Unique Data Center Energy Management Solutions

Posted on November 28, 2012 by Gento

According to a recent New York Times article, U.S. Data Centers consumed 76 billion kilowatt-hours in 2010. At the Green Data Center Conference in Santa Clara, there was a great deal of discussion on data center energy waste. But there was more on just how far data center managers are willing to go to make their data centers more efficient without sacrificing performance. Raritan has several inventive solutions to achieve this.

Solutions to optimize comatose servers

One of the main issues data center managers want to solve is how to manage their servers so they’re not running when they’re not needed. It’s a knowledge and management issue. In fact, comatose servers are one of the largest wastes of energy in the data center. Kenneth Brill of the Uptime Institute estimates that 15% - 30% of data centers’ servers could be turned off.

Raritan’s Power IQ, when coupled with our intelligent Power Distribution Units (iPDUs), helps you identify under-utilized infrastructure so you can save energy. The solution gives an entirely new meaning to the phrase “knowledge is power.”

Over-provisioning of power and infrastructure

• 99.5% SLAs.
• Complaining end users who have come to expect an “always-on” readiness.
• Demanding industries like finance and gaming where transactions happen in milliseconds and nanoseconds.

In the face of end-user demand, try telling a data center they can’t over-provision.

Over-provisioning power supply to data centers is simply standard practice. Data center managers know it’s the only way to keep up with service level expectations and to manage spikes in traffic.

Or is it?

When data center managers have granular data down to the outlet level, and have the software to make sense of all that data, they’ll begin to see where energy capacity is stranded and can be better used.

What data center managers need is a tool set to visualize the data center from an asset standpoint - how much capacity any particular device is using in real-time. Such a solution would allow them centralized monitoring, email notifications, graceful OS shutdown, outlet control, powerful analytics of your costs, line capacities, carbon footprint and rack temperature.

Capacity vs. Consumption

As you probably know, most colocation facilities charge customers for energy by the capacity they provide rather than the real-time consumption per month or per year. Because of this, the customers say “Well, if I am paying for a certain capacity, then it’s up to me to utilize as much as possible so that I don’t leave any money on the table.”

Raritan has a solution for customers in this situation. We created power strips that will be rated for the maximum available rating allowable for each particular need. Meanwhile, most of the industry has chosen to stay under a certain rating to minimize the PDU cost. What it leaves the customer with is a less than optimal solution. Raritan’s solution gives data centers the maximum rating regardless of the capacity needed. It’s within ANSI guidelines and comes in at a reasonable price. Also, this eliminates the need to create more cabinets in the data center or in the colocation facility and saves a considerable amount of money and effort.

Large deployments of iPDUs

If your data center embraces all of this and suddenly decides to install hundreds, or even thousands of iPDUs, you could have a real mess of credentials, unique policies and alerts. Raritan came up with a simple way to help:

A USB Stick.

You provide the USB stick with a list of PDUs and power strips and their corresponding credentials and policies.

One of your techs goes to each PDU, plugs this USB stick into them and the PDU will read what’s in the stick and configure itself with the required credentials.

When an iPDU is configured correctly, we put a smiley face on the screen so you know it went well and it took the right credentials.

Further Reading

Bill Brill Interview - WallStreetAndTech.com

Energy Star Data Center Energy Efficiency Initiative

New York Times - Data Center Energy Waste