Data Center Energy Blog

Data Center Energy Blog

The Viridity Data Center Energy Blog is a forum for the latest news, trends and best practices, in data center energy management and efficiency.

describe the image

Subscribe via E-mail

Your email:

Current Articles | RSS Feed RSS Feed

Data Center Efficiency Is More Practical Than Altruistic

  | Share on Twitter Twitter | Share on Facebook Facebook | Submit to Digg digg it |  Add to delicious  delicious |  Submit to StumbleUpon StumbleUpon |  Share on LinkedIn LinkedIn | Submit to Reddit reddit 

describe the imageViridity Software is a green technology, so you might imagine that the inspiration for our company was to help data centers be environmentally correct. And you’d be right; that was part of it. At Viridity Software, we try to be environmentally correct. (You will find a few SUVs in our employee parking lot, and I have seen the cleaning crew use bleach to clean the bathrooms.) But truth be told, the inspiration for us at Viridity was more practical than altruistic. Green tech is often labeled as expensive--its implementation difficult and inconsistent with existing infrastructure.

Our path couldn’t be more different. We’re all for the greater good. But our motivation with Viridity has really always been about solving business problems—specifically about data center efficiency. 

Prior to co-founding Viridity Software, I was hired as a consultant for a software company that worked closely with the data centers of large Wall Street companies. I was actually working on a storage project in a data center; we really had nothing to do with data center efficiency or energy resource management.

Still, you couldn’t help but notice the chaos running amuck at these organizations. For the guys in IT, it was all about business applications. On the facilities’ side, it was all about where to put them. Too often, IT had no idea of the strain inflicted on the data center with the addition of new computers and applications. At the same time, data center operators had only vague notions about available remaining energy at their sites. They couldn’t really tell you how close they were to running out of power. Sometimes, the whole enterprise seemed to be hanging by a thread.

Ultimately, all that inefficiency—coupled with communication breakdowns—has to catch up with a business. And it did. At one of those very large Wall Street data centers, the IT department put in a requisition for $25 million in new storage arrays—a significant investment, but certainly worth it from a business standpoint. So the requisition was approved, and the equipment was ordered, purchased, and delivered.  It wasn’t until after the arrays were actually installed on the data center floor that someone said, “You know, you can’t actually plug those puppies in, there isn’t enough power….”  

Data and business systems were clearly the dominant concern here--energy less so. Until they ran out of power. No one checked with the data center, which already ran beyond capacity. That $25 million investment sat there on the data center floor, quietly not doing anything useful but annoy people, indefinitely. Rockin’ good use of dough.

Suddenly, energy resource management became a huge concern at the site. To make room for the new arrays, our group switched focus and spent the next six months chasing down and analyzing applications and hardware from an energy-demand standpoint, trying to figure out what applications we could move to other data centers to make room on the grid for the arrays. Unfortunately, we never made much headway. The data center maintained such a tangled collection of applications and servers and racks from different providers that the discovery process was next to impossible. We left before the job was finished. I wonder sometimes if those arrays were ever powered on ;-)  

What kills me is that this is not some isolated situation. It’s common. When I talk to clients and analysts about Viridity, I always bring up that story. It’s such a vivid anecdote about the need for energy resource management tools, and the connection between IT and facilities that it demands. And every time I bring it up, the response from customers is the same. I get a familiar nod of recognition. It seems data centers everywhere face the same problem. They need tools to make better use of the energy they have and to make informed business decisions. 

Green is the gravy.  

-- Michael Rowan, Co-founder and CTO, Viridity Software

It Was a Big Week for Data Center Energy News

  | Share on Twitter Twitter | Share on Facebook Facebook | Submit to Digg digg it |  Add to delicious  delicious |  Submit to StumbleUpon StumbleUpon |  Share on LinkedIn LinkedIn | Submit to Reddit reddit 

There was a lot of interesting news on data center energy this week.  Here's a look at some of the articles that caught our attention.

DatacenterDynamics ran a nice piece on how the rising cost and demand for energy in the data center combined with impending government legislation means energy efficiency metrics are taking hold and evolving at a rapid pace.

SearchDataCenter News ran an article by Robert McFarlane on the new EnergyStar ratings for data centers.  In the article, he points out that the new Energy Star rating will have its detractors because it is based on the already controversial PUE (power usage efficiency) which is becoming the global standard for measuring data center efficiency.  He points out, the metric is "not valid as a comparison among different data centers" and that the problem with PUE is that it looks only at how much energy is used and not how that energy usage is contributing the business. 

Melissa Alvarez blogged about The Three Biggest Myths About Green IT. The first myth is that the case for Green IT is a clear one.  She writes that "the biggest problem in making the case for energy efficiency has nothing to do with the technology. The issue is who realizes the benefits."  The second myth is that Green IT is achievable.  Her thinking here is that Green IT isn't so much a destination as a journey that involves creating processes and policies that define a way of operating over the long-term.  And the third myth is that everybody cares about Green IT.  She quotes the results of a recent survey from CIO magazine that asked IT management their top priorities where Green IT didn't make the list.  Her thinking is that "Green IT is really about doing away with waste and inefficiency -- and inefficiency costs money."

At Viridity Software, we believe that you can't manage what you can't measure.  It's great to read that energy efficiency metrics are taking hold across the industry and that people recognize that reducing data center energy usage benefits the whole organization.

Top 7 Reasons to Remove Orphan Servers from the Data Center Floor

  | Share on Twitter Twitter | Share on Facebook Facebook | Submit to Digg digg it |  Add to delicious  delicious |  Submit to StumbleUpon StumbleUpon |  Share on LinkedIn LinkedIn | Submit to Reddit reddit 

If a business could identify the servers doing productive work and turn off its non-productive ones, it could save 20 to 40 percent of energy use while reducing the infrastructure cooling load.  When you consider that each watt of IT energy requires at least as much energy in infrastructure support, a huge potential for savings and energy reduction exists without a risk to day-to-day operations.

Here are seven reasons to remove orphan servers from the data center floor:


1.     Be Net Productive – nonproductive servers are net consumers of power, space, and cooling rather than producers.  They increase energy costs while providing little to no return on investment (ROI).  Clearly, shutting them down reduces energy waste and increases the value of the data center.


2.    Quick Win Project – eliminating unused servers is a short-term project that requires no additional hardware investment.  It becomes a “quick win” project that yields a measurable increase in ROI, data center efficiency, and an extended future savings in energy costs.


3.    Increase PUE and DCiE Accuracy – organizations rely on industry-standard efficiency metrics like power usage effectiveness (PUE) and data center infrastructure efficiency (DCiE) to determine the effectiveness of their energy management.  Terminating unnecessary servers is an easy way to improve PUE and DCiE reports.


4.    Reducing Total Assets to Manage – removing orphaned servers makes the remaining servers in the data center more effective.  With less assets to keep track of, managers can spend time attending to the servers critical to running business applications.


5.    Eliminate Hot Spots – because servers can run hot, removing them reduces “hot spots”.


6.    Green Friendly – reducing the amount of hardware running in the data center helps an organization reduce its carbon footprint.


7.    Gain Space – consolidating equipment frees up space in the data center.
 

Download our white paper “Nonproductive Servers in the Data Center” to learn more.

Michael Tresh, Director of Product Management

All Posts