Following the November 7 announcement of Rackspace’s tantalizingly named (and OpenStack-powered) Rackspace Cloud: Private Edition, I’ve found that I have descended into doing something I don’t normally do – getting all hot under the collar over what amounts to really nothing more than a clever positioning statement.
First, let me clear a few things up. I have no dog (left) in the OpenStack fight, nor do I really care to challenge Rackspace’s latest positioning as it relates to analyzing their strategy and what it may or may not mean. I am, and always have been, a believer in open source, a great fan of the OpenStack concept and have written about it many times in a positive light. However, as some of you may recall, I have always been vocal about the need for there to be more traditional enterprise customers (I have really gone past caring about SPs/MSPs/Telcos) to be involved in its development and today, with little evidence of that actually occurring, I remain convinced that the current OpenStack “product” is nothing more than a promising science project – making the announcement that Rackspace are now ready to fling open it’s genetically-modified doors to, gasp, the enterprise, more than a little bemusing.
It is a well known fact that Rackspace has poured a considerable amount of time and money into OpenStack and has by far and away (since the acquisition of Anso Labs) the largest percentage of contributors to the community. Hardly a surprise, you may conclude, considering they have bet their proverbial house on using OpenStack to power the next generation of their public cloud offerings, and with such potential revenue at stake in the great multi-billion dollar cloud-grab, they simply can not afford for it to fail – at any cost.
Is that why, perhaps, a peek behind the curtain may reveal a community governed by a benevolent dictator (BD) rather than one of a joyful meritocracy ? Is that a good thing ? Do we need a Cloud Spring ? Well, let me help you decide.
A BD project is led by a benevolent dictator and managed by the community. That is, the community actively contributes to the day-to-day maintenance of the project, but the general strategic line is drawn by the benevolent dictator. In case of disagreement, they have the last word. It is the benevolent dictator’s job to resolve disputes within the community and to ensure that the project is able to progress in a coordinated way. In turn, it is the community’s job to guide the decisions of the benevolent dictator through active engagement and contribution.
In the context of this post, this hypothesis alone gives me a very uneasy feeling. Anyone who has paid even the slightest passing interest in OpenStack’s progress will have witnessed a steady throng of the technology vendor hoi-polloi clamoring to contribute, desperate for a slice of the action and all in the name of the community. However, given the familiar nuances of the Apache License (where there is no requirement for any modified code, enhancement or additional special sauce to be returned to the community) and the fact that we are already seeing a growing number of both startups and large OEM vendors creating their own combinations of architectures, distributions and services on top of a frankly wobbly footing, I am beginning to wonder if this bright star may burn out before it has chance to really shine.
Despite what many of the regular twitter-contributing experts have said, I don’t believe that this trend in “calling everything Open” is represented in a level playing field. In fact, it’s positively lunar. In my extremely humble opinion, the parallels drawn between let’s say Facebook’s OpenCompute platform and OpenStack, are misleading and wholly inaccurate. They do not, for example, take into consideration the number of contributors (few vs many), the reasons (the real ones) for the many contributors involvement and last but not least, the revenue potentials of each. I remain to be convinced that there is a significant amount of money to be made from the release of the blueprint of efficient data center design and throwaway hardware – certainly not compared to why we’re already seeing a great jostling for position in the marketplace for the stuff that sits on top of it.
So, back to Rackspace for a moment. I actually think their “we can’t fail, so you can’t fail” approach is incredibly clever but it falls woefully short in the reality stakes and borders just a little on cloud fantasy. Other than at a few tech-savvy organizations, there has been little practical production deployment of OpenStack in large organizations and I can’t imagine (especially given my beta experiences of this offering which were, shall we say, less than mindblowing) that, and I quote, customers looking to build medium-to-large scale private clouds are going to be beating the door down to implement this from a reference architecture which, among other bizarre recommendations, suggests a hypervisor that:
a) is not one that most traditional enterprises will have any experience of
b) is not the best for running Windows workloads which make up the lion’s share of enterprise server installed base
c) is not the same as the one that Rackspace use today (so quite what expectation that sets for customers is a mystery)
Sobering to think that at the bottom of the cloud pile, unfortunately, sits the poor unassuming customer, increasingly perplexed by meaningless terminology and unwittingly drowning in choice. He dreams of openness and the idea of no lock in, yet the more I see the unwelcome combination of hidden agendas and greed-fueled fragmentation rearing their gorgonesque ugly head, the more I fear those dreams will never be realized.
Customers won’t buy what they don’t understand and perhaps my biggest, and until now most secret fear for OpenStack is that the fragmentation (not forking – people who say forking are either posh Brits swearing or people who don’t understand how open source projects work but like to think they do) reaches such a level that one can no longer be sure of the integrity of the OpenStack components themselves.
Can you imagine it ? “I am running Crackhouse’s version of OpenStack”…”ah yes, but I heard that’s not real OpenStack”. What a terrifying thought for the very compatibility OpenStack promised to do so much to solve.
So, my final word on this is somewhat rhetorical – overall and on balance, there have been relatively few major open source successes compared to those that offered promise and petered out weakly and cling to life as shadows of their once-promised greatness as ubergeek utilities wallowing in a sea of cool. I hope that in OpenStack we don’t end up missing a golden opportunity because the benevolent dictator was broadcasting when he should have been be tuning in.