There is a big fight going on about who will provide the ‘operating system of the cloud’. Participants are a number of large names offering their own closed solutions, like Microsoft, VMware and Amazon. There seems to be a huge demand for an open source cloud platform and a number of projects are trying to fill this gap. According to recent research of the size en involvement of their respective communities, OpenStack is now clearly in the lead.
In the meanwhile an increasing number of cloud users are concluding that there are significant risks associated with attaching themselves to one solution provider. Cloud providers would like to have a software platform that allows them to effectively compete with the closed solutions.
Because of these two developments there seems to be an enormous demand for an open source alternative that is expected to develop itself into the Linux of the cloud. A number of projects has started in this area over the last few years and the question now is which alternative will become dominant. Large companies and governments are not expected to be willing to work with different open source cloud systems so one project will probably become the clear market leader.
There are currently four serious projects that want to become the basis of the cloud-infrastructure. The most important factor that will determine the winner is the size and involvement of the community: How many people are using the project and how many people are active contributors. Qingye Jiang of Eucalyptus, one of the four competitors, produces a quarterly report that compares the size and involvement of the four communities.
OpenStack – After the project was started in juli 2010 by NASA and Rackspace, this project attracted a huge number of well know players in a very short time. The project develops itself quickly through a half yearly release-cycle. OpenStack does not attempt to link up with any closed solutions like Amazon’s.
CloudStack – This is a young project that is launched in februari 2012 by Citrix on the basis of the cloud.com platform that they had acquired shortly before. Citrix tried to integrate CloudStack in OpenStack first, but reported that OpenStack was behind in its developement at that point. Sources within OpenStack report that Citrix was dissepointed when OpenStack did not want to absorm CloudStack in it entirity however. Whatever the case: CloudStack went further on its own and sinds April 2012 the project is also supported by Apache.
OpenNebula – This project originated in the European accademic world. There are a number of important deployments active, but they are with the somewhat less well known companies and research outfits.
Eucalyptus – Eucalyptus was first released in May 2008. The large advantage of this solution is that it integrates very well with Amazon. That is an advantage for some applications but the less than arms lenght connection with this large provider ensures that other serieus parties, that are not inclined to accept Amazon dominance, steer well clear of this solution.
Oingye has compared the projects in a number of different ways. First he looked at the number of different subjects that are discussed on mailing lists and forums (Figure 1 of Eucalyptus report). OpenStack has a clear advantage on other projects here, but a subject generate more answers when it is posted with the CloudStack community (Figure 3).
The number of people that participates in discussions is a lot higher with OpenStack than with the other projects (Figure 4). This seems especially the case since the last Grisley-release in April of 2013. It is likely that OpenStack is being tried out by a large number of new users (Figure 6 & 7). CloudStack also features a decent number of new participants, but has not been able to keep up with OpenStack for the last half year.
Participants that join the CloudStack as well as the OpenStack communities often stay active. Although OpenStack also has an advantage with 36.3 percent versus 24.2 percent for CloudStack. This percentage is a lot lower with Eucalyptus and OpenNebula (Figure 8).
Finally Qingye looks at the actual software contributions to the projects. Looking at this OpenStack has an enormous lead over other projects with CloudStack again at number 2 with some distance. This is the case for the number of contributions (Figure 10), as wel as for the number of contributing parties (Figure 12 & 14).
It seems that OpenStack and CloudStack have attained a large lead over the other two projects. These alternatives are likely to develop themselves as niche players. Looking at the two serious candidates OpenStack seems to have the advantage. This is especially the case is we look at new parties that start to use the platform and new parties starting to contribute to the project. If the new ‘Havana’ version of OpenStack, that is expected later this month, will generate enough community interest and involvement to keep this momentum going then OpenStack seems to be in a very good position to win the race.
What also draws attention is that CloudStack seems to have more traction than on other continents. A relatively large number of European and Dutch parties contributes to the project and the next collaboration conference will be located in Amsterdam. OpenStack seems to have a significantly stronger presence in the US and Asia. If OpenStack maintains its global advantage the OpenStack community in Europe is very likely to start quickly as well.