The Death of the Data Center Is Greatly Exaggerated

Every year… check that, it’s almost every day we hear something that says it will be the next best thing. Huge campaigns are created and flashy ads and commercials tout how it will eliminate the need for x or y. Now I could be talking about IT or the auto industry or any number of other industries. But for this blog, I will be using auto industry analogies to discuss the data center and the requirements for private enterprise data center ownership for the foreseeable future.

You may say, just one second there old timer, but bear with me. Even in Texas and Colorado where I split my time, ranchers still use horses when the automotive and ATV industry cannot fill the need. And much the same, we still have legacy mainframe, Unix and Windows, all which were supposed to have been end of life by the latest and greatest at some point in the last 20 years.

Zettabytes of Data Traffic

Now I am not saying there will not be a change in how we do business, but what I am saying is that the complex nature of the data center as we know it is going to continually increase, as digital business growth pumps more data through enterprise systems. And that makes for some interesting traffic on the information highway.

First, let’s explore this by examining some recent research. A data center study by AFCOM in 2016 had some interesting statistics:

  • Annual global IP traffic will reach 3.3 zettabytes by 2021
  • Global IP traffic will increase nearly threefold over the next five years

You will find several other vendors out there who have predicted roughly the same or are using this same information. It’s true, most likely, and another thing that is predicted is the majority of that traffic will be generated by non-wired devices.

Enterprise Data Centers Becoming a Foundation for Many Cloud Services

We know non-wired devices will comprise most of the traffic just by looking at the statistics on social media accounts or the number of people staring at their smartphones. It’s tempting to think that with all this power in our pockets, is there still a need for back-end data systems?

At first glance, you could almost deduce that there is good reason to support the claim that data centers will go away, but no, it is just the opposite. Eventually all that information pumped to the devices has to be stored and processed somewhere. Your smartphone just doesn’t have that capacity… yet. Just like my local grocer can’t supply all its needs on its own, nor the local gas station; large truck loads of supplies arrive daily. Yes, there are local trucks and individual local providers but the analogy here is it takes a lot of different contributors and a huge, largely unseen supply chain to provide the whole experience.

How do I know this?

When there is an outage on Amazon, Facebook, your company’s web portal—you pick the medium—it doesn’t get blamed on the device in your hand. It was a data center, or component of, that failed somewhere.

Enterprise Data Center Investments on the Rise

Now you might say, yes, but with the rise of these public cloud providers, aren’t they the ones who are providing the back-end data centers instead of enterprise data centers?

Well, let’s look at a few more statistics from that AFCOM study:

  • 58% of the respondents owned between 2 and 9 data centers
  • 19% owned more than 10
  • Respondents also stated that roughly 5.3 of their data centers will be renovated
  • The 3-year forecast showed that they plan to increase number of data centers

This data suggests that enterprises are continuing to invest in their data centers and even increase them, instead of reducing them.

These are not small data centers either, with the study showing that 48% of respondents said they are between 5k and 50k square feet, and 16% said they were between 100k and 500k square feet. These same respondents were made up of entities who had their own facilities, used colocation data center providers and private and public cloud.

Private Cloud Keeps Growing

Let’s remember that private cloud is still in the enterprise data center. IDC predicts that private clouds will keep growing. While some of this growth comes as traditional application environments and are superseded by internal private clouds, these clouds also will support many new digital business applications.

Cloud IT Infrastructure Revenues Surpassed Traditional IT Infrastructure Revenues for the First Time in the Third Quarter of 2018

Additionally, some workloads are being re-patriated from public to private clouds. How big is this trend? One analyst firm estimates that nearly 50% of all entities who have moved workloads to public clouds have made efforts to move workloads back to their enterprise data centers due to cost, security or latency.

Diverse Workloads are Here to Stay

Another twist to this plot is that some cloud providers, notably Amazon and Microsoft, have announced on-premises public cloud, which will increase capacity requirements of enterprise data centers. Even the public cloud providers find benefits from locating in enterprise data centers.

It’s hard to escape the conclusion that enterprise data centers will be around for a long time.  However, they will have a broader and more diverse set of workloads, including private and public clouds, as well as traditional systems.

To correlate the longevity of the enterprise data center to our automotive analogy, some entities have their own car, some rent, and some use other transport and ride-hailing services, such as Uber. Some have pickup trucks, and some have semis (owned or leased). However, this growing diversity in transport vehicles and consumption options has not resulted in fewer vehicles being sold nor in less traffic and congestion. There were some optimistic assumptions that the rapid rise of ride-hailing services would mean that private car ownership would decline or perhaps disappear altogether, but this hasn’t happened yet, even in some of the most tech-savvy and traffic-dense cities with a range of transport options.

Defying earlier predictions, Harvard Square continues to see growth in the number of cars despite 6 subway stations, and several lines and bike lanes. (Photo credit: Juliana DiLuca Photography)

A recent Boston Globe article cites the continued growth in automobiles in Cambridge, Massachusetts, the home of Harvard and MIT, defying earlier predictions, as the city continues to grow, despite having 6 subway stations, many bus lines and bike lanes. Interestingly, this growth in Cambridge is largely due to technology innovation in such industries as biotech and software. The point here is that just as no single transport model has emerged, neither will a single computing model by a panacea for IT organizations. Diverse enterprise data centers will increasingly continue to play a key role in multi-cloud strategies for most IT organizations.

Perhaps a better way to look at this is to think less about the physical location (enterprise data centers vs. public cloud) and more about how workloads are structured and consumed (physical or virtual assets tied to particular applications vs. public or private cloud services shared across multiple applications). I’ll explore this more in another, upcoming blog.

Is Data Center Investment Worth it?

With all the cloud talk, you might think that would mean a decline in data center investment. On the contrary, from a facilities standpoint, there’s a lot of new investment in data centers. According to a study by real estate firm CBRE, which was focused on tracking the investments of the actual buildings and facilities, 2017 was a banner year for US data center investments, with nearly $18.2 billion being spent. In most cases, the life span of a data center is typically 10 years at the top end.

A commissioned study conducted by Forrester Consulting on behalf of Dell Technologies, focused on the data center infrastructure itself, analyzed how long it takes for the investments to pay off. Forrester concludes that the payback period of 3.5 years and benefits seen at the 5-year mark are worth the investment.

Source: The Total Economic Impact of Dell Data Center Modernization and Migration Services

This still gives us a few years to watch this ever increasingly complex organism mature into what it started as many years ago, when the web exploded, and smart devices flooded the market. By the way, one of the reasons cited for this explosion in growth was the requirement to store data for self-driving cars.

Summary: The Data Center Isn’t Dead, it’s Becoming Repurposed and More Powerful

Will we continue to make advances in the data center space? Absolutely. Will that landscape change? Absolutely. Just like in the automotive space. But just like in the automotive space, new methods and technologies get added, but the base still stays. And so, it will be with data centers. After all, there are, according to the US government, nearly 3 million data centers in the US alone, so transitioning away from them would require a massive shift that will take years.

By any reckoning, enterprise data centers are here to stay. As Mark Twain famously quipped, reports of his death have been greatly exaggerated. So, treat reports of the death of the data center with similar caution.

To learn more about our perspectives on Data Center services and how we address these challenges, including whitepapers, descriptions of our tools and services, visit our website here.

About the Author: Ted Streck

An Experienced Manager with over 30 years industry experience delivering consulting services and IT solutions. Prior to joining Dell Technologies, Ted was a Vice President at a national ISP in charge of their IT Systems and operations. Prior to that Ted was the Vice President of a software development organization, where he was responsible for a 30 million dollar subsidiary. Additionally Ted was part of an acquisition team that acquired over a dozen smaller companies in the course of 25 months. At Dell EMC, Ted has successfully delivered projects that include Business Continuity/ Disaster Recovery strategy development, business requirements analysis and systems and application requirements analysis. Ted is experienced in operational management as well as project management and IT infrastructure strategy, implementation and support. Ted has considerable experience in Data Center Strategy, Server and Storage Consolidation and the migration of legacy technology, processes and organizations to new models. Ted is one of the original authors of the Dell EMC data center migration methodology.