Experts Predict The End Is Nigh For Enterprise Data Centers, Long Live The Hyperscaler
A recent blog post by Gartner analyst Dave Cappuccio offers a startling prediction: by 2025, 80% of enterprises will have shut down their traditional data center, vs. 10% that have already shut down data centers to date. Cappuccio backs that prediction by identifying the evolutionary changes in how workloads are processed on business needs and not physical constraints. In other words, the drive towards digital transformation is shifting processing priorities to the cloud, making the traditional data center less relevant for conducting business.
The rapid demise of the enterprise data center directly correlates to the rapid growth of the cloud, where many workloads are migrating to and shifting where compute is being done. This effectively limits the importance of the traditional data center. However, that shift to the cloud comes with some unique challenges that many organizations are struggling to resolve. For example, cloud service providers (CSPs), or hyperscalers, are encountering an exponential increase in data generation from their clients, which in turn is forcing CSPs to grow their infrastructure at unprecedented rates.
Resent research indicates that 68% of cloud services delivered today are being delivered by hyperscale providers. With 2.5 quintillion bytes of data being created every day and Gartner estimating over 20.4 billion Internet of Things (IoT) devices to be deployed by 2020, data generation will only accelerate. As this data moves to the cloud, hyperscalers will look to the data center industry to provide world-class infrastructure at an unprecedented pace.
PLENTY OF ROOM FOR GROWTH
While the demise of the enterprise data center as we know it remains debatable, it no longer remains a question of “if,” but when. Those purporting a rapid demise are relying on cloud adoption indicators and workload shifts as evidence, yet there are still numerous unknowns in the world of enterprise IT that may lend relevance to traditional enterprise data centers. For example, privacy requirements and compliance regulations such as GDPR and HIPAA may have a lasting impact on data storage and workload processing, impacting enterprise’s decisions to move line of business workloads into the cloud.
That said, certain infrastructures and workloads are best suited for cloud migration, reducing costs, enhancing agility, and simplifying the deployment of mission critical applications for the enterprise. For example, clouds can more easily address issues of scale, and are equipped to scale up or down on demand. What’s more, they can better provide distributed services and incorporate business continuity solutions. The distributed nature of hyperscalers also brings forth regional solutions, where workloads can be localized to reduce latency and provide an improved user experience.
It is those benefits and countless others that make moving workloads from an enterprise data center to the cloud appealing to enterprises, thus creating opportunities for both hyperscalers and data center providers — as long as they adapt their build and deployment models to keep pace.
THE GREAT NORTH AMERICAN MIGRATION
The migration of enterprise workloads to the cloud has become rampant worldwide and is accelerating. Nowhere is this more true than in North America, where Goldstein research is forecasting a compound annual growth rate (CAGR) of 16% until 2024. Goldstein also reports that colocation and hybrid services demand is continuously growing in North America, with more than 60% of enterprises estimated to increase their implementation of cloud hosting data centers and colocation in the coming few years.
Cloud adoption growth, additionally fueled by digital transformation, has forced hyperscalers to up their game and embrace new ideologies and technologies that can help sustain the ability to meet rapidly growing workloads. That, in turn, has led many hyperscalers to look beyond traditional geographical locations and choose alternatives. Case in point: Montréal, which has become home to many hyperscalers. Many of these hyperscalers service Canadian companies, which have become more focused on retaining data within national borders.
Montréal become a destination for some of the world’s largest internet companies and one of the best choices for data centers in North America thanks to a number of factors. For example, Canada’s consumer-focused privacy legislation, which prioritizes citizens and companies of every size rather than protecting the interests of big businesses and the state. Many Canadian companies have also been prompted to keep data and compute in-country by concerns related to revelations of NSA surveillance programs, and are now focused more than ever on data privacy, as well as security. Furthermore, Montréal has an ample supply of inexpensive hydroelectric power and local government incentives and support. It offers ready access to green technologies, low-cost energy, and room for growth, allowing Montréal to displace other locations that have become heavily saturated and difficult to expand.
In the last year, Google, Amazon Web Services (AWS), and Microsoft announced decisions to select the province of Québec for their cloud operations. Operators like these global leaders are bringing their data center operations to Québec for the sustainability benefits, strong privacy laws, and low cost compared to other major data center regions.
THE HYPERSCALER EQUATION
Hyperscalers have become the go-to providers for larger enterprises moving workloads to the cloud. However, hyperscalers are facing increasing pressures to lower costs, increase scale, meet shifting demands, and take on new workloads. These requirements have forced them to implement the right infrastructure to support exponential growth. Key considerations for selecting infrastructure providers include data center deployment time, design flexibility and innovation, scale, low power cost, and more.
For example, data centers that serve hyperscalers will have to consider design changes that support rapid deployments. Design changes must incorporate modular approaches which means a shift in how cabinets, floors, power distribution, cooling, and physical security elements are designed.
Each of those potential requirements introduces operational challenges, and hyperscalers may find it difficult to go it alone. Partnering with colocation providers may prove to be the most efficient methodology to embrace new ideologies and mitigate the problems of this rapid growth.
THE SHORT TERM
There is no denying that data creation is on an exponential growth trajectory, fed by factors ranging from industrial Internet of Things (IoT), to streaming data, to connected vehicles. Adding to that growth are the mounting expectations of today’s consumers, relying on technology to be available anywhere and at any time, while also offering a great user experience.
Hyperscalers will have to remain at the top of their game to meet predicted growth; failure to do so may hamper growth both short- and long-term. What’s more, new technologies, such as augmented reality, AI-driven analytics, and Big Data processing, may further increase the growth potential of the cloud.