10 Best Practices for Power Management in the Latest Data Centers

1 of 11

10 Best Practices for Power Management in the Latest Data Centers

by Chris Preimesberger

2 of 11

Get Beyond PUE

PUE (power usage effectiveness) is a metric used to determine the energy efficiency of a data center. This is determined by dividing the amount of power entering a data center by the power used to run the computer infrastructure within it. PUE is a great start, but managers need to understand the IT power efficiency of your equipment. With a fixed cooling infrastructure, upgrading IT equipment to lower the power consumption will make your PUE go up.

3 of 11

Managing a Data Center Is a Balancing Act

In data center management, managers need to focus on balancing utilization across facilities (power and cooling) and IT (CPU, Memory, Network and Storage). There are some good apps for that!

4 of 11

Its Cool to Be Warm

Increase inlet temperatures for servers to 80.6 per ASHREA (American Society of Heating, Refrigerating and Air-Conditioning Engineers) recommendations. Use hot/cold aisles to decrease cooling requirements and optimize air flow.

5 of 11

Facility Utilization an Operational Key

Eliminate all those bottlenecks that get built into the infrastructure through disconnected silos of servers or storage arrays that come in months or years apart. There is plenty of good data centerware now available that can get various departments to talk to each other, share resources, and, in turn, save costs.

6 of 11

PM Integration Needs to Be in Place

Power management needs to be integrated directly into capacity and performance management. This is ultimately about the transactions per kilowatt hour. Understanding server efficiency is an important metric.

7 of 11

Carefully Planned Refreshes a Major Factor

Maximize capacity and reduce power consumption through intelligent technology refresh decisions.

8 of 11

Get Updated and More Efficient with IT

Virtualization is an important factor here. Consolidating servers and storage arrays to use far less power and increase capacity utilization can have a huge positive power impact right away--not only over the long haul.

9 of 11

Know That Application Usage Drives the Data Center

Realize that application service levels drive the entire vehicle—including data center power, capacity and performance-level decisions. Always keep this as priority No. 1.

10 of 11

Keep Looking Ahead at New Possibilities

When planning ahead for a new data center or a DC addition, research new concepts such as tiered data centers, application quality-of-service grouping, storage pooling, and active/active multi-site configurations. This is where the data center is going in the future, so you might as well be on board early.

11 of 11

Use Power Chargebacks in Your Billing

Implementing chargebacks to share power costs amount the various corporate departments receiving IT services make the issue more visible to the managers who can make a difference. Including power usage into a chargeback mechanism is also a better method to allocate virtual machine charges to eliminate over-provisioning and under-utilization. Otherwise, many users will simply take advantage of the resources for as long as they can get away with it.

Top White Papers and Webcasts