Common reasons I heard for why cloud computing would never amount to much included:
Everything’s fine just the way things are. Why would we need to make a change?
Most applications can’t benefit from the cloud. They don’t change very often and have stable use patterns. So who needs the cloud?
Cloud costs too much. I can buy a server for a lot less than what a cloud provider charges for computing services.
And again and again: security. There’s no way a cloud provider can match internal IT for security of applications and data.
But I stuck with it, convinced that the value of immediate availability, low costs, and the absence of a need for long-term commitment would prove irresistible. And over time, this has been borne out.
Users have come to embrace cloud computing in large numbers, driven by the desire to move faster and save money. Today, everyone accepts that cloud computing will be a significant part of their application portfolio footprint. And those concerns about cost and security? They seem to have fallen by the wayside.
Based on my work and discussions with people across the industry, both vendors and users, here are five things I believe we can look forward to in 2018.
Digital transformation shifts into high gear
Most companies have launched digital transformation efforts. They may call them by that name or may refer to the initiatives as digital marketing, mobile-first, or multi-channel customer engagement.
Next year will be a time when digital transformation becomes a core focus of companies. The reason is simple: People want to interact via digital means. Smartphones have 81 percent market share of all mobile phones. Consumers are increasingly shopping online, turning the retail landscape into a nightmare. Streaming media is killing off existing offerings.
Next year will see this trend move from consumer-forward offerings to mainstream B2B interactions. Once you’ve experienced convenience in one part of your life and can see that it’s possible, say, to substitute a five-minute online task for a lengthy errand, you begin to expect that in all aspects of your daily existence.
Why does it take two weeks to get medical test results? Why can’t I get real-time insight into the status of my industrial machine order? Why is buying a house a chaotic process that requires physical interaction and is loaded with endless, ridiculous fees?
Every IT organization will be tasked with streamlining the way its company does business. And woe to those IT leaders who fail to step up to the challenge. According to McKinsey, impatient CEOs will find outside organizations that can help them become a digital-forward company.
Machine learning and predictive analytics leave winter behind
Artificial intelligence has had a rocky history. Heralded in the ‘50s and predicted to emerge within 10 years, it endured two different “winters,” as applying AI concepts proved harder than expected.
In large part, that was because insufficient computing power and storage existed to support AI algorithms. AI theory outstripped researchers’ available computing capacity needed to train and test AI systems.
Guess what? Cloud computing solves the capacity problem, big time.
As a result, we’ve entered a renaissance of AI, now commonly referred to as machine learning or predictive analytics.
One example? Speech recognition. I am a huge devotee of Google Voice and have used it for years. But even five years ago, it was amusing to post funny transcription messages on social media.
Today, it’s a completely different story. Transcriptions are very accurate; I can always understand the gist of the message, and messages are frequently word-perfect.
Those transcriptions are just one example of how voice recognition has improved in the past three or four years.
It is also just one example of the ways machine learning is applied to improve business processes. One of my favorite examples is how the Hong Kong public transport system schedules maintenance tasks based on machine learning.
Next year will see companies in every industry evaluate how machine learning can enhance their operations and improve operating results. And the technology will be applied in areas that might seem, at first glance, to be resistant to use of machine learning. Like sales. Or partner management. Or teaching.
Cloud growth will go vertical
Cloud computing is increasingly becoming the default infrastructure for enterprise computing tasks. Cloud adoption is driven not just by convenience and low cost. Increasingly, cloud computing is serving as an innovation platform, allowing companies to implement transformative applications that would be impossible to deliver on traditional infrastructure.
I estimate that the biggest providers are growing at around 60 percent per year. While some expect cloud’s growth rate to slow, I hold the opposite opinion. I believe that the growth rate is going to accelerate.
Because I align with noted disruptive innovation scholar Carlota Perez, whose framework posits two phases of each technology revolution. The first phase demonstrates the power of the technology, while the second phase witnesses the adoption of the technology into general usage.
Notwithstanding the sizable cloud computing market of today, we are about to see it move into a decade-long growth phase. As many people never tire of pointing out, cloud computing represents only a small portion of total IT spend. So it has a lot of room to grow.
Five years from now, we’ll recognize that 2018 represented the inflection point, where cloud computing went from an infrastructure adjunct to the primary form of computing.
Focus on the application — aka DevOps stat!
It is an engineering truism that removing one bottleneck exposes another, which must then be addressed to improve throughput.
In a cloud computing context, that means that changing infrastructure availability from months to minutes highlights the inefficient application life-cycle practices typical of most enterprise IT organizations.
While a large percentage of these organizations have adopted agile development practices, that hasn’t improved overall application delivery cadence.
That’s because everything after development remains mired in artifact rebuilds, dissimilar environments, and manual activities. This is the land of “I don’t understand why it isn’t working ... it worked on my machine.”
Next year, as cloud computing moves into the mainstream of IT activities (see previous prediction), it will confront these slothful practices. The pressure to improve overall application delivery times will build and build.
Whether you call this improvement process DevOps, reengineering, automate-all-the-things, or one of a dozen different names, the aim is the same: restructure the application development and delivery process to remove inefficiencies.
Next year will see application life-cycle restructuring become a senior IT leadership priority, as the need to improve overall throughput becomes a critical need.
Talent will becomes a critical success factor
Every one of the previous four predictions illustrates the enormousness of the upcoming change in IT and how each of them requires really talented staff for successful implementation.
Unfortunately, most enterprises continue to treat technical personnel as interchangeable cogs managed for cost control, instead of recognizing staff for what it is: the engine to power the shift to the digital enterprise.
Awareness of the critical nature of technical skills to pull off the kinds of changes described above will build throughout 2018. Many enterprises will come to appreciate that having the right skills on board will make the difference between future success and, well, failure.
Expect to see increased emphasis on staffing IT organizations with a range of cutting-edge skills relating to application design and operation, along with specific technologies like machine learning, IoT hardware, real-time event processing, and many others.
One can also expect groans as enterprise IT organizations confront the fact that leading-edge talent expects leading-edge salaries, and trying to wedge these people into existing salary structures won’t work.
As Clay Shirky noted in his nearly decade-old blog post on the newspaper industry, inflection points in technology are messy, hard to understand (not to say even recognize), and uncomfortable. With our natural human instinct to impose narrative on situations, we assume that tomorrow will resemble today, only a little bit different, instead of recognizing that we are in the middle of disruptive change.
As I stated earlier, cloud computing can be thought of as an innovation platform. In their recent book “Machine, Platform, Crowd,” Andrew McAfee and Erik Brynjolfsson begin the platform section of the book by noting just how much change has occurred in our society and economy over the past 20 years.
Just 20 years ago, most people talked via landline phones (which, of course, weren’t referred to as “landline,” since that was pretty much the only kind of phone one could have). If they wanted to listen to music, the only way to get access was to go to a store and buy a physical thing. News was delivered by physical things called newspapers or by a broadcast medium called television.
The Internet has unmoored that world, making information available across a network and getting rid of “thing-ness” in the process.
Cloud computing reflects the advancement of this process to computing itself, and the changes it brings to the use and application of computing are now reaching a disruption point. In my opinion, 2018 represents the inflection point in this process, and massive change looms for all of us.