With 2014 looming, energy managers in every industry have an opportunity to think about the past year and how we’re going to continue to improve in the year to come.
Nowhere is more the case than in data centers, where rapid technology change and exploding demand are bringing both new tools and huge challenges to energy management.
The simple fact is that today’s data centers are extremely hungry, with the Uptime Institute estimating that they collectively consume as much as three percent of all global electricity production. Energy has become a top-line operational expense, making efficiency a critical part of how companies will build a long-term competitive advantage. Considering that data centers can waste 90 percent or more of the electricity they pull off the grid (according to a 2012 New York Times investigation), that competitive advantage can be massive.
Understandably, this has driven a lot of recent innovation. The last couple of years have seen a whole array of new solutions emerge to help operators measure and manage their energy better. That includes everything from virtualization, to Data center Infrastructure Management (DCIM) solutions, to Application-Aware Power Management (AAPM) software.
As more companies adopt these tools in the coming year, they’ll find that they’re suddenly equipped with much deeper insight into their energy consumption, server and application performance, and energy productivity. With measurement and monitoring happening not only at the facility level, but also at the level of individual servers and applications, energy managers are going to have a lot more tools in their toolkit than ever before.
Let’s take a look at five key areas where these new metrics and new tools are already helping data centers to improve reporting, find inefficiencies, and radically cut energy costs.
1. Transaction costs
With detailed energy metrics, companies are finally able to accurately track the operational costs associated with individual transactions and different transaction types. This allows for much closer alignment of costs with broader business objectives, especially for companies that are highly margin-sensitive.
As just one example, eBay has done a great job tracking their performance in this area. They realized a long time ago that energy has a major impact on their cost to process every transaction. To start improving efficiency, the application, business, and data center teams all need to be looking at the same efficiency metrics, which is why eBay uses a simple dashboard to share their performance. (You can check it out for yourself at dse.ebay.com.)
2. Application efficiency
Do you understand which individual applications have the most impact on your costs? Inefficient applications can drive up the energy consumption of the servers running them, without also driving up revenue or performance enough to justify those additional costs.
With application-level insight into energy performance, you can align application performance and costs more closely with revenue, allowing you to scale back or replace applications that are dragging down your bottom line.
3. SLA costs
Some Server Level Agreements are simply more expensive to deliver, based on transaction costs, application efficiency, and uptime and backup requirements. Data centers have always worked to accurately model the cost of meeting certain SLAs. Detailed energy metrics makes this process much more reliable and easily reportable.
4. Workload and idle servers
Are you paying the same energy bill regardless of whether actual demand is up or down? A classic example of this is data centers that support online retail: most pay to run plenty of excess capacity 365x24x7 just to support demand spikes that generally happen on just a few key days per year. The rest of the time, all of those idling servers waste a lot of power without generating revenue.
How much waste? Server utilization rates are typically very low, currently averaging in the 6–12 percent range. When you consider that a completely idle server still draws 60 percent of its maximum power, it becomes fairly obvious that we’re talking about a very significant chunk of year-round energy costs.
Intelligent power management solutions allow data centers to track seasonal, daily, and even hourly fluctuations in demand, and then dynamically adjust the power state of idle server in response. In many data centers, this can mean an immediate reduction in server power costs by as much as half.
5. Comatose and legacy server costs
The new year is also a perfect time to get a head-start on your spring cleaning. With detailed, per-server energy metrics, it’s relatively straightforward to determine the computing power to energy consumption ratio of legacy servers, which in turn makes it much simpler to make a business case for strategic upgrades or virtualization. It’s also a great time to look for comatose servers that haven’t processed a transaction in a long time but keep quietly drawing power.
Of course, only larger data centers with well-developed energy management programs are likely to be able to improve in all five of these areas at once. Yet no matter the size, it all starts with a simple wishlist for 2014: deciding where you’d like to improve first and what new metrics are most critical to those goals. Then it’s just a matter of seeking out the solution that will equip you with the metrics you need.
Aaron Rallo is the founder and CEO of TSO Logic. Aaron has spent the last 15 years building and managing large-scale transactional solutions for online retailers, with both hands-on and C-level responsibility in datacenters around the world. He can be reached at arallo@tsologic.com.