Back to Basics
And, as I examine the features and columns that comprise this issue, I find that we have assembled a rather glitz-free issue, aside from just a few photos. In fact, only the cover story itself really purports to take a high-level look at how we ought to run data centers.
The other stories, by contrast, take a look at the business case for reliability (p. 64), managing air flow (p. 44 and 54), and DCIM (p. 60). These are all issues that go to the heart of controlling costs and monitoring data center operations.
This is not to say that we are describing only low-hanging fruit. DCIM, in particular, can be an expensive proposition on a first-cost basis. Even monitoring PUE can be costly if legacy equipment must be upgraded to gather data.
Some time ago, I moderated a webinar during which the Uptime Institute’s Julian Kudritzki spoke about the intent of the Tier system and how it was inappropriate to label a Tier IV center as better than a Tier III. Certainly, he said, a Tier IV center should have greater availability than a Tier III facility, but it was more important that the facility meet the business purpose of the enterprise that owns it. Bad marketing had led some to misunderstand the intent of the Tier system.
I liked what Julian had to say because it implied that someone cared about the purpose of the data center and that building a right-sized data center (or wrong Tiered) data center for an enterprise would be a waste of money and resources.
In a world where bigger is better, we often mistake bigger for better and sometimes more expensive for better. But, if we have learned anything in recent years, it is that we can no longer afford to paper over inefficiencies by spending more money or drawing more energy.
PUE is a way to measure what we are doing, and DCIM is a way to control what we are doing, and managing air flow is a way to minimize the energy we use.