The short days of December and the promise of the new year inevitably present a time for reflection and future expectations. I’ve long done an annual piece on cloud predictions and am happy to share the lessons I’ve drawn from 2022 and what I expect to see in 2023.
Here are the things I think you should be paying attention to going forward, cloud-wise.
Cloud Providers Aren’t Innovating, and That’s a Good Thing
Within the analyst and commentariat there seems to be a feeling that cloud computing has settled into a rut. They seem to feel that, yes, cloud is a powerful trend, lots of companies are moving applications to the cloud, and AMG (Amazon, Microsoft, and Google) are, well, really big businesses, but…it’s gotten kind of boring.
The recent AWS Reinvent conference is a case in point. My Twitter stream was full of reactions to the keynotes that could be summarized as “meh — this is all just incremental improvements to what already exists.” There was a definite sense of nostalgia for the old Reinvents where AWS would announce unexpected innovations like Lambda serverless functions or a satellite base station service.
Fairly typical of this reaction to Reinvent was my long-time friend Brian Gracely’s CloudCast podcast, in which he shared that he felt that Reinvent wasn’t that exciting. As I understood the discussion, his theme is that the foundation of cloud computing is built out now and unlikely to show real innovation going forward. Again, not to say that there’s not lots of activity in incremental improvements and product extensions, but that cloud’s exciting days are past. The analyst group Forrester echoed that point of view with a blog post describing the extensions to core services that AWS announced at the show.
I don’t disagree with this perspective. There was nothing mind-blowing launched at Reinvent. I interpret this less-innovative time differently, though.
I think it’s good news that AMG aren’t launching fantastic new services left and right anymore. Because that means, as Brian pointed out, that the core capabilities of cloud computing are now in place and stable. In turn, that is great news for users.
Why? Because when the infrastructure into which one will place applications is transforming rapidly, it’s difficult to create reliable plans — after all, one might design an application architecture today and find out in six month’s time that it’s obsolete and needs to be reworked. That’s not fun and, more to the point, tends to bring an attitude of wait-and-see into play.
Why not just live with the way things are until one can be more confident that the architecture will stick and the application can be deployed with a lengthy lifespan expectation?
Now that the fundamentals of cloud computing are in place, users can feel more assurance about the stability of their cloud environment, and that’s a good thing.
Data Centers Empty as Users Climb the S Curve
The S curve is a well-established theory of technology adoption. Essentially, a new offering experiences slow growth over time as the market learns about it and evaluates how best to adopt it. Eventually the market reaches a critical mass of knowledge and the growth curve climbs steeply as more and more users adopt it. Finally, once most everyone who can benefit from the new offering has adopted it, the growth curve once again flattens out. Each of these curve slope changes is referred to as an inflection point — the point at which growth prospects change significantly.
While AMG experienced rapid growth for the past decade, from my perspective the industry is just entering or is early into the steep growth part of the curve. I say this despite having recently discussed their impressive quarterly financial results, which show plenty of current growth.
I maintain that AMG growth is going to increase because of my sense that many companies — large companies — are refiguring their previous strategy in which they planned to move some portion of their application portfolio to the cloud while keeping a large proportion on-prem. Now I’m hearing more about companies deciding to wholesale evacuate their data centers in favor of running everything in the cloud.
I think the reason for this is that these companies have concluded that owning their own data centers is a poor use of capital — data centers are not a differentiating capability and devoting expensive capital to owning them harms financial results and reduces future opportunity. After all, a dollar spent on a data center is a dollar that cannot be spent on streamlining a supply chain, improving manufacturing productivity, or launching into new markets.
The upshot of this trend is that data centers will empty to the benefit of AMG.
Containers and Cloud-Native: The New Must-have Accessory
Cloud computing originally offered up more easily provisioned virtual machine-based computing resources — thus the moniker “infrastructure as a service” (aka IaaS). This streamlined a lethargic part of the application value stream while leaving undisturbed other elements of the stream. Users saw an immediate benefit to their IT practices without having to disrupt their existing processes and tooling.
We’ve left those days well behind us. The emerging best practice for application development is centered on containers, which require fewer compute resources to operate, instantiate far more quickly, and are more portable than virtual machine-based compute environments.
That’s great. However, the shift to containers is a forcing function for wholesale change in application lifecycle processes and the tooling used to move code from development to production. I wrote about this shift last year in a piece I called Why Cloud Means Cloud-Native.
In that piece, I said:
“Over time, the cloud-native cohort has developed a set of best practices for lifecycle management, spanning the use of a sophisticated code management platform through to automated monitoring and management of application components to provide scale and resilience. Every process and milestone has been streamlined to provide fast, automated execution and enable touchless production placement once a developer’s fingers leave the keyboard.”
If IT organizations aspire to match the cloud-native cohort, it means they will need to completely transform their existing application lifecycle practices. I see many organizations struggle with this, as they look to incrementally improve piece parts of the life cycle rather than recognizing the need for a comprehensive transformation.
The upshot of all this is that cloud computing has gone far beyond IaaS and if an organization wants to meet the top performer benchmark it will require a focused organizational effort.
Cloud Providers Are Innovating, and That’s a Good Thing
Wait, wut? Bernard, you just got done telling us that cloud providers aren’t innovating, and that’s a good thing? Now you say they are? And that’s a good thing?
What’s going on here?
It’s true that the cloud providers have, by and large, built out their core services and have settled into ongoing incremental improvement in those services.
Nonetheless, there is a lot of innovation going on in the cloud providers. It’s just that innovation is built on and around the foundation of cloud computing.
Let me offer a couple of examples.
At Reinvent, AWS announced a new “omics” service aimed at genetics analysis. Omics is a pre-integrated platform to allow life sciences companies to import massive amounts of data. Omics automatically formats it into useful schemas, runs analytics on managed infrastructure, and stores important variant information relevant to disease analysis and new treatment development.
Omics even enables secure sharing to allow collaboration between disparate participants and organizations, crucial in an era where innovative treatments require cross-domain collaboration. The new service relieves users from the burden of a lot of complex set up, configuration, and software management, and allows them to get on with the job of improving our lives.
This demonstrates an appreciation of the detailed requirements of an entire industry and creation of a solution broadly applicable to many participants.
A more focused platform for innovation came with a Microsoft/London Stock Exchange announcement. The London stock exchange is making a very large long-term commitment to Azure, and the two organizations will collaborate on creating new solutions for the financial services industry based on a collection of Microsoft tools like PowerBI, Excel, and Azure Machine Learning.
The clue that this is something more, and heralds real innovation, is a discreet statement deep down in the blog post: Microsoft has purchased four percent of LSEG and installed its head of Azure on its board. That implies a much deeper collaboration and bespeaks a commitment to joint development of specialized tools designed to give LSEG competitive differentiation.
The way to view this innovation is as part of an inevitable upward move as the providers engage with customers to provide innovation at the frontier of what can be achieved based on the foundation of previously-absorbed core infrastructure innovation. We can expect to see plenty more of this type of innovation going forward.
In other words, you think the cloud providers have delivered a lot of innovation? You ain’t seen nothing yet.
The War for Talent Becomes the War for Market Share
Competition for cloud talent is an evergreen topic in the industry, with many observers bewailing the shortage of skilled personnel capable of building, yes, cloud-native systems.
Just how important that talent is can be understood by a paper Amazon (the ecommerce division, not the AWS cloud computing division) just published. The paper’s authors describe how they applied deep reinforcement learning to inventory management.
The net result? They improved inventory management 12% with no sacrifice in product availability. In the world of retail, where margins are pennies on the dollar, reducing inventory working capital requirements by that amount is a huge improvement.
If you’re a competitor to Amazon’s ecommerce division, now you have to figure out how you can respond. Obviously, you need access to massive amounts of compute, necessary for effective machine learning, which means hyperscale cloud computing. You need the technical staff to build the machine learning pipeline and data management. And, of course, you need machine learning specialists who can build and test ML models to optimize outcomes.
So table stakes is the talent pool to implement the technology stack and operations to meet the new benchmark of inventory management. (If you want to assign your company’s boffins to implement the Amazon systems, here is a link to the paper).
But technical staff isn’t enough. You also need to integrate the output of the inventory management upstream into your ordering processes and downstream into your physical warehouses. Even further upstream you need to have business managers monitoring the overall system to determine product mix and promotion mechanisms.
In other words, technical talent is necessary but not sufficient to meet the war for market share that a system based on technology along with integrated business processes can create.
Sounds hard, eh? Well, AWS, bless its heart, has decided to help. At Reinvent it announced a preview of AWS Supply Chain, which appears to implement an ML-based inventory management system.
Of course, taking full advantage of it will require treating the entire commerce value chain as an integrated system, and no cloud service can deliver that. That’s a business management problem. And the successful companies in the future will have to apply end-to-end value stream practices to win the war for market share. But its foundation is the right technical talent delivering cloud-native capabilities.
There you have it. My predictions for 2023. Summed up: the early foundation of cloud computing is complete and explosive growth of cloud infrastructure is directly ahead. Looking just beyond that are a set of innovative cloud services and business processes that build upon that foundation to create transformational economic outcomes.