25 May, 2018
In a climate of constantly accelerating digital development that is redefining customer expectations, every individual business unit must possess the capability to react quickly to market changes and stay ahead of the competition.
This will almost inevitably result in increased IT demands – demands which data center managers and capacity planners are charged with accommodating, usually within a legacy infrastructure. But balancing different interdepartmental IT demands against the capacity and cooling limitations of the existing data center suite is no easy task. And with revenue loss and damage to reputation amongst the potential consequences for failure of critical infrastructure, the stakes are certainly high.
The good news is that virtually prototyping the data center environment can play a key role in optimizing capacity provision, allowing data centers to reliably serve business needs and support growth.
The Capacity Conundrum
Macro-level capacity planning is a prerequisite for facilities managers in order for them to understand current utilization, and map against available capacity for their entire portfolio of data centers. Because demand fluctuations are usually driven by the needs of individual business units, they are correlated to the necessary IT equipment required to meet application requirements, and then matched against the available space, power and cooling.
At the highest level, capacity limitation is based on the lesser of two resources: UPS power capacity or overall cooling capacity. If a shortfall in available capacity against demand is identified, this allows facilities managers to remediate the situation. This could involve adding more cooling and power, migrating to a colocation provider, or even building a new data center. But given the expense involved with such procedures, effective capacity planning allows for more drastic measures to be planned for and delayed until absolutely essential.
However, capacity planning that is conducted on an analytical basis (spreadsheets) will not remain accurate long term – as it fails to take physics into account.
Go with the (air)flow
Analytical capacity calculation may work initially, but over time, the projections will become less and less accurate. Cooling capacity is driven by both temperature and volume of air. Whilst the overall airflow delivered to a data center may remain the same, over time the amount of air delivered will deviate from the original design, as the internal environment is altered by hardware upgrades and reshuffles.
Consequently, if capacity planning is based on cooling figures from the original data center design, inaccurate predictions will be reached. A more concise figure is therefore required – one which incorporates how effective the airflow is effectively cooling the IT equipment.
In the example below, the amount of IT equipment powered vs. cooled has deviated over time. The gap that exists between the power and cooling is the driving factor behind reduction in usable capacity.
Otherwise, the consequences of inaccurate long-term capacity planning could be a false sense of security regarding infrastructure performance, and lower resilience to failure than anticipated. In the worst cases, this could result in the need to build a new data center earlier than first thought, and at considerable expense. So how can facilities’ evolution over time be taken into account during the planning process?
Mapping Against a Changing Environment
Incorporating virtual prototyping into the capacity planning process facilitates more accurate calculations. Our data center management tools allows you to create a baseline virtual facility, which can be used to experiment with any alterations in a safe environment – giving you the ability the plan 30, 60 or 90 days in advance.
Using the Virtual Facility, you can implement any necessary remedial action as future loads are planned out. This may involve experimenting with different cabinets, trying different layouts or solving for airflow delivery based on matching projected load. In doing so, a much more accurate picture of real-time and future cooling demands can be established. This revolutionizes the capacity planning process and ensures that existing facilities can continue to accommodate business development long-term.
For more information about how our solutions can help you with the capacity planning process, please visit our website.
Blog written by: Akhil Docca, Corporate Marketing & Product Strategy Manager
Other Recent Posts
Designing High-Density Data Centers in 6SigmaRoom
Will Standard Cooling Methods Suffice in High-Density Designs?In this blog, we’ll explore whether…
Simulation School for Our Young Machine Learners
As humans, we are hurtling towards a not-so-distant future entirely supported and optimized by data…