Your efforts on planning space utilization should come from the start of the project. Many IT planners might have something in mind, with arrangements based on configurations they are used to supporting as well as upcoming deployments. There are plenty of obstacles and pitfalls that can trip up the process as well as after you’ve already settled into the space. When considering advice from experts, you can save on costs and misery in the long term.
The total cost of ownership (TCO) is a commanding metric to this process; when things aren’t clear about the goals of the operation, you might not be dedicating time and workforce energy to the right efforts within your organization. When weighing cost and efficiency you need to consider the factors that go into the TCO calculation you are making for the space or facility. If you want efficiency and are seeking to achieve it within an older data center, the quick solutions of the past decade still apply – with hot/cold aisle containment being among the most cost effective solutions.
Beyond this, there are other mistakes that data center planners and operators can commit. One of these is not knowing how to run your cooling units for a space, as running these units without a controls plan to understand the energy per unit and whether as a whole (with variable speed drives) might be better based on location. Even in these modern times, you can still hear about CRAC units with frozen coils and cycling on and off too much (causing wear that leads to additional failures).
Having the correct fire suppression system plan for an existing data center should be comprehensive, including not just the general main data hall but also the subdivided zones. Containment zones need to have a tie-in with the main fire alarm system so that any soft wall partitions or other blockers can drop to allow the sprinklers full range. This will allow the containment to be compliant with any codes as well as appease the local fire marshal.
One of the other things to consider is that data center halls are usually not filled to capacity, especially when opening. As it grows you need to be cognizant of the supporting equipment efficiency so that the systems are not over cooling, providing cooling where it is not needed, and creation of hot spots or low air flow dead zones. Being able to understand this airflow management and the changing power and It equipment demands is important to prevent this issues from cycling out of control during peak demand times.
It’s important to see a diagram that shows the placement of all of the gear, connections, and equipment across a data hall. Vendors and manufacturers will be aiming to hold to your standards but also be able to point out where there might be layout gains or misses from a maintenance or operations aspect. Updating these can be paramount to making the most of your existing facility and not triggering the need to expand or build something new too early. Of course there are many DCIM tools that are on the market to help with these aspects, and they are starting to merge to be more comprehensive with the possibility of becoming a ‘digital twin’ in the future.
Avoiding mistakes in data center space planning
A common data center headache is slacking on planning space utilization and following the plans and rules established at the start. There are plenty of obstacles and pitfalls that can trip up the process after the data center has settled into full operation.
There has been the tendency toward hot-aisle containment when aiming to maximize the space layout while allowing for flexibility on equipment placement. Planning around the CRAC units and fire suppression systems are also important, as the layouts above and around the data center hall can dictate locations and rack row spacing if it was not part of the original plan.
Data centers are not like other facilities – they are usually not full on the first day and may take years to reach full deployment across the entire facility. And even then equipment may be swapped out many times with new configurations and requirements, changing the needs of the original space plan with a myriad of exceptions. Even as that is done the teams need to understand how the original standards evolved so that the normal power densities are thoroughly thought out with regards to the space and cooling.
Then there are certain loads that may be high capacity, requiring specialty in-row, liquid or rack-rear cooling. Suddenly a data center can accept much higher density, but the cost can be a considerable amount of space. PDUs are also involved with the space planning to make sure that power distribution can be comprehensive while not having to make long power runs across the data center to specific server racks.
And there comes a time when the data center needs a comprehensive audit, by an experienced third party. They can evaluate the status and what investments may be worthwhile to increase capacity and space as well as your true limits based on the current data center equipment. They will be realistic about what the data center can do and prevent over-planning. From there it is up to the data center manager to verify and keep the provisioning on track.