Home > Benchmarking

Benchmarking

Metrics

Self-benchmarking

Case Studies

Lawrence Berkeley National Laboratory

Benchmarking

Energy benchmarking is one of the key tools for achieving increased energy performance in data centers. As a top-level indicator of savings potential, benchmarked results for actual data centers show a 10-fold variation [chart] in electric power requirements per square foot of floor area. Other studies [chart] have found similar results. Benchmarking begins to be more valuable when extended to the end-use level [chart | table]. An example of more specialized benchmarks is the percentage of total data center energy use devoted to HVAC.


Read an overview presentation [download] on the approaches and uses for benchmarking.

The key to benchmarking is selecting appropriate metrics.

Benchmarking serves several important purposes:

  1. to establish a baseline for typical facility energy use
  2. to provide comparison to other similar facilities and/or similar systems highlight energy intensive systems and components
  3. to highlight areas of potential improvement
  4. to identify maintenance or operating problems
  5. to establish best practices for use in future designs

We have developed in-depth benchmarking case-study reports for 14 data centers.

A Self-benchmarking Guide is tailored for use in data center environments and will enable you to implement your own benchmarking process.