Day 1, Keynote Address
The keynote given by General Hugh Shelton (Ret.) on Monday morning focused on leadership in contrast to so many examples of selfishness and unethical behavior to which we’ve witnessed recently in our world at large. I really liked the distinction that General Shelton drew between leadership and management. Certainly, the ability to coach, teach and mentor can have a major impact on communicating and disseminating culture in an organization. Also, and I’ve seen this in so many organizations of which I’ve been a part, leaders set the moral and ethical tone in any organization by leading from the front and setting an example. By following a few steps, leadership can be established:
1) Have and act with integrity…your word is your bond.
2) Lead from the front…set the right example.
3) Meet employee expectations and vision…establish an organizational, inclusive climate (culture) and create a team.
4) Set standards and communicate them…follow the Golden Rule.
5) Coach, teach and mentor…stand as a role model, provide nurturing feedback to those you lead.
6) Ensure that Teamwork relies on collaborative leadership.
The flow between General Shelton’s keynote address and the first presentation of the day couldn’t have been more perfect. The panel of data center superstars including Mike Manos (AOL), Jack Glass (Citi Technology Infrastructure), Joe Kava (Google) and moderated by Dave Ohara, Data Center Meme from GreenM3 (www.greenm3.com). The panel focused on “The Hunt for Talent” in the data center environment and IT landscape. What I found really fascinating was the direct application of General Shelton’s keynote to the hunt for talent. Furthermore, it was captivating to get a glimpse into each panelist’s perception of the quest for talent and the nature of the different environments over which they reign was captivating. Mike Manos leads the transition of AOL from Internet Portal to Media Company; he is focused on his current employees and support of that initiative. Jack Glass manages data center planning for Citi and Joe Kava leads the data center team that designs, builds and operates Google’s data centers. Amazingly, Google’s enterprise level data centers only started six years ago.
The most important attributes of prospective candidates for each varied as well. Manos stated that in the past, candidates needed specialized skill sets determined by specific discipline. Now, candidates need a blended skill set with much broader experience and a wider range over various silos. Glass posited that in the past, his employees needed an understanding of the power side of the equation, whereas now, there’s a greater focus on mechanical and systems engineering, which has drawn them to candidates from the Maritime academies. Kava took a different tack on the subject suggesting that at Google, it is more attractive to have a genuine curiosity about how systems interact and a general knowledge of various technologies (being an IT Geek) is what attracts them. Also, Kava noted that there is a distributed leadership model at Google. That really makes sense to me since people who take a leadership role typically strive proactively to find solutions rather than waiting for one to be handed to them. In terms of offering tools to manage, Kava suggested defining and continually reinforcing core values within a culture of continuous learning. Manos concurred that finding teachers is difficult and that most people are resistant to change. Certainly worth considering and remembering was Glass’ insight about the amount of data that we receive is overwhelming, that we analyze very little and take action on even less.
Mike Manos revisited the stage to present on the transformation of AOL that he is sheparding. He started with some keen observances:
1) We are living in the most disruptive period in data center history.
2) We are barraged by an enormous number of messages.
3) Worse, management is barraged by the same messages.
4) Data Centers are not all alike.
5) Why? Because of capital expenditures, regulations and information protection.
The AOL footprint is large: over 60,000 devices; 15MW of capacity with primary facilities in North America, EMEA; a Tier one Broadband dark fiber network that is undergoing a massive and rapid transformation from a technology company to a media company that doesn’t just offer dial-up any longer (though dial-up is still a large portion of its revenue). Manos supported his assumptions above by references to General Shelton’s keynote as well as the fact that most people are resistant to change and won’t bet on new technologies unless they have to, are told to, or they wait until it’s not new anymore.
Day 2, Keynote Address
Steve Fairfax, President of M Technology gave a riveting and somewhat controversial presentation entitled, “Hidden Threats to Data Center Reliability.” Steve’s perspective is so interesting because he doesn’t take anything for granted. This even translates to the psychology of our brains… how we’re wired. I think he’s right in stating that our brains respond to imminent threats such as animals, poisonous plants, etc., but that we also give priority to recent or frequent threats and take elaborate measures to avoid the threats that we’ve experienced. This lesson doesn’t always pay off and may force us into making decisions that create less than reliable systems. Steve suggested that although more reliable, 2N systems typically have twice the number of failures as N+1 systems. This man is a wealth of information about Diesel fuels, fuel stability and other factors that promote generator failures. Unlike the fireworks that erupted between Pitt Turner and Steve at the 2010 Fall Conference, there was only some mild disagreement, but all of the questions posed and answered were quite interesting.
Day 2, Presentation 5
Pitt Turner’s presentation on “Global Data Center Deployment" provided a very unique and broad view of what’s happening in the data center sector in other countries. With the exception of information that I’ve received from other enterprise users about developing data centers in other [third world] countries, this is really the first balanced and objective perspective that I’ve heard. Pitt’s opening remark was particularly poignant and accurate: “Data Centers have gone from real estate investments to financial investments.” And with all financial investments, “Past performance is not a guarantee of future success.” That’s really a brilliant way of connecting two statements that have become true. Even more important, Pitt asserted that the way that we’ve done business in the past won’t be enough to meet tomorrow’s business needs or demands.
I was amazed to learn that 50% of growth in our industry over the coming years will occur in the BRIC (Brazil, Russia, India & China) countries. When one stops to think about this for a moment, there should be no surprise as these countries are on an explosive population track and are beginning to join the rest of the world from a consumer perspective. Also interesting to note is that the global data center market is not homogeneous. Data centers differ in each location and market drivers differ as well.
Obviously, there are challenges in data center development in other countries that the North American markets have already overcome: Lack of Commissioning; Portable equipment; and Lack of craftsmanship top the list. In 2006, there were no purpose built data centers in Russia. Every data center development was being designed by vendors in converted office buildings. By 2009, consulting engineers were beginning to emerge and there were a number of projects in planning. By 2011, there were a number of projects underway and the first Tier III certified facility was completed. The bulk of current data center developments are hybrid projects consisting of half colo space and the other half enterprise space. There are still many concerns and considerations present including: public health; carbon footprint; reporting; privacy and data protection; location of processing; wider climatic temperature ranges; free cooling; voltage to IT devices; avoidance of cooling towers, etc. But seemingly the most important concerns are selecting sites in cooler climates where there is network connectivity; existing entitlements to power (more than 30% of projects must contend with this limitation); and adaptive reuse of existing buildings.
Day 3, Presentation 1
By far the most the most frightening presentation was given by Kevin Kealy, who postulated that control systems are the target for most modern day hackers. The first example was a thorough discussion and description of the Stuxnet virus that ravaged the Iranian nuclear enrichment program computer software. Although I have read several articles about the virus previously, I had no idea that it had been active for nearly a year before it was even detected in June 2009. This highly sophisticated effectively weaponized malware successfully targeted a particular control system of nuclear production centrifuges, increased the speed of the centrifuges, sent messages to the operator of the control system that in essence were incorrect that ruined the computer program and the materials. The software used zero-day vulnerabilities that are computationally extremely expensive to program and, to this day, we still don’t know what else it can do. This virus even had its own malware anti-virus program built in and destroyed itself 40 days after it entered the system and became active.
Scarier still, on October 14, 2011, security experts detected what is now being termed, “Son of Stuxnet”, DUQU. The primary mission of this new virus is to search for information related to control systems. We have no idea who has disseminated the virus, or what their intent. Further, terminates itself 36 days after insertion. This is based on the same code as Stuxnet and is a Random Access Trojan (RAT).
All control systems are targets, not just military control systems. Of particular interest is civilian infrastructure as disruptions in systems that control dams, power grids, water barriers, cooling systems, generators, nuclear power plants and shunting yards are critical to our country and economy.
Kealy cited an article from October 10, 2011 about the control system of a UAV being infected with a key logger virus that prevented the military from controlling a drone aircraft. Most of these control systems are operating on Windows operating systems.
As mentioned above, civilian infrastructure is primary and that has also attracted the attention of the military, which has placed military personnel, at no charge to the operators, within such enterprises under a program called Training with Industry (TWI). The intention is to understand how the control systems work, harden security related to operations and run predictive exercises to avoid outages in the case of penetration.
Kealy made some terrific observations and suggestions as to how to avoid a security breach:
1) Hackers are lazy, except for those that are termed advanced persistent threats.
2) We are living in a target-rich environment for hackers.
3) Control systems are the next big target to be hacked.
4) Cyberwar is no longer hyperbole.
5) Most people don’t bother to secure their systems or infrastructure adequately (or at all).
6) Don’t be like most people, install and use a firewall (demilitarized zone).
7) Include control systems in the rest of your security protocol and adhere to PCI DSS standards.
Day 3, Presentation 3
The final session that I was able to attend was given by Darryl S. Brown and Robert DeVita of TelX. Entitled “Mobility and the Hybrid Cloud”, they quoted some very interesting statistics including the projections that iCloud will funnel approximately $750M in annual revenue into Apple’s pipeline, that 70% of the monthly revenue will be shared by the creators and copyright owners of the music that is downloaded and that there this will attract over 150,000,000 users. Personally, based on my prior career in the music industry, I really appreciate those projections and hope that they become a reality. Additionally, Mr. Brown and Mr. DaVita also stated that the first four of the primary drivers for cloud adoption are: technology; mobile applications; social communications, collaboration and media; and, video streaming.
Accompanying every discussion that I’ve heard about the Cloud, rapid deployment time and low capital expense related to deployment are always stressed. That’s great news, but the elephant in the room that everyone continues to ignore is the fact that unless one is willing to migrate to the public cloud, the monthly recurring costs associated with private cloud adoption and usage are substantial and, in many cases, can be a multiple of colocation costs for equivalent technology.
By all measures, this was a great conference. I spoke with many attendees that thoroughly enjoyed the conference for the information they were able to glean, the contacts they made, relationships they reinforced and experiences that they had.