Upgrading a data center can be like the “butterfly effect,” and not always in a good way. Any one decision – whether it concerns hardware procurement, network infrastructure or security – can branch out in numerous ways, with a slew of unplanned trickle-down consequences that happen in response to an earlier action. Obviously, it’s not as if the decision to use a chilled water system over a glycol-cooled system for data center cooling will cause an epic data center failure, but single decisions can have widespread effects.
In order to assure that upgrades and transformations to data centers lead to premium performance and forestall any unpleasant surprises, owners and operators can focus on ACE – availability, capacity and efficiency. Without these objectives – and tools to measure how effective equipment is in living up to them – data centers can produce less than optimal ROI. The data center components that operators design, procure, install and manage likely affect each of the three ACE qualities, but it’s important to figure out how to quantify and forecast the effect a new change or asset would have on data center operations overall.
Eliminating ‘What Ifs’ Through Measurement
Say it all together now: “You can’t manage what you don’t measure.” This adage gets repeated time and time again in the data center industry because detailed data reporting and metrics are absolutely essential to assessing – and improving – data center functionality. Without being able to isolate different variables within data center operations, organizations have a hard time targeting the areas in which improvement would be most beneficial.
Of course, availability, capacity and efficiency are, by themselves, terms for evaluating performance. They’re not single numbers – rather, they stand for overarching assessments that blend dozens of different metrics together. This gives evaluators and designers a complicated task – each potential asset or design change, from servers to cooling to IT load, must be assessed for how it functions in isolation and how it operates as part of the larger data center infrastructure. Only then can it be effectively managed, reconfigured and improved.
Future Facilities developed the ACE metric as means of assessing data center key performance indicators, intending it to complement existing industry metrics, such as power usage effectiveness, and ultimately removing the “what ifs” involved in forecasting data center performance. A recent report by 451 Research discussed the need for more comprehensive measurements and KPIs in order to determine how changes impact the data center.
“The key to ACE is that it enables managers to see how their data centers are scored against three important parameters (which are often competing) in terms of design goals or impact: availability, capacity and efficiency,” the report stated. “[I]ts adoption may encourage wider use of models to simulate and improve data center performance.”
The Potential Benefits of ACE
Whereas PUE measures the functionality of one aspect of data center functionality – and has been criticized for its limitations – ACE proposes to evaluate how disparate elements of the data center interact. On its website, Future Facilities outlined a few examples. If a designer could use monitoring tools to predict power and cooling availability, for example, he or she could use that information to design more efficient servers. Airflow and temperature measurements could factor into cooling efficiency decisions.
The potential benefits of ACE are significant. As rising computing demands and rapidly expanding environments thin the margin of acceptable leeway from optimal efficiency and performance expectations, data center decision-makers need tools at their disposal that improve visibility. It could help create a brighter future for data center development.