Data Center Design: Essential Components and Best Practices

Did you know that a smartphone, a toothbrush, and a car all share a common design principle?

They all solve problems by first using human-centered design and functionality. This principle also applies to enterprise data center designs. Organizations can build data centers only after knowing their data processing, storage, and connectivity requirements.

As an IT infrastructure manager, you probably use data center infrastructure management (DCIM) software to run on-premises data centers that host legacy enterprise applications and virtualized workloads.

Now, if you're considering expanding on-premises capacity, enhancing public cloud services, or adopting a hybrid cloud strategy, you must first look into data center planning and design. This will help with seamless data management and operations while you work on improving your company's data center networking capacity.

What is data center design?

Data center design involves conceptualizing and designing a data center's infrastructure, architectural layout, and IT resources. It helps logically model a data center before its implementation.

Purpose-built data center design decisions require more than understanding tier level, cost-to-build, or watts per square foot. One wrong step can increase operational costs, waste capital, and prevent uptime performance goals from being achieved.

This article walks you through critical considerations to keep in mind while designing a data center.

Why is data center design important?

Data center design is crucial for housing network equipment and IT infrastructure systems for gathering, processing, storing, and sharing business data. The right design means having a dedicated space with the necessary cooling, power, reliability, and security systems to support an organization's computing demands.

Enterprises take modern data center design seriously because it impacts how well they deliver resources to partners, employees, and customers. Small or mid-sized companies often struggle to keep up with capital investments, operational expenses, and maintenance that data centers require. They often opt for colocation data center facilities that rent server rack spaces.

Want to learn more about Data Center Infrastructure Management (DCIM) Software? Explore Data Center Infrastructure Management (DCIM) products.

Key data center components

Depending on its size, a data center can occupy a part of a building, entire premises, or multiple buildings. These spaces create a robust environment of computer systems and servers that ensure uninterrupted data operations even during power outages.

The physical facility and the state-of-the-art IT infrastructure are vital components of any data center construction.

Data center facility

A data center facility is a physical building space where an organization deploys the IT infrastructure. It hosts complex storage, network, and computing devices for smooth data processing. Organizations must consider the following factors for a data center architecture while designing a facility.

• Space is the square feet or square meters of floor space a company needs to accommodate its IT infrastructure. Some may divide the space for different purposes.
• Power supply includes traditional, renewable, and auxiliary power sources that a business needs to operate all IT devices without disruption. Common infrastructure providing power to IT devices includes floor power distribution units (PDUs), busways (self-contained overhead power distribution systems), and rack PDUs.
• Cooling equipment like traditional heating, ventilation, and air conditioning (HVAC) systems, computer room air handler (CRAH) units, and computer room air conditioning (CRAC) systems prevent IT assets from overheating.
• Security measures like electronic door locks, alarms, video surveillance systems, and biometric scanners protect assets from theft and damage.
• Building management involves using building management systems, facility management solutions, or smart buildings software for monitoring and managing power, temperature, lighting, humidity, and security logging.

IT infrastructure

Data center infrastructure is a collective term for all the IT equipment that a data center building uses to run applications and provide data services to business customers. It usually includes the following elements.

• Servers support data center functionality by hosting applications and executing computing tasks. Facilities generally store servers in server rack cabinets.
• Storage devices like disk arrays not only store applications and business data but also protect them.
• Networking connectivity equipment creates a network of cybersecurity elements such as firewalls, routers, data center switches, and load balancers for efficient application data processing.
• Cables are wires that connect IT devices, physical servers, and other gear in a facility. An efficient cabling infrastructure ensures smooth data transmission while preventing unexpected downtime.
• Backup power supply systems, such as uninterruptible power supply (UPS) devices, emergency power systems, and generators, ensure infrastructure stability in case of power outages.
• DCIM solutions aid organizations with data center infrastructure configuration, monitoring, and reporting.

How to design a data center

Organizations create data centers based on their workload and unique needs. They can follow existing data center standards to design fault-tolerant, scalable, and safe facilities. Plus, periodic upgrades can keep facilities up-to-date and efficient for handling evolving business requirements. Companies usually focus on the following factors while designing data centers.

• Cooling units
• Space planning
• Electrical layout
• Conceptual design
• Facility maintenance
• Leak prevention systems
• Backup power generation
• Lightning protection equipment
• Switching and power distribution
• Site reliability and security systems
• Building construction requirements
• Building and environment monitoring systems

Data center design standards

The four major data center design and infrastructure standards are:

• Uptime Institute’s tier standard outlines the infrastructure a facility needs for smooth data center operations. It specifies four tiers based on data centers' topology and operational sustainability. Each data center tier defines specific business functions' fault, maintenance, power, and cooling capabilities.
• ANSI/TIA 942-B by the Telecommunications Industry Association (TIA) engineering committee is another globally recognized data center infrastructure standard. It specifies a data center’s physical infrastructure requirements for location, construction, fire protection, security, telecommunication, and maintenance. ANSI/TIA 942-B also features four reliability levels developed by Building Industry Consulting Services International (BICSI)-certified professionals.
• EN 50600 series, developed by the European Committee for Electro-Technical Standardization (CENELEC), is the European transnational standard for data center planning, construction, and operation. In addition to cable and network design, it also focuses on energy management and environmental viability.
• ASHRAE by the American Society of Heating, Refrigerating, and Air-Conditioning Engineers is a popular energy standard for data centers. It shares HVAC&R guidelines for using sustainable technology in a building environment.

Data center regulatory and operational standards

Other regulatory and operational standards include:

• SAS 70 Type I or II
• Sarbanes-Oxley Act
• Gramm-Leach-Bliley Act
• ISO 9000 Standard (for quality)
• ISO 27001 Standards (for information security)
• ISO 14000 Standards (for environmental management)
• EN 50600-2-6 Standards (for operations and management)
• Health Insurance Portability and Accountability Act (HIPAA)

Data center design best practices

The right design practices transform a space into an efficient data center facility adept at hosting and operating IT infrastructure. Enterprises emphasize physical space organization, cooling systems, powering utilities, and security for smooth operations.

Physical space and organization

Designing an enterprise data center is more complex than just knowing the space you have. It's critical to find the right-sized data center facility so you can accommodate future upgrades according to growing business needs. On the other hand, you don't want unused space either. Consider the following factors while planning a data center space design.

• Floorspace is the facility area where you store IT equipment. While you can’t plan for extra space while remodeling an existing data center, consider doing so while designing a new facility. That way, you'll be better positioned to accommodate future organizational needs.
• Lighting is essential for video surveillance. Consider using environment-friendly lights that turn on with motion sensors as people walk in. Also, it’s best to keep the lights on during maintenance or repairs.
• Noise is another component that needs the attention of data center owners. The hum of servers and cooling fans creates noise, because of which you may need hearing protection.
• Temperature usually remains low with cooling systems in place. So, support technicians may need protective clothing while working in data center facilities.
• Weight is another factor to consider because of the IT equipment load, especially at data centers with raised floors for passing cooling airflows.
• Cabinets and racks are essential for housing IT gear. Some companies add empty racks to accommodate more servers for future needs.
• Cables are key to network connectivity and power distribution in a data center. For safety and efficiency, consider using a mix of fiber optics and copper cables with the smallest diameters. Also, an overhead rafter system can be deployed or the wiring should be kept below the floor to clear pathways.

How to find the right data center size

Data center size varies based on an organization’s needs but boils down to factors like server configurations, technologies you use, network architectures, and budget.

• Density measures the amount of power a facility uses for data center operations. It is becoming increasingly important as companies use consolidation technologies like server virtualization to do more with less. You can calculate a data center's size and density from the peak kilowatt load and compute space.
• Physical size is the square footage of space in your data center. With newer technology, a data center room can hold more servers and storage arrays than before. That’s why it’s essential to consider the space for layout, capacity planning, and density discussions.
• Technology you use plays a key role in finding the right data center size. For example, companies using legacy technology can manage with traditional network and server architecture from smaller data centers. However, they must consider processing technologies, server consolidation, and density before upgrading their computing capacity.

Data center security

Besides protecting the facility, organizations should also protect their network, data, and IT infrastructure. Security in a data center involves the following aspects.

• Physical security or access security refers to managing human personnel so that only authorized executives can access the facility and its gear. Common data center physical security measures include deploying video surveillance, keeping employee logs, staffing security guards, and limiting access to racks and servers.
• Facility security includes various predictive maintenance measures that protect a data center environment from humidity, temperature, energy, and other environmental malfunctions.
• Network security protects a data center network with intrusion detection systems (IDS), firewalls, and access control lists (ACLs) to grant or deny permission to those requesting access to digital environments.
• Data encryption and backup keep sensitive data secure with password protection and data backup.

Data center security software

Organizations combine the following solutions to protect data center facilities and environments.

• Firewall software
• Antivirus software
• Log monitoring software
• Network security software
• Configuration management tools
• Intrusion detection and prevention systems (IDPS)

Data center cooling design systems

The energy a data center uses to power servers and other IT equipment converts into heat. When not controlled, heat and humidity damage electrical devices and may lead to malfunctions. That's why data center operators and managers use different cooling systems and ventilation techniques to maintain a facility's temperature and humidity levels.

Data center designers focus on a facility’s physical size and HVAC systems to measure its cooling needs. They calculate this by converting power demands in watts to cooling capacity in tons. Without the right-sized cooling system, a data center will either pay extra for the cooling it doesn’t need or not have enough capacity to support IT systems.

Data center facilities also must pay attention to how they handle cooled and heated air. Home and office rooms release cold air from one vent and remove warm air using another vent, allowing average temperatures by mixing cooled and heated air. This approach doesn’t work in data centers as it contains racks with extremely hot temperatures that require careful cooling.

Designing a facility for optimal airflow is crucial for cooling IT systems and lowering utility costs. That’s why data center designers use containment schemes like hot aisle or cold aisle layouts to handle server room air.

What is a hot aisle or cold aisle?

A hot aisle or cold aisle layout uses a common hot aisle to collect heated air from a server rack's rear end while exposing the front side of these rack rows to cold air. In this arrangement, all rack rears face each other.

Another commonly used method is liquid cooling. This approach removes all heat sources and fully or partially immerses IT gears in chilled, dielectric liquids like mineral oils. Liquid cooling technologies are more efficient in heat transfer than air cooling. However, when not done cautiously, these technologies may lead to flooding, liquid intrusion, part corrosion, and human safety-related challenges.

Data center power and performance

Utility costs are perhaps the major operating expense for enterprise data centers. Consider the following factors to maximize uptime and lower downtime with reliable, cost-effective power options for your data center technology.

• Use clean power free from electrical noise, spikes, and surges.
• Improve power reliability to prevent blackouts, brownouts, and disruptions.
• Install power backup solutions to minimize downtime during power failures.
• Boost power usage efficiency by monitoring power needs for IT and non-IT tasks.
• Ensure adequate power capacity for running the entire data center infrastructure.

Energy efficiency and data center sustainability

Organizations won’t be able to share, store, and access data at the click of a button without data centers. However, the limited supply of fossil fuels and increasing carbon dioxide (CO2) emission levels are pushing companies to be more conscious about energy efficiency. That’s why they are becoming more mindful about reducing wasted electricity and power usage effectiveness (PUE).

Data enters often end up wasting electricity because of using uninterruptible power supplies (USPs) to run their infrastructure. Their energy efficiency reduces as the input energy doesn’t result in useful work. That’s why they use measures like PUE to estimate energy efficiency.

Principles of sustainable data center design:

• High-efficiency cooling is used to limit the energy cooling systems' use.
• Digital tools and technologies for facilitating energy flow distribution.
• Low-carbon energy usage and adoption of renewable energy sources.
• Software optimization and automation for tracking and correcting energy waste.

PUE is the ratio of the amount of power entering a data center and the power used by the IT infrastructure in it. Companies with ratios closer to one are considered to have more efficient data center infrastructure. One way to achieve the best PUE score is to reduce non-IT energy uses and create an energy-efficient space.

Enterprises are also aiming to achieve net zero carbon emissions with green data centers that use energy from renewable sources. For example, they can use solar, wind, geothermal, or hydropower energy to power IT infrastructure in facilities.

Data center design checklist

Organizations can’t take a one-size-fits-all approach while designing data centers. That’s why it’s best to set them up for efficiency with the following tips.

• Use PUE metrics to measure power efficiency.
• Manage and optimize airflow for proper cooling.
• Boost power handling to improve power distribution.
• Strategize rack layout for better capacity planning.
• Create safety guidelines to keep the data center environment safe.

Build vs. buy a data center: Key considerations

Your data center design also depends on whether you’re building a new one or buying and upgrading an existing one. Building one requires a significant amount of capital, planning, and time to complete. On the other hand, you can use colocation services to lease servers and pay comparatively less operational costs. Not sure what’s best for you? Consider the following factors before finalizing your data center design.

• Consider time to market. The design and construction timeline also impacts the total cost of ownership (TCO), especially if you’re building a facility. For example, it may take up to 30 months to fully build out data centers. Working with data center infrastructure providers reduces this time to a year since they often have sites that you can build on.
• Capacity planning is important because older IT equipment is cheap to acquire but may add to your cost even after upgrades. You don’t want the computing equipment to outgrow the supporting infrastructure. Consider opting for modular build-outs to extend the life of the data center.
• Energy efficiency isn't a symbolic gesture anymore to show off your corporate social responsibility commitments. Data centers consume about 1,00 kilowatt hours (kWh) per square meter, about 10 times what a typical American household consumes. Data centers are estimated to assume up to 4% of global electricity consumption by 2030. You can use energy-saving technologies easily in a new data center, but energy costs will still be the largest expense.
• Data center operations management is important. Organizations need to consider whether they have trained professionals to maintain critical data center systems around the clock. Some companies work with third-party service providers to enjoy the support of skilled teams that can fast-track repairs and minimize outages.

Apart from these factors, the ultimate evaluation should be based on whether building a data center gives you a strategic advantage or restrains your resources in the long term.

Questions to ask while evaluating build vs. buy

Discuss these questions internally to have more insights into whether building or buying is more suitable for your organization.

• Experience: Has your organization operated a data center in the past? Has it built one previously?
• Relationship: Have you already worked with IT infrastructure consulting providers who can help with infrastructure design and maintenance?
• Project scope and scale: How does the proposed facility size and scope compare with your last?
• Previous budget and results: Were you able to keep the last data center project within the approved budget? Was it completed on time? Also, was the team satisfied with the results? Were there changes that impacted the cost and time to complete the project?
• Mistakes: What can you do to avoid previous mistakes made during the last data center project?

Data center construction challenges

Organizations must pay attention to the following factors to ensure seamless facility operations while building a data center.

• Insufficient expansion space: Business needs may differ in the future. That’s why it’s best to decide the square footage requirements after calculating the space for IT equipment, office space, and staging areas.
• Power and cabling issues: Most power failures or unplanned outages occur because of poor power backup planning. Moreover, data centers should also focus on reducing latency with network monitoring software.
• Total cost of ownership: In addition to capital expenses and energy costs, data centers require regular maintenance. Creating a budget is important for aligning these costs with an organization’s risk profile.
• Performance management: System reliability and high availability are key to a data center’s success. Enterprises can’t operate efficiently without a calculated approach to manage both.

Data center infrastructure management (DCIM) software

DCIM solutions are critical for data center management, asset monitoring, and performance tracking. Enterprises use these tools to boost uptime and availability while streamlining data management and documentation.

Best DCIM solutions

1. Site24x7
2. ManageEngine OpManager
3. IBM Turbonomic
4. Nlyte
5. Cisco Nexus Dashboard

Designing right is hard, but not doing it is disastrous

We’re less likely to notice whether a design is good or bad if it manages to serve our needs. However, data centers are complex facilities that can quickly become unmanageable without a well-planned design. Without it, you’ll compromise the efficiency of your cloud computing system, storage devices, and networking systems—all key to fueling business operations.

Start streamlining data flow between data centers and infrastructure assets with data center networking solutions.

This article was originally written in 2023. It has been updated with new information.

Sudipto Paul

Sudipto Paul is a Sr. Content Marketing Specialist at G2. With over five years of experience in SaaS content marketing, he creates helpful content that sparks conversations and drives actions. At G2, he writes in-depth IT infrastructure articles on topics like application server, data center management, hyperconverged infrastructure, and vector database. Sudipto received his MBA from Liverpool John Moores University. Connect with him on LinkedIn.