Skip to main content

Data centers: Alive and well, and living at the edge

A survey of almost 900 IT practitioners and data center operators provides insight into how data centers are adapting, especially around edge computing, and the challenges they continue to face when it comes to staffing and diversity.

A recent Uptime Institute Data Center Industry Survey shows that data centers are rapidly adapting to the changing landscape in terms of how they are used and the services they deliver. The bad news is that outages or severe service degradation have increased over the past three years, according to half of the respondents, and of those problems, 80 percent were preventable.

Unsurprisingly, the majority of data centers are seeing their primary emphasis shift to hybrid IT operations. Making use of both private and public cloud infrastructures, data centers remain relevant because they become the pivot point for hybrid IT. They enable enterprises to cost-effectively deploy services and applications and provide the flexibility once thought available only from pure cloud deployments. The survey also indicates that hybrid IT adoption is the easy part of updating data centers. Future growth in data center operations will need a significant investment in data center processes and operations.

What's on the data center horizon

Perhaps the most obvious of the upcoming challenges is edge computing. More than 40 percent of respondents said they see the edge as the future of the enterprise environment and that computing will be handled by data centers focused on edge tasks that the businesses directly own or operate. Fewer than 15 percent see this as a task that can be completely outsourced to a public cloud provider. However, the nature of edge data centers remains unclear, with arguments being made for a range of models, including colocation, smaller edge data centers, micro data centers, solely owned and operated facilities, and shared services. But there is little doubt that edge services are a significant part of the future of the data center.

Another significant change is how data centers are being used. Previously, rack density was considered a major indicator of data center use. The introduction of cloud, artificial intelligence, and high-performance computing (HPC) services seems to have changed the focus. The energy consumption model that rack density indicated has bifurcated. Rather than a general increase in density, the survey indicates that the model is now split in two directions.

Data centers where a lot of hybrid IT and general-purpose computing workloads occur, the average rack density is starting to decrease. However, this is offset by the rack density increase in data centers with a focus on AI and HPC. Almost 20 percent of the respondents with these types of environments reported rack density exceeding 30 kilowatts per rack, with 5 percent noting they have rack densities of more than 50 kW per rack. That's almost double the rack density found in previously deployed high-availability data centers and more than five times the bottom third of data centers, which see densities of less than 10 kW.

Related reading: The evolving enterprise calculus of public cloud versus private infrastructure

Management tools are the future

The prevalence of data center infrastructure management (DCIM) tools is a key finding. While there was a significant failure rate in the successful adoption of these tools—more 35 percent reported—better than 50 percent saw them deployed across their environments. Also significant is the understanding that this isn’t going to be enough and that in the future, there will be a major need to find tools that allow management across the hybrid environment, reaching into the privately operated data center and out into the public cloud. These types of management tools will be critical as hybrid IT expands into edge computing.

Staffing and diversity remain a problem

Finding and hiring good staff is challenging. Almost 40 percent reported this as an ongoing issue despite the fact that a third of respondents reported past or pending staffing reductions. While a third said they have problems finding people to fill specific technical roles, the most common problem, cited by more than 50 percent of respondents, is finding solid candidates for the basics of data center staffing, operations, and management roles.

The report also shows a significant lack of diversity in data center staffing. More than half reported that women make up less than 6 percent of data center staff, and more than 70 percent don’t perceive this as a problem.

The need for additional skills in the data center was identified as important for business growth. However, the data center staff shortages will continue. Improving diversity can only result in a better candidate pool.

Data centers have become more energy efficient

Last is an issue data centers have long been disparaged for: energy usage. While its overall value as a metric is in some dispute, using power usage effectiveness (PUE) as a long-term measure of energy efficiency is an effective way of measuring how successful data centers are in reducing their energy footprint. Uptime Institute’s research shows that over the past decade, average PUE reported by its members has dropped from 2.5 to 1.58. This reflects the data center industry’s commitment to delivering services in a more cost-effective, environmentally friendly, and energy-efficient way.

This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.