We wrapped our first full day of talks here at DataCentre2012 and I have to say the content was incredibly good. A couple of the key highlights that really stuck out in my mind were the talk given by Christian Belady who covered some interesting bits of the Microsoft Data Center Strategy moving forward. Of course I have a personal interest in that program having been there for Generation1 through Generation4 of the evolutions of the program. Christian covered some of the technology trends that they are incorporating into their Generation 5 facilities. It was some very interesting stuff and he went into deeper detail than I have heard so far around the concept of co-generation of power at data center locations. While I personally have some doubts about the all-in costs and immediacy of its applicability it was great to see some deep meaningful thought and differentiation out of the Microsoft program. He also went into a some interesting “future” visions which talked about data being the next energy source. While he took this concept to an entirely new level I do feel he is directionally correct. His correlations between the delivery of “data” in a utility model rang very true to me as I have long preached about the fact that we are at the dawning of the Information Utility for over 5 years.
Another fascinating talk came from Oliver J Jones of a company called Chayora. Few people and companies really understand the complexities and idiosyncrasies of doing business let alone dealing with the development and deployment of large scale infrastructure there. The presentation done by Mr. Jones was incredibly well done. Articulating the size, opportunity, and challenges of working in China through the lens of the data center market he nimbly worked in the benefits of working with a company with this kind of expertise. It was a great way to quietly sell Chayora’s value proposition and looking around the room I could tell the room was enthralled. His thoughts and data points had me thinking and running through scenarios all day long. Having been to many infrastructure conferences and seeing hundreds if not thousands of presentations, anyone who can capture that much of my mindshare for the day is a clear winner.
Tom Furlong and Jay Park of Facebook gave a great talk on OCP with a great focus on their new facility in Sweden. They also talked a bit about their other facilities in Prineville and North Carolina as well. With Furlong taking the Mechanical innovations and Park going through the electrical it was a great talk to created lots of interesting questions. An incredibly captivating portion of the talk was around calculating data center availability. In all honesty it was the first time I had ever seen this topic taken head on at a data center conference. In my experience, like PUE, Availability calculations can fall under the spell of marketing departments who truly don’t understand that there SHOULD be real math behind the calculation. There were two interesting take aways for me. The first was just how impactful this portion of the talk had on the room in general. There was an incredible amount of people taking notes as Jay Park went through the equation and way to think about it. It led me to my second revelation – There are large parts of our industry who don’t know how to do this. In private conversations after their talk some people confided that had never truly understood how to calculate this. It was an interesting wake-up call for me to ensure I covered the basics even in my own talks.
After the Facebook talk it was time for me to mount the stage for Global Thought Leadership Panel. I was joined on stage by some great industry thinkers including Christian Belady of Microsoft, Len Bosack (founder of Cisco Systems) now CEO XKL Systems, Jack Tison-CTO of Panduit, Kfir Godrich-VP and Chief Technologist at HP, John Corcoran-Executive Chairman of Global Switch, and Paul-Francois Cattier-Global VP of Data Centers at Schneider Electric. That’s a lot of people and brainpower to fit on a single stage. We really needed three times the amount of time allotted for this panel, but that is the way these things go. Perhaps one of the most interesting recurring themes from question to question was the general agreement that at the end of the day – great technology means nothing without the will do something different. There was an interesting debate on the differences between enterprise users and large scale users like Microsoft, Google, Facebook, Amazon, and AOL. I was quite chagrined and a little proud to hear AOL named in that list of luminaries (it wasn’t me who brought it up). But I was quick to point out that AOL is a bit different in that it has been around for 30 years and our challenges are EXACTLY like Enterprise data center environments. More on that tomorrow in my keynote I guess.
All in all, it was a good day – there were lots of moments of brilliance in the panel discussions throughout the day. One regret I have was on the panel regarding DCIM. They ran out of time for questions from the audience which was unfortunate. People continue to confuse DCIM as BMS version 2.0 and really miss capturing the work and soft costs, let alone the ongoing commitment to the effort once started. Additionally there is the question of once you have mountains of collected data, what do you do with that. I had a bunch of questions on this topic for the panel, including if any of the major manufacturers were thinking about building a decision engine over the data collection. To me it’s a natural outgrowth and next phase of DCIM. The one case study they discussed was InterXion. It was a great effort but I think in the end maintained the confusion around a BMS with a web interface versus true Facilities and IT integration. Another panel on Modularization got some really lively discussion on feature/functionality and differentiation, and lack of adoption. To a real degree it highlighted an interesting gulf between manufacturers (mostly represented by the panel) who need to differentiate their products and the users who require vendor interoperability of the solution space. It probably doesn’t help to have Microsoft or myself in the audience when it comes to discussions around modular capacity. On to tomorrow!