OpenStack (Photo credit: Wikipedia)
Last week at OSCON (CloudAve Coverage), OpenStack project (previous CloudAve coverage) celebrated its second birthday with much fanfare. Even though I missed the OpenStack day at the conference, I did get a chance to talk to OpenStack team, developers, practitioners, well wishers, etc. during the event. I think it is time to do a reality check on the project and, for a moment, I am taking out my open source evangelist hat. Before I talk about the good, the bad and the ugly, I want to wish a very happy birthday to OpenStack and I also hope that the project thrives big so that the future of infrastructure market is a federated cloud ecosystem (something I have been advocating for years now).
The Good
- Excitement about the project is unabated. I could see this excitement on the developers side and also on the consultants side. OpenStack marketing could be a major factor in sustaining this excitement but let us not underestimate the open source developers and they don’t fall way too easily for marketing campaigns. I would like to put a * here because I am going to touch this point on the ugly side too.
- The story about an university in the State of Oregon using OpenStack to build a cloud with over thousand nodes. One of the biggest criticism about OpenStack in the past was on scaling and this case study goes on to show that it is possible to trust OpenStack cloud for important academic research projects
- The support for ARM architecture is very interesting and it is going to keep OpenStack relevant in the future evolution of cloud computing. In 2009, I wrote a short report about P2P cloud and suggested it as an alternative model for cloud so that we avoid the disruptions that were (are) plaguing the public clouds. Dave Asprey from Trend Micro is also a big advocate of mobile cloud. The support for ARM architecture is interesting in that context.
- The process for OpenStack Foundation is complete. Now anyone can join the foundation. It is not perfect but it is a start. Marten Mickos of Eucalyptus calls OpenStack Soviet Union. I like to see it more like European Union, a noble idea, some execution issues and future unknown.
- Rackspace is moving their cloud to Nova compute. If the second largest cloud provider trust Nova for their cloud, I would say the code is maturing faster than what many were expecting
The Bad
- The last point in the previous section is the first point in this section. Concerns about reliability and scalability still persists. I had a chance to speak with someone from the service provider side and they are still apprehensive about OpenStack’s reliability issue
- This is something I need to dig in more to understand but I had superficial chat with Jonathan Bryce on it. Even though he tried to convince me that the Release Management committee is independent of powerful interests and aligned with the technical goals of the project, I am still uneasy with the fact that it is dominated by big vendors. Hopefully, I get to understand this better in the coming months.
The Ugly
- Yesterday, we publicly came to know about how OpenStack developers from the original Anso Labs team are quitting Rackspace to Nebula, it raises some troubling questions. OpenStackers dismiss this news by saying that they are going to another OpenStack company and hence it doesn’t affect the project in any way. Their argument is that if at all it has any impact, it will be on Rackspace and it could be a problem with the culture there. I may agree with this assessment but though my interactions with various stakeholders, I have found that there is still a criticism among the developer community that most of the money is spent on marketing efforts than engineering efforts. I don’t discount the importance of marketing even for an open source project but any implication that engineering problems can be swept under the carpet purely through their marketing muscle is dangerous for the project in the long run. We need more openness in the transactions than in the past. Secrecy is not in the DNA of any open source project and OpenStack can be no different. I think having a Linus Torvalds, a guy who only focusses on the long term success of the Linux project and doesn’t hesitate to say no to anyone on anything if he sees it as detrimental to the project, for OpenStack will be in the best interests of the project.
Now that we have done the reality check, I am going to wear my open source evangelist hat and sing like Steve Ballmer “OpenStack, OpenStack, OpenStack, ………”.
Hey, ya know, maybe Linus is interested in adding to his project portfolio of awesome? 🙂
I’m honestly confused by all the “concern” over the departures of some of the former Anso Lab crew. Think about it guys…they were part of an acquisition that I’m sure not everyone was thrilled about, they were not located at Rackspace HQ and therefore not immersed in the culture as most Rackspace employees and they were being wooed with offers from a former boss/colleague who can offer some lucrative benefits at this stage in his company.
It’s not hard to see why this happened at all and I honestly don’t think it has anything to do with Rackspace directly or their culture. Nor do I think this will have any material impact on Rackspace. These guys were there for a relatively short period of time and likely had minimal impact on any internal projects within this timeframe (if any due to their focus on OpenStack directly).
I’m not saying these aren’t great developers but who couldn’t see this coming a mile away? Rackspace has some sharp talent, this isn’t going to change anything for them and these folks will still be involved with OpenStack.
Just my 2 cents as a former Racker looking in from the outside at this…
Thanks Chad for your perspective on it.