Effective power design in data centers is a critical component of their ability to work successfully and accurately without encountering a range of potentially disastrous issues, including instances of total power failure that lead to significant losses.
Mistakes are pretty standard when creating optimal data center power design, especially when developed by undertrained data center staff rather than industry experts and professional data center advisors like C&C Technology Group.
The following sections explain a handful of the most common mistakes that tend to occur in creating power design that data centers need to actively try to prevent to keep their systems running as effectively as possible while avoiding total system shutdowns, overloads, and other significant issues.
Common Data Center Power Design Mistakes
Choosing Improper Data Center Power Equipment
There are many components and pieces of technology that go into creating effective data center power design. If even one piece of equipment is chosen incorrectly, it can lead to an under or overstimulation of the power load that can cause critical problems for the entire system. For example, an uninterruptable power supply (UPS) with a power rating of 500 kVA would deliver 300 kW along with a leading power factor of 0.8.
This will provide 400 kW with a 0.8 lagging power factor. Any mismatch between the values of the servers and the UPS power factors would trigger a system overload that causes the UPS to trip, leading to several issues for the data center.
Poor Cabling Between Servers and UPS
Suppose the data center power design should accidentally undermine the size and capabilities of a power cable connecting servers with the power distribution unit (PDU). In that case, there can be significant harmonic distortion that can trigger instances of excessive heating, dropping of voltage, and eventual overloading of the entire system.
Developers of power design need to carefully consider these values of distortion and select the most effective cabling to avoid such issues. General requirements state that to have a 100sqm cable working at regular performance, another cable of at least 130sqm will need to be selected as an alternative backup to help ensure the first cable doesn’t fail and accidentally shut down the entire system.
Inaccurate Capacity Estimations of Data Center Power
If the power capacities of the data center and its equipment are improperly estimated during the power design process, they may be underutilized and lead to eventual power supply failures. It’s essential to carefully study the historical data of the center’s power consumption and its general power use trends to determine the optimal power requirements for the system.
Power design creators should also take the time to carefully examine the nameplate power ratings of all their equipment and develop an automatic failsafe for the system generators to turn on and off in the event of an issue. Such a failsafe should help protect the integrity and performance of data center systems and equipment. Ensuring accurate estimations of the data center’s power capacities is essential in creating a balance between the system’s equipment to protect against both overloading and underutilization.
Are you on the hunt for top-quality data center power solutions, planning, monitoring, and more? Reach out to the expert specialists at C&C Technology Group today to learn about everything they can do for you and your data center.
System Failure Due to an Excessive Load at a Single Location
Another significant mistake that data center systems tend to encounter due to ineffective power design is overall system failure due to an excessive power load occurring at a single point. When multiple components and pieces of heavy-duty data center equipment are connected to a single point in the power line, overloads and malfunctions can quickly occur, causing everything to shut down in a sudden burst of power failure. If these systems and pieces of equipment have high power requirements and are connected to a single PDU, then all systems would immediately stop functioning in the event of a PDU breakdown or power failure.
To effectively overcome this significant issue of power failure at a single point of equipment connection, data center power designs that effectively utilize two PDU systems connected to data servers in a dual bus architecture design should be adopted. This style of system architecture can help ensure that, should a single PDU fail, other PDU systems would take over its role to prevent an entire system shutdown and the range of problems that can develop from said shutdowns.
Related: Data Centers: What They Do & Why
Not Establishing Clear Project Leadership
As part of the data center design process, you should establish clear project leadership and ownership for all tasks and responsibilities associated with building the data center. You can have multiple stakeholders like IT staff, contractors, and business owners that will all be involved with the data center’s design and construction.
As part of the early design process, designate which stakeholders will be responsible for each task and who will have the final say on design and construction. This will establish a clear line of accountability for the project.
Choosing the Site Before Finishing the Design Process
If you select the site for your data center before finishing the design process, you may face a host of spatial constraints because the space doesn’t accommodate your business’s data center needs or future growth. By finishing the design process, you’ll know what size capacity you’ll need for the data center and its future growth.
This will save time and money by finishing the design process first. You’ll be able to assess sites for the data center more accurately. You won’t have to get creative with organizing your hardware or reorder prefabricated components to fit the room constraints. And you won’t have to move sites as your business grows because you accommodated the future growth.
Not Planning for Future Growth
During the design process, future growth needs to play a prominent role when estimating the data center’s capacity. By anticipating and leaving room in the data center for future growth, it will enable the business to expand without moving sites, incorporate new technology, and run operations efficiently.
The average lifespan of a data center is 10-15 years. But your organization can extend that lifespan if growth is built into the design. Then your operational costs will be significantly less since you will only need to replace hardware as the technology demands.
Failing to Identify Dependencies
When designing a data center, equipment dependencies are often overlooked or underestimated. Most of the equipment has interoperability and compatibility requirements and constraints that can affect the design of the infrastructure.
You’ll want to research the equipment extensively during design and plan for compatibility issues or expansion requirements. Otherwise, you may have additional costs or interoperability issues that will delay construction.
What Factors Should You Consider When Designing a Data Center?
To avoid making any of these mistakes, you should set a few key factors as a higher priority when designing a data center. In addition to the site selection, hardware, facility layout, cooling, network, and power distribution, your organization needs to focus on:
- Availability: The proportion of uptime the data center is operable.
- Reliability: The ability of the center to perform under specific conditions for a designated time.
- Serviceability: The ability to diagnose and solve any issue in the data center.
- Capacity: The size of the facility and maximum load the data center can run.
- Scalability: The ability to expand with planned load growth over the years.
- Manageability: The ease of managing the facility and all its infrastructure.
- Criticality: The importance of the data center in providing service to end customers.
Want expert help designing your data center to avoid many of these mistakes? Learn more about our data center design advisory services.
Related Link: Data Center Capacity Planning [Ultimate Guide]
Inaccurate Designs for Automatic Transfer Mechanisms
Transfer switches are specially designed to help move power from the utility power supply to backup generators within the data center. For this transfer of power to occur successfully and without triggering any potential system issues, data center staff will need to manually disconnect any neutral terminals if they happen to employ the use of power contractors or four-pole circuit breakers.
This requirement is because most currently adopted UPS systems require the use of an input-neutral connection that can enable the transfer to occur without the use of a transformer. When the generator powers on and the UPS doesn’t have any neutral terminal for reference, floating neutral is created, and a floating load can develop as a result, leading to the server tripping or freezing. The solution for this issue involves using an automatic transfer switch with an overlapping neutral mechanism that will help safeguard against potential instances of system tripping.
Do you own or manage a data center that needs specialized power system designs, monitoring, protection solutions, and more? Reach out to the experts at C&C Technology Group today to learn about their range of top-quality products.
Last Updated on May 30, 2023 by Josh Mahan