Datacenter Trends

Datacenter Predictions: AI Moves In and Cloud Moves to the Edge

The datacenter of 2020 and beyond will look quite different from the datacenter of today, according to researchers at IDC.

• By John K. Waters
• 11/09/2018

The future of the datacenter is fraught with change driven by a range of emerging trends, from the continuing advance of digital transformation (DX) efforts and the race to innovate, to an increasing focus on standardization and new levels of facilities automation.

Those drivers (and a few more) are behind a new set of predictions offered in a recent edition of IDC's FutureScape Web conference series, "Worldwide Datacenter 2019 Predictions." Rick Villars, vice president of IDC's Datacenter and Cloud group, and Jennifer Cooke, research director of the firm's Datacenter Management group, presented 10 predictions (listed at the end of this article) during the webcast, and laid out the factors that led to those predictions.

"IDC believes firmly that digital transformation is the leading driver of what's shaping all decisions around technology, and that includes the datacenter," Villars said. "In that same contest, the race to innovate...is one of the key elements of DX. The pace of transformation of the business is accelerating...which means the underlying systems need to support that. The amount of time you have to deliver new capabilities and respond to new functions is much shorter."

Villars also cited increasing tech standardization and a growing focus on platforms as key drivers of datacenter change. Add to that the spreading implementation of sensing and automation in environments, which impact datacenter workloads and capabilities; new AI-enabled, self-managed operations; and the "deluge of data" from the Internet of Things (IoT).

And then, there are new risk factors to consider.

"In this new world," Villars said, "we know data is increasingly critical, but concerns about how it's being used and delivered -- who can access it, how it's being manipulated -- is a growing challenge that is shaping risk. And we continue to deal with the risk of inertia, the technical debt that has been built up in the datacenter over the years, and how that prevents us from really taking advantage of new things."

Villars and Cooke drilled down into four of the firm's predictions that they considered critical to datacenter operators. The first is an answer to the question, "How will companies respond to that digital deluge and greater demands for data access from more and more people and devices?" The answer, said Villars, is with "trusted data bunkers."

"One of the key challenges we hear about when we talk to enterprises about their digital transformation efforts is the explosion of data they're dealing with," he said. That explosion of data, which has been coming mainly from mobile devices and the cloud, will increasingly come from sensors in cities, factories and hospitals. In the near term, this won't just be about collection or caching, but local processing and cleansing and real-time reaction at the edge. Using these data bunkers as staging points could provide real economic and performance benefits, he said.

"Any time you are evaluating a colocation partner, ask if they offer these kinds of complementary data services," he said.

Because of the expansion of where data and applications reside, the next big challenge is connectivity, Villars said. Consequently, datacenter leaders are investing time and effort into developing a much more dynamic connectivity environment. In fact, over the next three years, 70 percent of enterprises are likely to adopt "dynamic software-defined branch and network solutions that deliver security and flexibility across cloud, datacenter and edge interactions," according to Villars.

"Connectivity is going to be a bigger part of the investment portfolio," he said. IDC recommends that CIOs and IT leaders develop investment strategies to support next-gen workloads, stay informed about offerings that speed deployment, and choose colocation partners investing in software-defined networking (SDN) and virtual network functions (VNF) to address challenges.

"This is a very dynamic market right now," he said, adding that the launch of new services and capabilities is much faster than it has been a decade. "So pay attention to innovation and invest in solutions that allow you to take advantage of it," he said.

A third prediction, the advent of autonomous facilities, is a direct result of these trends, said Cooke. The datacenter is being asked to do more -- a lot more -- because of the increasingly distributed nature of IT and the need to support IT workloads at edge locations, but also because of the advent of the compute-intensive workloads generated by AI, augmented reality and virtual reality.

But while compute-intensive technologies such as AI are putting new pressure on datacenters for increased levels of support, they're also offering capabilities that datacenter operators can exploit to provide that support.

"Let's turn the lens around and think about how AI is impacting the way we operate datacenters, and the opportunities for transformation if we shift datacenter management from existing processes and expertise of staff...to instead leveraging massive amounts of data to become more proactive and predictive," Cooke said. "We believe that AI will allow datacenters to run with minimal human intervention and control."

By 2022, 50 percent of IT assets in enterprises' datacenters will have the ability to run autonomously through the use of embedded AI functionality that leverages smart IT and facilities systems, Cooke said.

"Our environments are becoming very complex," she said, "which requires a shift away from gut-feel or human decisions to more data-driven decisions. And there's a lot of data available within our datacenters today that we're not leveraging."

AI and autonomous operations will change the kinds of skills that are sought and valued within IT organizations and from colocation partners, Cooke added. "IT staff will not only be leveraging new tools," she said, "but will need a whole new toolbox to accomplish their job."

IDC recommends investing in trusted platforms to "pave the way for remote management," and to seek areas to test greater reliance on AI and autonomous operations.

The fourth prediction Cooke and Villars drilled down on is the emergence of new private clouds. "This is one of the fundamental changes in the private cloud conversation, which has been very data-centric in the past," Villars said. "But increasingly, 'private cloud' is more of an edge conversation."

After 2020, IDC predicts that "most enterprises will shift internal investments to deploying dedicated cloud platforms from which to launch innovative services, rather than investing in modernizing existing datacenters."

"Over the next two years, a lot of investment in private clouds will be focused on modernizing existing datacenters, but after 2020, most of that investment is going to shift to delivering innovative new services in edge locations," Villars said. "We are bringing cloud to edge...This is the next big step that we see."