Eucalyptus announced that they will be consolidating the different flavors of their cloud management solution into one open source edition. A lot of outlets including this blog have given plenty of attention to OpenStack which is the flavor of the day for Cloud Management tools. Eucalyptus has been around for several years and is currently on their 3.1 release looking to expand their level of integration with Amazon AWS. Eucalyptus is far and above further along in maturity level vs. OpenStack. However, Eucalyptus up to this point has had used an open source/closed source approach to their product offering. With the release of 3.1 they have completely embraced open source with the exception of their plugins. I made the comment on the above referenced Gigaom article that the value of Eucalyptus is in their integration with Amazon and their hypervisor management plugins. The author, Barb Darrow, commented that the announcement is significant as “anyone can develop plugins for eucalyptus so if you don’t want to use theirs, you can BYO…”
I question if the enterprise customer actually desires to make the investment to bring their own plugins. I’ve done infrastructure management for quite some time and there are very few organizations that understand the complexity of cloud infrastructures and cloud management tools. Virtualization has just become main stream so, cloud management is not a topic easily decomposed to system requirements that can then be developed in an enterprise software development shop. It is a very steep learning curve to start to uncover the requirements for running your own cloud let alone write the code to extend functionality of cloud managers. The skill needed to develop or extend the capability of these platforms is a very specialized skill that doesn’t exist in your normal enterprise or the market at large. I can see the Zynga’s, Netflix’ and Navisite’s of the world getting excited about having access to the code base for solutions such as OpenStack and Eucalyptus but, what about the McDonald’s and Allstate Insurances of the world? How can they leverage the advantages of an open source cloud manager platform?
OpenStack is clearly positioning themselves as a service provider/OEM solution to this point but Eucalyptus is much closer to an enterprise focused solution. I like to equate the complexity of cloud managers to ERP systems. They can do whatever you want them to do just as long as you have the expertise to make it happen. This is why the Eucalyptus plugins are so critical to the traditional enterprise. Without the plugin, organizations are left with just 70% of the solution and need to develop the other 30%. My experience has been that organizations are looking for a turnkey solution that currently doesn’t exist (even from VMware). So, I would infer offering the source code to your cloud manager really doesn’t appeal to most enterprises. At the most enterprises are interested in the API’s to these systems so that they can be integrated with other infrastructure management solutions. As for heavy customization, it’s more a feature to be used by technology organizations and service providers.
If you are not securing revenue from cloud services or building an infrastructure from the ground up, how do you leverage the source code of these open source solutions in the traditional enterprise?