We recently annoyed Peder Ulander of Cloud.com by suggesting when Cloud.com joined OpenStack it was a Turkey waiting for Thanksgiving. It wasn’t personal, but we do understand that being compared to a fat bird with a big neck can cause offense. To calm things down we spoke to Peder, and we thought Virtualization Practice readers might be interested in the conversation.
Mike First, to clear the air a bit I think we’d better focus on a few things we agree on. I believe OpenStack is a very important initiative that offers the real possibility to develop a widely-adopted non-proprietary infrastructure for both public and private clouds, and is more likely to deliver effective interoperability through a common implementation than the current slow and fragmented standards initiative which seek only to provide a common API.
Peder I completely agree. I don’t necessarily think that innovation has been slow in the cloud space, it just hasn’t had the combined focus and collaboration that OpenStack brings to the table and, as such, has been somewhat fragmented. I was talking to a customer the other day who made the comment that it seemed that before this the big guys (AMZN, VMware and MSFT) could take their time in getting their cloud strategies together – specifically pointing out the delay with vCloud and Azure because the efforts were being driven by smaller players in the space.
With our initial emergence from stealth mode, coupled with our participation with the OpenStack program we have seen significant traction with customers that have become frustrated with the wait, lack of transparency and limited innovation from the big guys and are looking for companies just like Cloud.com. Today, the collaboration and promotion by a number of leaders in the open source cloud space specifically focused on bring together and advancing an entire cloud ecosystem (not just an API as you pointed out) based on open standards and driving interoperability has put significant pressure on the old, proprietary regime and pressing them to rethink their strategies. Innovation will continue to be fast and exciting in the cloud space, but now the added wrinkle of openness will help benefit customers significantly.
Mike I think one of the most important aspects of the initiative is that it is multi-vendor from day 1, it does not use the GPL license and therefore offers a broad range of commercialization options
Peder I know there are a bunch of discussions happening around open core vs open source, GPL v LGPL v AGPL v Apache, etc. I think they somewhat distract from the effort. Seriously, it is like a group of democrats talking with each other around who is more left. The reality is that the licensing model that the OpenStack team chose, Apache, is one of the most open and developer friendly licenses in the open source community. It offers a significant amount of room for companies and developers to not only contribute, but also commercialize the technology to provide additional value to their customers. Highlighting the success of their choice in Apache is the rapid growth of the community from individual developers to established companies… like Cloud.com.
I’d also like to point out our license as you called us out a little bit in your post. While I agree with you that there are a number of vendors using an open core model to deliver a crippled open source product and forcing users to their enterprise model, we don’t believe in that approach. As such, we have delivered about 99% of our code in open source and actually have our developers building real time in our community repo (we are nor managing two). We will be using this effort as an integration point with the OpenStack community as well … developing “in” the community is just as important as developing “with” the community.
Our GPL Community Edition is pretty much identical to our SP and EE editions. The “closed” elements of our code are a result of 3rd party integration with proprietary hardware (not mine to open source) and our metering module. The metering module aggregates all activities and spits our a file that can integrate easily into a billing or chargeback engine. Since there is money involved here, we feel that this is a great point of value add and differentiation. Other than that, the community product is fully deployable in either a test or a production environment (you just have to support it yourself)
Mike Yes, I am familiar with the problem you mention with proprietary drivers, and Billing Integration (for real money in external clouds and for chargeback in internal clouds) is a very important area, one where there is definitely a rationale for a proprietary solutions element. What were Cloud.com’s overall commercial objectives in joining the initiative and how do you think it benefits your customers?
Peder We were approached by the OpenStack team in early summer around their efforts in building and launching an open source movement focused on bringing together customers, developers and technologists to drive open standards and interoperability in the IaaS cloud computing space. At that time, it was nothing more than a broad vision but we could smell the makings of the next Eclipse happening and felt that it was absolutely important to participate. As we got into more detail we saw a number of areas to engage on a commercial level that would benefit both Cloud.com and our customers.
First, from a technology point of view, we saw the opportunity to fill a product gap we had in the S3 market by adopting the Open Block Storage project into our commercial product to extend our solution to include both compute and storage in the cloud service. We are also going to be adopting the API schema and tools approach to deliver a truly interoperable cloud environment – by that I mean the ability to create, manage and run applications that support APIs from clouds like Amazon, Rackspace, vCloud, Citrix, etc.
While we have done a lot of work in this area and deliver it today in our product, we will be contributing this to the OpenStack initiative and the end benefit is an acceleration of innovation and implementation to truly get to an interoperable cloud. Both of these have clear benefits to our normal commercial business as well as our customers. Extending beyond that, as a not-for-profit foundation, similar to Eclipse, enterprises and SPs looking to build clouds from this effort will require support and customization for their specific environments and will not be getting this from the OpenStack project. As a key commercial vendor for building and delivering an IaaS stack, Cloud.com is in a very prime position as one of the only orchestration players in the effort and have already seen a positive influx of business as a result of our participation.
Mike One of the crucial issues, however, is how organizations can transition their business model from their position at the point they engage with the Ecosystem through to the point when the ecosystem matures. I remember the carnage in the development tools market starting in 2002 when Eclipse started to dominate: Borland acquired WebGain who had acquired TogetherSoft. All of these organizations attempted to engage with Eclipse but failed and ultimately the consolidated entity was sold to Embarcadero in 2008 for a fraction of its 2002 value. I understand that Cloud.com is nimble enough to change strategy, but how do you intend to transition your business model to surf the wave rather than to get drowned?
Peder Are you calling me a drowning turkey now Mike? We are nimble and in a position of influence as this effort gets under way. There are a couple of ways companies can engage in projects like an OpenStack.
- You can engage, innovate and influence with your code and capabilities – ultimately helping the success of the project.
- You can participate, but ultimately take the code in with little modification and serve as an integrator.
One is in a leadership position with strong differentiation, the other is in the followers seat. If you are in the first bucket, even as the technology matures, you are constantly pushing and innovating in the market and in a power position with customers. Customers like working with leaders.
Aside from Eclipse, we can draw parallel to RedHat and their efforts around Linux. They didn’t own the kernel or many of components of the distribution, but innovated where it mattered around package management, application ecosystem, testing and support. As one of the main contributors to the effort, customers wanted to engage with them vs. go with other partners that were creating, in essence, an built linux package. So, back to the question, how do I soar with the eagles vs. drown with the turkeys? Time will tell, but focusing on adding value and innovation to the project and identify areas where we can differentiate will help us stay above the fray.
Mike Now you are engaged with the initiative, where do you see it going in the medium term, and what would you see as the key success factors?
Peder Short/medium term there is a lot of work going towards getting the project scoped, organized and executing. The launch response was amazing and the contribution level is a testament to how important this effort is. We need to make sure we are all continuing on the momentum and excitement, while delivering builds and components that fulfill the promise of a build-able, manageable cloud environment. The key success factors will be a direct result of the activities – meaning continued success with contributors and developers followed by actual implementation with customers in real life.
Mike How is cloud.com actively engaging to ensure those objectives are met?
Peder We are constantly engaged with the OpenStack group. We have regular calls and meet-ups (believe it or not, in person on a monthly basis with massive participation from the members) and are really pushing this forward. Part of that includes contributing to the scope and governance, part of that is having my engineers actively engaged in contributing and collaborating with the group. We are also actively hiring for OpenStack developers (http://cloud.com/main/main/company/jobs/foss-developer), so if you know of any please send them our way.
Mike Thanks, Peder. It has been good to talk to you, and we look forward to covering OpenStack and Cloud.com over the coming months.