Is the end of the traditional data center in sight?


We all know that the cloud is growing at the expense of server rooms and data centers. But Cisco predicts in its Cisco Global Cloud Index that by 2021 “hyperscale data centers” will represent 53 percent of all installed data center servers. For hyperscale, read cloud data centers and you see the future as Cisco sees it.

But what is a hyperscale data center anyway? According to CIsco, it  must have at least one of the following:  

  • More than billion dollars in Infrastructure-as-a-Service (IaaS) or Platform-as-a-Service (PaaS). Here we find Amazon Web Services (AWS), Rackspace, and IBM.
  • $2 billion in Software-as-a-Service (SaaS). This includes services such as Salesforce, Google Docs and Microsoft Office 365.
  • $4 billion in search or social networking, which brings Google, Yahoo and Facebook in play.
  • Or, finally, $8 billion in e-commerce/payment processing, which is where we find Alibaba, Amazon, and eBay.

Of course, most companies aren’t Internet giants, but looking ahead, many businesses will find more and more of their workloads moving to the big players’ data centers, or, more precisely, their clouds. Cisco estimates by 2021, 94 percent of workloads and compute instances will be processed by cloud data centers; 6 percent will be processed by traditional data centers.

These giant-sized data centers differ from their predecessors in other ways besides sheer size, according to Cisco: “Cloud data centers offer increased performance, higher capacity and greater ease of management compared with traditional data centers. Virtualization serves as a catalyst for hardware and software consolidation, greater automation, and an integrated security approach.”

These changes come from the use of not just virtualization, but virtualization transition into containers. At the same time, managing these more advanced data centers is moving from DevOps tools to Kubernetes cloud container orchestration.  

In other words, it’s not just more processors, it’s that the biggest cloud players are making the most of their hardware by using virtualization, containers and DevOps/Kubernetes management. Or, as Cisco puts it, “The workload and compute instance density (that is, workloads and compute instances per physical server) for cloud data centers was 8.8 in 2016 and will grow to 13.2 by 2021. Comparatively, for traditional data centers, workload and compute instance density was 2.4 in 2016 and will grow to 3.8 by 2021.”

Those last numbers are the real killer for the traditional data center. I recently had a young IT staffer look at me in amazement when I said many companies wanted to have their own physical servers at a data center or co-location facility. Older IT staff may still think that way, but the bottom line numbers don’t support this approach.

It doesn’t have to be that way. Smaller data centers need to look to software-defined services, DevOps, containers and container management to make the most of their existing infrastructure. If they adopt to the newest ways of computing, they can continue to do well.

Still, per Cisco’s count, by 2021, 73 percent of the cloud workloads and compute instances will be in public cloud data centers, up from 58 percent in 2016, and 27 percent of the cloud workloads and compute instances will be in private cloud data centers, down from 42 percent in 2016.

I believe the private and hybrid cloud data models have a lot of life left in them. In particular, I think cloud bursting — a hybrid cloud approach where daily computing requirements are handled by a private cloud, but when demand spikes, the additional traffic demand, the burst, is handled by a public cloud — will play a large role in SMB cloud use.

What I can’t argue with Cisco about is that the traditional data center or co-location company is on the way out. The jokes about cloud computing being just someone’s else’s computers have worn thin. Tomorrow belongs to the cloud.


  1. Very interesting article about the nearest future, thanks.

Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.