There are many in the data center industry that have repeatedly called for change in this community of ours. Change in technology, change in priorities, Change for the future. Over the years we have seen those changes come very slowly and while they are starting to move a little faster now, (primarily due to the economic conditions and scrutiny over budgets more-so than a desire to evolve our space) our industry still faces challenges and resistance to forward progress. There are lots of great ideas, lots of forward thinking, but moving this work to execution and educating business leaders as well as data center professionals to break away from those old stand by accepted norms has not gone well.
That is why I am extremely happy to announce my involvement with the University of Missouri in the launch of a Not-For-Profit Data Center specific organization. You might have read the formal announcement by Dave Ohara who launched the news via his industry website, GreenM3. Dave is another of of those industry insiders who has long been perplexed by the lack of movement and initiative we have had on some great ideas and stand outs doing great work. More importantly, it doesn’t stop there. We have been able to put together quite a team of industry heavy-weights to get involved in this effort. Those announcements are forthcoming, and when they do, I think you will get a sense of the type of sea-change this effort could potentially have.
One of the largest challenges we have with regards to data centers is education. Those of you who follow my blog know that I believe that some engineering and construction firms are incented ‘not to change’ or implementing new approaches. The cover of complexity allows customers to remain in the dark while innovation is stifled. Those forces who desire to maintain an aura of black box complexity around this space and repeatedly speak to the arcane arts of building out data center facilities have been at this a long time. To them, the interplay of systems requiring one-off monumental temples to technology on every single build is the norm. Its how you maximize profit, and keep yourself in a profitable position.
When I discussed this idea briefly with a close industry friend, his first question naturally revolved around how this work would compete with that of the Green Grid, or Uptime Institute, Data Center Pulse, or the other competing industry groups. Essentially was this going to be yet another competing though-leadership organization. The very specific answer to this is no, absolutely not.
These groups have been out espousing best practices for years. They have embraced different technologies, they have tried to educate the industry. They have been pushing for change (for the most part). They do a great job of highlighting the challenges we face, but for the most part have waited around for universal good will and monetary pressures to make them happen. It dawned on us that there was another way. You need to ensure that you build something that gains mindshare, that gets the business leadership attention, that causes a paradigm shift. As we put the pieces together we realized that the solution had to be credible, technical, and above all have a business case around it. It seemed to us the parallels to the Open Source movement and the applicability of the approach were a perfect match.
To be clear, this Open Source Data Center Initiative is focused around execution. Its focused around putting together an open and free engineering framework upon which data center designs, technologies, and the like can be quickly put together and more-over standardize the approaches that both end-users and engineering firms approach the data center industry.
Imagine if you will a base framework upon which engineering firms, or even individual engineers can propose technologies and designs, specific solution vendors could pitch technologies for inclusion and highlight their effectiveness, more over than all of that it will remove much mystery behind the work that happens in designing facilities and normalize conversations.
If you think of the Linux movement, and all of those who actively participate in submitting enhancements, features, even pulling together specific build packages for distribution, one could even see such things emerging in the data center engineering realm. In fact with the myriad of emerging technologies assisting in more energy efficiency, greater densities, differences in approach to economization (air or water), use of containers or non use of containers, its easy to see the potential for this component based design.
One might think that we are effectively trying to put formal engineering firms out of business with this kind of work. I would argue that this is definitely not the case. While it may have the effect of removing some of the extra-profit that results from the current ‘complexity’ factor, this initiative should specifically drive common requirements, and lead to better educated customers, drive specific standards, and result in real world testing and data from the manufacturing community. Plus, as anyone knows who has ever actually built a data center, the devil is in the localization and details. Plus as this is an open-source initiative we will not be formally signing the drawings from a professional engineering perspective.
Manufacturers could submit their technologies, sample application of their solutions, and have those designs plugged into a ‘package’ or ‘RPM’ if I could steal a term from the Redhat Linux nomenclature. Moreover, we will be able to start driving true visibility of costs both upfront and operating and associate those costs with the set designs with differences and trending from regions around the world. If its successful, it could be a very good thing.
We are not naive about this however. We certainly expect there to be some resistance to this approach out there and in fact some outright negativity from those firms that make the most of the black box complexity components.
We will have more information on the approach and what it is we are trying to accomplish very soon.