practice within computer science – the open source community can play a major role in making a software development project or practice achievable and successful. History has proven that open source development practices can lead to robust and well-supported and integrated software systems. The Mozilla project is one great example of open source saving the day from a closed-source, limited software.
Specific to cloud computing, it seems as if the largest contributors and proponents of the cloud computing architecture are large technology multinationals with gargantuan research and development budgets as well as distributed data centers. Wouldn’t it make sense that distributed software architecture such as cloud computing platforms be distributed amongst a large quantity of software developers? One would think so, that open source development and the architecture of the software would seem to coincide; however, in the most recent developments such as Amazon’s EC2 (and other suite of cloud services) and IBM’s Blue Cloud aren’t exactly what one would call an “open” platform - at least not yet.
There is hope for open source’s role in the cloud computing platform with developments such as the Eucalyptus project, and 10gen. Such open cloud computing platforms show promise for cloud computing. The necessity for open source software for the cloud-computing platform is highlighted by recent discussions about being “locked in” to a cloud provider. It seems as if running software in a commercial provider’s cloud is convenient; however, it presents various sensitive issues with intellectual property, security, and privacy. If I run my businesses application within Amazon’s EC2 cloud, does that mean they have access to my application’s information? I haven’t been one to spend my free time reading EULA’s, but I am sure if one were to pick through it, there is a high probability of a clause in it stating that your data can be the provider’s data if they so choose.
Open source clouds prevent the necessity for relying on commercial providers for cloud access and gives promise to those who would prefer to run their own cloud platform, making modifications to the cloud infrastructure as desired. It makes sense to invest more into the open source initiatives with a cloud architecture. One potential obstacle for making such developments is that open source developers may not have access to physical servers (the cloud has to run on something) worldwide in which they can develop, test and tune the infrastructure that is the cloud computing platform. Perhaps a more P2P-style architecture would be suitable for the development of open source development for the cloud for those without access to regionally dispersed systems. Perhaps it is not an issue at all, and a better question for the Eucalyptus forums to see how they address this obstacle.
Perhaps the open source community is allowing the large tech firms with large budgets to do the experimentation first, and then once some headway has been made the open source community will join them in the cloud. The entire practice is developing and unfolding as I write this article, and it is too new to predict exactly what will happen as of yet.
As with most software development projects, there will be a proprietary, closed-source, tech-elite faction offering the next best cloud, and the open-source community sharing and contributing to the robustness of the cloud on a much larger scale.
Comments
People
Apparently, my comment about IBM’s usage of proprietary systems in their cloud architecture was a bit askew after standing corrected by this article: http://blog.xen.org/index.php/2008/08/05/ibm-cloud-computing-and-xen.
To clarify, IBM’s platform may not be open-source, but the systems which they deploy it on are apparently Xen!