This was my second time attending the Symposium in Santa Clara California. The event took place from May 14 through May 17, 2012. The theme of the conference was Digital Infrastructure Convergence with a significant emphasis on modular data center design and construction.
Day 1 - Highlights
Notwithstanding the “advertorial” aspect of the presentation given by Dean Nelson of eBay, he did quote some incredible statistics that are definitely worth repeating.
- In 2009, there were 5 Million downloads of the eBay mobile application. In just the first part of 2012, there have been 80 Million downloads.
- In 2009, the value of the trades on eBay was $600 Million. In just the first part of 2012, the value of the trades has exceeded $8 Billion.
- Due to online and mobile shopping, retailers are being disintermediated. Evidence for this is the demise of Borders and Circuit City, which illustrates how retailers are becoming showrooms for online shopping.
- Just as the axioms penned by Intel’s Gordon Moore and David House, John Donahoe, eBay’s CEO, recently noted that ‘We will see more change in the next three years than we’ve seen in the past 15 years.’
- eBay has accomplished what most companies are still trying to achieve: an alignment of IT and Facilities. This is leading the company to use Total Cost of Ownership as a decision-making tool and PUE as a design driver.
- Using their Data Center Matrix model, eBay is achieving some amazing PUE numbers averaging 1.043 from their Project Mercury container solution, especially considering the 119° F temperature on the day the measurements were taken.
Somewhat underscoring Dean Nelson’s presentation, George Slessman, CEO of i/o, again predicted that the last “snowflake” data center (meaning of unique “stick” design) will be built no later than in 2013. This is understandable based on i/o’s laser-focus on the i/o container product, but still not credible when one considers number of enterprises, developers, colo service providers and vendors that all have their own unique ideas about data center design. He also postulated that IT is exponential.
This seems almost self-evident. The storm of cloud-based services and storage needs for data, the proliferation of mobile devices and ensuing never-ending creation of applications, the global IT footprint will, at least until there is a manufacturing and computing-based design, continue to grow. Data centers and the need for power will continue to expend as well. One of the fascinating statistics quoted was that between 2005 and 2010, worldwide data center power usage increased by 56% to 17GW. CPU computing power has increased every year and the cost has been steadily reducing every year while data center construction costs have been static. Support for this statement was the assertion that data center infrastructure is the fastest growing cost in deploying any IT landscape. Of course, this line of argument led to the conclusion that containerized data centers are the future of the business.
Day 2 – Highlights
Matt Stansberry, Uptime Institute Director of Content and Publications, Program Director of Symposium 2012, gave a great overview of cloud adoption statistics. Using data from Uptime’s survey of 1,100 data center operators, we learned that:
- 8% of respondents have upgraded their data centers in the past 5 years;
- 69% have implemented private cloud services;
- 31% have implemented public cloud services;
- 54% have used SaaS; and
- 5% have used colocation services.
Furthermore, over the next 12 months:
- 49% plan to utilize private cloud services;
- 37% are considering using private cloud services;
- 25% plan to utilize public cloud services;
- 30% are considering using public cloud services; and
- 82% are considering implementing energy conservation strategies for financial reasons.
I continue to meet with clients that are considering a migration of a part or all of their IT landscape to the cloud until they understand the monthly recurring charges that accompany a cloud-strategy. While recently assisting a client in evaluating options for a 500TB storage requirement, the pricing we discovered from a variety of providers is all over the map, ranging from $0.04 to $0.20 per GB. In some instances, considering all of the additional charges for on/off boarding and transport, the differential is even more extreme. I’m sure as with so many other solutions in the data center and IT fields, competition leads to commoditization, which will in turn lead to higher adoption rates and quicker pacing.