Blog post

Data Center Space Efficiency Metric

By Dave Cappuccio | July 02, 2013 | 4 Comments

Data Centers

The DCSE Metric: A simple way to look at data center space is to analyze the effective use of space by existing IT equipment, relative to the total available space for IT. The DCSE metric factors in both Horizontal Space Utilization (HSU) and Vertical Space Utilization (VSU).

Vertical Space Utilization (VSU)

VSU = Installed IT Equipment (RU) divided by (Total Rack Space Equipment (RU) * Optimum Target)

Where VSU is the ratio of the total quantity of installed IT equipment in terms of “RUs” (i.e. standard rack units – 1U or one rack unit, 19″ [48 cm] wide and 1.75″ [4.45 cm] tall) to the total number of “RUs” available; both in all racks inside of the Data Center facility. Optimum target is simply the maximum utilization level allowed for racks within the data center. Ideally this number would be close to 100%, but in many older data centers it is not possible due to cooling or power constraints at the rack level. By applying optimum targets to this formula we can chart a metric that is relevant to any data center design.

Horizontal Space Utilization (HSU)

Where HSU is the ratio of the total quantity of installed racks divided by the maximum quantity of racks supported by the Data Center facility.

DCSE Example

The organization supports a small data center of 1,600 square feet, but currently the floor space is near 85% of capacity. Therefore the current number of racks = 45, and the maximum number of racks = 53 (assumes 30 square feet on average per rack).

HSU = 45 / 53 = .85

The average rack utilization is 70% and the estimated “optimum” utilization is 80% due to power and cooling limitations. Therefore the actual installed equipment or rack unit count is 1,333 (45 racks X 42 RU per rack X 70%). The Optimal installed equipment count is 1,523 (53 Racks X 42 RU per rack X 80%).

VSU = 1,333 / 1,523 = .88

This shows a data center nearing it’s logical capacity, but with room to grow if configured properly. We multiply both VSU and HSU variables together and by applying a geometric mean to their product we insure that small or large changes in one variable do not have an unbalanced impact on the overall score (see Evidence). Therefore, the Data Center Space Efficiency index becomes:

DCSE = Geomean(.85*.88) = 86% capacity

Given the criteria above the data center is operating at 86% of it’s potential capacity for equipment and space utilization – no great surprise.

What DCSE points out rather quickly is the potential growth available within this existing configuration. With a combination of higher virtualization levels and increased rack densities it’s likely this rack environment will support existing growth rates for quite some time. And yes, we must assume that both power and cooling are available to support these higher densities., but by factoring in the Optimal rack density much of the power and cooling issues can be mitigated. If not, an analysis of the cost to add additional power and cooling vs. the cost to build out a new data center might in fact change the overall decision making process.

Calculating the Impact of Technology Refresh

An interesting use of DCSE is it can also be used for what-if analysis on potential upgrades. As an example, given the environment above, let’s assume the plans are in place to upgrade ½ of the existing installed base of servers from 2U devices to 1U. The results are as follows:

HSU stays the same since the rack count does not change.

The current installed RU count is 1,333 but with 1/2 being upgraded to 1U servers the total used RU count would become 1,333 X 0.75 = 1,000.

VSU changes to 1,000 / 1,523 = .53

Therefore: DCSE = Geomean(0.85 * 0.53) =  68% capacity

Using this model is becomes clear that by using a technology refresh of ½ of the servers the data center space is now at 68% of possible capacity rather than 86%, providing a logical method to increase productivity while deferring capital expense on a new data center for years to come.

Bottom Line: DCSE is not the end-all for data center space planning, but was designed to give IT Managers a view of capacity levels within their data centers, and a means to compare that level to realistic potential (optimal) capacity levels, rather than just using a hypothetical maximum. Using DCSE on an ongoing basis will yield a clear view of how space and capacity targets are changing over time, and how an organizations overall data center efficiency is improving.

The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.

Comments are closed


  • Mark Thiele says:

    Hi Dave,

    I like where you’re going with this, and have been working on a proposed new metric myself for several months now. However, I’m focusing more on the larger picture of the entire data center facility, it’s utilization, sustainability of design/construction and it’s lifecycle. I think focusing on an individual rack vs. total power against utilized floor space is a better model for instance. I also see Sustainability as a key goal in the sense that the more you can fit into one data center the fewer data centers we have to build. If we build less data centers we reduce the tons of carbon created through the manufacturing and build process (Cement, steel, glass, carpet, insulation, pavement, etc.)

  • Hello just wanted to give you a quick heads up. The words in
    your article seem to be running off the screen in Internet explorer.
    I’m not sure if this is a format issue or something to do with browser compatibility but I thought I’d post to let you
    know. The design look great though! Hope you get the issue
    fixed soon. Thanks

  • This really is a good blog post.quite great.excellent.Thx you will really a lot of

  • Thank you. We’ll look into it.