Monty Python fans will instantly recognize the line, "I'm not dead yet!" as deriving from the (in)famous classic, "Monty Python and the Holy Grail." In one scene, we see a man insist another is dead - or will be in a moment. The allegedly 'almost dead' man argues vehemently that he isn't dead yet. To the contrary, he feels just fine.
This scene depicts, fairly accurately, predictions that the data center is dead - or at least on its death bed. The claim has been made many times since the ascendancy of cloud to the majority industry mindshare. And it continues to be wrong.
One of the reasons for this inaccurate prediction was the purposeful conflation of SaaS with IaaS. SaaS is the inevitable evolution of packaged software from build to buy to subscribe. Before they were declared "cloud," they were hosted by application service providers (ASP). Nothing really changed early on but the moniker. We can argue about how SaaS has grown into the cloud moniker, mostly by extending into what's really PaaS offerings, but in the early days there was little to commend SaaS as cloud.
Lumping adoption of SaaS in with IaaS caused a great deal of speculation that "cloud" was eating IT. Combining it with IaaS inflated expectations of cloud's adoption rates. Pundits warned that data centers would disappear. Predictions were made that no one would invest in building out their out data center facility anymore. It was cloud - all the way or bust.
But the reality is that data center facilities are still in operation. They are still being built and expanded and run by enterprises around the globe. The cloud hasn't managed to - and likely never will - kill the data center.
That's the conclusion of the latest Data Center survey from the Uptime Institute, in which it found that "the large, privately owned enterprise data center facility still forms the bedrock of corporate IT and is expected to be running half of all workloads in 2021." As for the cloud, "about a third of all workloads expected to be contracted to cloud, colocation, hosting and Software as a Service (SaaS) suppliers by 2021."
That's in-line with our own findings in the State of Application Services 2019, in which nearly half of respondents (46%) currently operate less than 25% of their applications in public (IaaS) cloud. Nearly one-third (31%) operate more than half their workloads on-premises, in a private cloud model. Even colocation performs better than public (IaaS) cloud, with 16% of respondents operating more than half their workloads in colocation data centers. Only 3% of respondents told us they were operating more than half their workloads in a public (IaaS) cloud.
To be fair, 7% of respondents were operating more than half their workloads in a private (off-premises) cloud. So technically they're likely using public (IaaS) cloud, but cordoning it off to keep out noisy (or nosy) neighbors.
Still, the numbers are anemic compared to predictions from the early days of cloud. And they may lose more in the near future.
Early in 2019, "an IDC executive told channel partners at the IGEL Disrupt conference that over 80 percent of companies surveyed by the analyst firm expect to repatriate public cloud workloads, and that 50 percent of those workloads could be repatriated."
Drivers for this decision are numerous and include the common culprits of security, visibility, and performance. Factors that enable repatriation include improving availability of multi-cloud operational tools and an increasing move toward application architectures that rely on more portable technology such as containers. It's easier to move containerized (cloud-native) applications from one cloud to another, after all, which gives organizations the flexibility to focus on the application and its needs to determine where it is best operated.
Containerization and cloud-native architectures, too, offer the efficiencies that were once found only in the public cloud. Its elasticity and variable use of resources made it possible to operate an application much more efficiently. Which is what containerization offers today and is likely why architectures relying on the technology are cropping up everywhere to dominate the app development landscape. Because cloud-native / containerized apps offer the same benefits without leaving the data center.
The reality is - and will continue to be - that the data center is still very much alive and feels just fine.