Tiers and Beyond: Understanding the Multiple Ways to Classify Modern Data Centers
Different data center classifications impact reliability, cost, and performance. Understanding these distinctions is crucial for your IT strategy.
At first glance, the way data centers are categorized may not seem especially important. After all, all data centers do essentially the same thing: provide space for hosting IT equipment.
It may not appear to matter much which type of data center a business chooses. However, the various data center categories can play a major role in determining how well a data center fits a given need.
Some types of data centers offer higher degrees of availability and uptime than others, for instance. Some offer special services, like "white-glove " management, that aren’t available across all categories of data centers.
Understanding these classification systems is crucial for making informed decisions that align with your business requirements, performance expectations, and budget constraints. Let’s explore the key frameworks that define today's data center landscape.
Primary Data Center Classification Methods
There are multiple approaches to categorizing data centers. Here’s a look at the most common types.
1. Tiers
The most widely recognized categorization system for data centers is the tier system, developed and certified by the Uptime Institute, which provides guidance on increasing IT availability and reliability.
The tier system focuses on a data center’s reliability and uptime guarantees. The higher the tier (ranging from Tier I to Tier IV), the more redundant systems are in place, resulting in greater resilience against downtime from power failures, cooling issues, or other infrastructure problems.
Related:AI Data Centers: A Popular Term That’s Hard to Define
Data centers seeking official tier certification must undergo rigorous analysis and validation by the Uptime Institute, which assesses their design, constructed facility, and operational sustainability before assigning a tier ranking. It’s important to note that while some data centers may claim "Tier 5" status, this is not part of Uptime Institute’s official classification system, which recognizes only four tiers.
When evaluating facilities, businesses should verify whether a data center has received formal Uptime Institute certification or is simply using tier terminology self-referentially. This distinction could significantly impact the actual reliability guarantees you receive.
2. Power Capacity
Another popular way of sorting data centers is based on how much energy they consume. The figure is usually measured in megawatts – hence why you might see one company talk about its latest "100 megawatt" data center project, while another is labeled as a "10 megawatt" facility.
This method of ranking data centers makes sense in the respect that the more power a facility consumes, the larger it tends to be. Thus, paying attention to power ratings can help businesses select data centers with the appropriate scale for handling the amount of IT equipment they need to deploy.
Related:Current Debate: Will the Data Center of the Future Be AC or DC?
The downside of power-based data center categorization is that total power capacity has little to do with other important aspects of data center operations, like reliability or sustainability.
3. Sustainability Certification
Speaking of data center sustainability , it’s also possible to differentiate between facilities based on their sustainability status. While there is no tiering system that assesses varying levels of data center sustainability, some data centers hold certifications related to sustainability, such as LEED and ISO 50001.
If sustainability is a priority, choosing a data center that has attained a specific sustainability standard makes sense. Keep in mind, however, that the overall sustainability of a facility doesn’t necessarily imply that individual racks, servers and workloads are configured in an energy-efficient way. Nor does it always mean that the data center sources energy from clean sources . The certifications primarily focus on energy efficiency.
Related:A Practical Guide to Data Center Yield on Cost (YoC)
4. Data Center Service Model
Data centers can be categorized based on the types of services they offer or the customers they serve. There are two main types of data centers in this respect:
Private data centers, which serve a specific company.
Colocation facilities , where multiple businesses can rent space and deploy IT equipment.
In general, colocation is the way to go unless your business’s IT needs are large enough to justify building an entire data center. Colocation also offers the benefit of making it easy to distribute workloads in various parts of the world (which can, in turn, improve latency for users based in different regions) by purchasing colocation space in multiple data centers. And colocation centers sometimes offer special services, like white-glove monitoring and management of IT equipment, which businesses would need to provide for themselves in a private data center.
5. Size
A final common approach to comparing types of data centers is general descriptions of their overall size. In this category, the common types of facilities include:
Hyperscale data centers, a generic term that refers to very large facilities.
Traditional data centers, meaning facilities that are large but not as huge as ‘hyperscale’ locations.
Modular data centers , which are typically small, although they can be combined to build larger facilities if desired.
Micro data centers , an even smaller type of facility, are often capable of supporting just one or two server racks.
All these ratings are subjective; there is no formal definition of how large or small a facility needs to be to fall into one or the other of these categories. Still, these descriptions can be useful for organizations seeking a quick assessment of a data center's total hosting capacity.
About the Author
Technology Analyst
Christopher Tozzi is a technology analyst with subject matter expertise in cloud computing, application development, open source software, virtualization, containers and more. He also lectures at a major university in the Albany, New York, area. His book, "For Fun and Profit: A History of the Free and Open Source Software Revolution ," was published by MIT Press.
You May Also Like