Cloud computing: An expert’s view on tomorrows hottest topic

Cloud computing has come a long, long way. I first began working with cloud computing in 2006, when Amazon Web Services launched what it called Infrastructure-as-a-Service; the phrase cloud computing hadn’t even been coined then.

From my first exposure to AWS, I recognized that cloud computing would become the next major IT platform, with profound implications for the entire technology industry, vendors and users alike.

The manifest benefits of cloud computing — immediate resource access, easy scalability, usage-based pricing — made traditional IT practices obsolete overnight.

However, that didn’t mean adoption occurred overnight. Far from it. Skepticism borne of doubt regarding cloud security, reliability, and cost advantages hindered adoption. And that went on for years.

But it’s all changed now. Cloud computing has not only reached but passed a tipping point; it’s clearly on the path to platform dominance I envisioned twelve years ago. Which means we’re getting to the interesting stage of cloud computing — what happens when cloud computing is the default choice for application deployment.

Next year promises to be a banner year for cloud computing. You can expect to see significant developments in the cloud ecosystem; I’ve captured what you’ll see as the biggest topics in 2019.

You’re probably saying, well, duh. If cloud computing has passed its tipping point, if it’s now the default deployment choice, of course enterprise IT will make it a strategic priority.

True.

But there’s a whole list of interesting implications associated with that. CIOs will direct their staffs to deliver cloud-based yearly goals. Procurement staffs will shift budget from traditional legacy vendors to cloud alternatives. Architecture groups will develop new cloud-based designs that incorporate both core infrastructure and higher-level managed services. Corporate training will develop new curriculum to build workforce skills. Legal groups will learn the ins and outs of CSP contracts. What’s more, enterprise IT will require focus and attention at all levels from the Board of Directors down to the developers, with more companies likely to list “cloud computing” as a material risk in their 10-Ks, the annual filing submitted by publicly traded companies to the Securities and Exchange Commission.

It’s not overstating it to say that 2019 will witness a shakeup in enterprise IT business-as-usual as most shops digest what it means to make a wholesale platform shift to the cloud.

Well, if cloud is now the default deployment platform, and enterprise procurement groups are shifting budget to cloud providers, what will that mean?

A lot of money. Flowing into the coffers of cloud providers. And in the US, that really means into the coffers of three companies: Amazon, Microsoft, and Google, which I dub AMG.

The three providers have had scorching numbers for the past few years, and 2019 will be no exception. AWS’s growth rate, which I regard as the most important financial number in the tech industry will remain above 40% throughout the year.

If you don’t pay attention closely, what this means is easy to overlook. AWS is likely to exit 2019 at a $35 billion run rate, growing at 40%+ per year. And Microsoft and Google, albeit on smaller numbers, will be growing at even faster rates. It’s no exaggeration to say that AMG are poised to become thedominant presence in the tech vendor space. The implications of this fact will play out over the next few years, but I don’t think it takes a genius to connect the dots. Additionally, AMG are expanding into areas previously unassociated with cloud computing by challenging other tech companies beyond the traditional “platform” companies (e.g., mainframe, server, database).

SDLC (Software Development Lifecycle) is a venerable term in IT and, in my experience, one that has fallen out of favor, replaced by CI/CD or DevOps. I was somewhat surprised, then, when I heard it used in a couple of different advanced development tools conferences over the past month.

I asked the organizer of one of these conferences (NewOps) why the term surfaced at his event. His explanation: too much of the recent past has focused on developers, without looking at longer-term application operability.

I welcome the comeback of SDLC, as I think it captures the totality of what it takes to bring a software release from inspiration to production, with an emphasis on the entire process rather than one stage of it.

You can expect to see SDLC emerge as a big issue in 2019. IT shops will complete their initial move to the cloud and then confront an unpalatable truth: cloud computing doesn’t mean fast SDLC. In fact, using cloud highlights inefficiencies in the application development value chain and acts as a forcing function to improve each stage along with inter-stage handoffs.

Part of using cloud computing and streamlining the SDLC is restructuring application architectures, both in terms of execution environments and architecture.

In terms of execution environments, containers provide benefits over traditional virtual machines — faster launch times, easier updates, and greater portability.

Commonly going hand-in-hand with the move to containers is adopting microservice architectures — functionally decomposing monolithic applications from single large binaries to small, single function components that communicate via APIs. Microservice applications allow easier scalability and facilitate more rapid application evolution.

The only thing is, coordinating all those containers, making sure they’re installed properly, have the correct security implemented, can communicate across the network, is pretty hard.

Which leads us to Kubernetes. Kubernetes enables declarative descriptions of applications, including access controls, automatic scaling, and network connectivity.

I’ve heard many describe Kubernetes as making container-based applications easy to deploy and manage. Enthusiasm for Kubernetes grew to epidemic-like proportions over the past couple of years.

Next year will bring home a truth: While Kubernetes-based applications may be easy to run, Kubernetes itself is no picnic to operate. It is an early stage product which is serving as the foundation of an ecosystem of supplementary products (e.g., Istio) all of which must be installed, configured, and managed correctly to make applications run properly.

In a recent podcast, I heard one participant state that with Kubernetes and Istio, traffic would just magically get routed to the right resource, without noting that someone would have to work really, really hard to make that magic happen.

Next year will be a “back to the future” time as IT organizations (re)learn that managing infrastructure for cloud-native applications is not easy. It’s hard.

However, the overall sense that containers and Kubernetes is a natural foundation for cloud-native applications is not misguided. So expect to see a further enthusiasm for consuming Kubernetes from one of the big cloud providers; they specialize in running complex infrastructure at scale, and most IT organizations will conclude they should let the specialists take over.

As I hope these predictions make clear, we are in the midst of an IT transformation. Every part of IT is changing as part of a drive to deliver applications better suited to the digital economy. This will require a major upgrade in talent, as organizations seek out the skills needed to build the more complex applications typical of cloud computing.

One approach, of course — and a good one — is to improve the skills of people already on staff. They’re smart, they’re motivated, and they have lots of domain knowledge of a company and its industry. I always advocate education as the right first step.

However, most companies will also have to step into the recruiting fray to hire from the outside. There’s just no way to meet the demand for cloud-native skills solely from existing staff.

And those companies will find this kind of talent is in short supply. Next year, you can expect to see many articles about how hard it is to find cloud-capable candidates, and how expensive they are.

Get used to it. The imbalance of cloud talent supply and demand will be constant over the next few years and will result in a salary increase spiral.

If ever you thought of IT as a cost center to be managed through budget squeezes and infrequent salary increases, banish those thoughts.

Technology adoption commonly follows what is referred to as an “S-curve.” Early on, adoption of a new technology is slow, as people struggle to learn about it and envision how it can be applied. Once a critical mass of proof points and knowledge is hit, adoption accelerates dramatically.

Cloud computing has now hit the steep part of its s-curve. It is now the de facto platform for new applications, and, increasingly, legacy applications are migrating to the cloud.

Next year will see the implications of this shift hit home, as different aspects of IT incorporate cloud computing and change their established practices. There will be much more discussion about the best way to take advantage of the cloud’s immediate resource access, easy scalability, usage-based pricing, and much less about its supposed drawbacks and shortcomings.

Get ready for a very, very interesting year.

Related:

• LISTEN — Point-Counterpoint, Going All-In On the Cloud
• How Cloud and Ops Teams Can Take Cloud Adoption From Bottleneck to Engineering Success
• Going Public on the Cloud