Cloud Rebalancing: What You Need to Know About Data Center Density

As companies optimize their IT environments to balance price and performance, many are considering moving some workloads out of the cloud. An IDC report from June 2024 found that about 80% of survey respondents “expected to see some level of repatriation of compute and storage resources in the next 12 months.”[1]

There are several factors driving this change: escalating cloud costs, security and privacy concerns, unmet expectations, fear of vendor lock-in and the need for more compute power. Whatever the reasons, organizations are relocating some workloads to on-premises and colocation data centers. For many enterprises, this is a move “back” to the data center. But for cloud native businesses whose applications were originally designed for cloud, it’s a move forward to a hybrid infrastructure model that balances costs, performance, flexibility and scalability.

Cloud repatriation, or “cloud rebalancing,” initiatives can be challenging for cloud natives since managing infrastructure is new territory. Some born-on-cloud companies, however, have become so large that they can’t afford to be exclusively in the cloud. As they reconsider their IT infrastructure investments, they’re exploring options. The goal is to find the right option for each workload, from on-premises to the cloud and everywhere in between.

Data center density is an important consideration for cloud rebalancing initiatives since it’s a big factor in infrastructure efficiency and costs. If you’re moving to an on-premises data center or into colocation, understanding data center density can help you optimize resources and ensure maximum performance and efficiency in your environment.

Data center density 101: What is it?

In the context of cloud rebalancing, repatriation or migration, “density” refers to the efficient utilization of resources, such as compute, storage and networking, within an environment. Data center density is often measured in terms of power consumption and cooling requirements.

Think of your data center like a high-rise apartment building where residents in every unit, on every floor, consume electricity and resources. The goal is to house as many residents as comfortably and resourcefully as possible. Likewise, in a data center, IT equipment consumes power and processing capability, and we aim to maximize the computing power and efficiency within a given space. Understanding data center density is like knowing how to best allocate apartments to residents, ensuring each floor is utilized to its fullest potential without overloading the building’s infrastructure.

Cloud native organizations didn’t have to think much about density when their infrastructure was exclusively in the cloud, since the cloud provider took care of it. But when moving to on-premises or colocation facilities, giving some attention to data center density can help these companies reduce costs and improve efficiencies.

5 key aspects of data center density

There are several ways to think about and measure data center density. Here are the most common ones:

1. Power density

Power density is what people most often think about when they hear “data center density.” It indicates how much power is consumed by equipment within a specific area. Higher power density means more power consumption per square foot. For example, a data center with a power density of 10 kW per square foot is much denser than one with 2 kW per square foot.

2. Compute density

Compute density refers to the number of computing resources, like CPUs, GPUs and storage devices, within a specific area. High compute density implies more powerful or abundant computing resources in the same physical space, which is critical for handling large-scale computations and big data analytics. AI, for example, has been driving up compute density for several years.

3. Rack density

Rack density refers to the amount of power used by equipment in a single server rack. While higher rack density can maximize the use of physical space by packing more equipment per rack, this approach isn’t always ideal. Often, strategic spacing of cabinets in a data center allows for better airflow, optimized cooling and more manageable power requirements. Having some built-in flexibility enables data centers to design solutions tailored to specific workloads and operational needs, ensuring that your infrastructure is not only dense where necessary but also scalable and efficient.

4. Cooling density

Cooling density is about the cooling capacity needed per unit of floor space. Higher density setups generate more heat, requiring diversified cooling solutions. Effective cooling is essential to maintain the optimal operating temperature of the equipment, ensuring the long-term reliability and longevity of gear.

5. Storage density

Storage density is the amount of data storage capacity within a given physical space. Abundant storage density means more data can be stored in a smaller area, which is particularly vital for organizations dealing with large volumes of data.

Why is data center density crucial to cloud rebalancing?

For organizations undergoing cloud rebalancing or repatriation, it’s important to optimize data center density to ensure that your infrastructure addresses all your business and technical requirements. There are several reasons why data center density is so important for orgs moving out of the cloud:

• Workload optimization: Increased density and strategic spacing allow for more effective workload placement, so you can put the right data and applications in the most appropriate location based on their requirements. In physical data centers, it’s crucial that we use the available space and power in a location efficiently since there are real physical constraints.
• Cost efficiency: Cost optimization is probably the biggest reason companies are moving certain workloads out of the cloud. Thus, when you balance density and flexible spacing well, you can reduce your total cost of ownership and avoid having to invest in new hardware too soon.
• Performance: Higher density deployments typically provide more compute resources in close proximity, so they can improve application performance and reduce latency.
• Energy efficiency and sustainability: Spacing cabinets appropriately can reduce energy consumption, contributing to sustainability goals and lowering operational costs.
• Management and maintenance: There are now advanced monitoring and management tools that can help maintain high-density environments in data centers effectively, thus reducing operational overhead.
• Cooling requirements: Effective cooling is essential to maintain the optimal operating temperature of IT equipment, ensuring reliability and longevity. Properly designed high-density environments can leverage innovative cooling techniques like liquid cooling to manage heat efficiently.
• Business continuity: By utilizing resources in a smaller space, organizations can build failover and backup systems within the same facility or nearby, ensuring a better mean time to recovery. This can help you better support disaster recovery and redundancy strategies.

Putting data center densities to work

Hopefully, this primer has shown you how data center densities can work to your advantage when you move workloads from public cloud to a private data center or colocation. If you’re set on moving to colocation or back on-premises to save money, the last thing you want to do is inadvertently add cost by drawing too much power. With the right infrastructure design, you can succeed in optimizing costs and performance with cloud rebalancing.

At Equinix, cloud native companies can migrate to colocation quickly with a business-ready cabinet that follows the most common cabinet specifications. This is just one of many approaches you can take to cloud repatriation. Because of the large ecosystem of clouds, SaaS providers, network service providers and other IT services on our platform, we make it easy to connect with your partners in all the right places using our virtual networking solutions.

Learn more about migrating to a high-performance data center by downloading Four key considerations when migrating to a High-Performance Data Center.

[1] Natalya Yezhkova, Assessing the Scale of Workload Repatriation: Insights from IDC’s Server and Storage Workloads Surveys, 1H23 and 2H23, IDC, Doc #US50903124, June 2024.