WP#34 - The Green Grid Data Center Compute Efficiency Metric: DCcE

Overview

WP#34 is a white paper published by The Green Grid on 20 December 2010, with Mark Blackburn (1E) as editor. It defines the Data Center Compute Efficiency (DCcE) metric and its sub-metric Server Compute Efficiency (ScE). The aim is to quantify how efficiently compute resources are used. The full text requires signing in with a free The Green Grid account; this record is based on the summary on the public landing page.

The metric's focus is the IT side rather than the infrastructure. While PUE measures the energy efficiency of facility infrastructure, DCcE is intended to do for IT what PUE does for infrastructure. This lets operators locate where inefficiency lies in compute resources. A facility may have a good PUE, yet if its servers run at low utilisation, inefficiency can persist in another layer.

According to the paper, the metric is meant for continuous refinement in the same way as PUE; it is not a one-off assessment but an indicator tracked and improved over time. As a design implication, the record suggests an optional IT efficiency panel in which the user enters server utilisation data and sees an ScE/DCcE-style indicator beside PUE. Given these caveats, the paper should be treated as a conceptual framework.

It is a 2010 study, and the calculation form of the metric, accepted threshold values and example results are not part of the public summary. This record therefore does not contain formula details; these must be verified from the paper itself. Whether the metric has been widely adopted in the industry cannot be inferred from this record.

Key points

1. Dated 20 December 2010; editor Mark Blackburn (1E).
2. Defines DCcE and its sub-metric ScE.
3. Lets operators locate inefficiency in compute resources.
4. Intended to do for IT what PUE does for infrastructure.
5. Intended to be used for continuous refinement, analogous to PUE.

Using it in design

• In an optional IT efficiency panel, allow users to enter server utilisation data.
• Display an ScE/DCcE-style indicator beside PUE.
• In low-utilisation scenarios, state in the report that a good PUE alone does not mean efficiency.
• Verify the metric's definition from the paper before implementing its formula.

Who should read it

Designers, operations managers and sustainability reporters who evaluate IT and facility efficiency together.

Publication details

Authors

Mark Blackburn; Dan Azevedo; Andy Hawkins; Zeydy Ortiz; Roger Tipley; Sven Van Den Berghe

Publisher

The Green Grid

Year

2010

Document number

WP 34

Document language

English

Source type

Industry body

Access

Citation only

Licence

Copyright retained; citation only — the publisher’s terms apply to the original document

Citation