The Cloud-only Era is Done

As ever in this job, it’s by putting multiple conversations together that I start building a picture. Over recent months, I’ve spoken to big enterprises and small businesses, global providers and small startups, new faces and old colleagues. It’s a privilege to spend time making sense of it all. 

Even better is that it affords the opportunity to test hypotheses. As I was listening to the radio last week, Young Ones comedian Ade Edmondson was talking about his days on the live circuit. Above all, he said, it allowed him the opportunity to find out what was funny, what his audience hooked into. 

Perhaps being at The Comedy Club isn’t that different to being an analyst, even if the jokes aren’t so funny. From the vantage point of being a panel guest at a recent CIO event, I found a room abuzz with today’s challenges and realities — money is no longer free, complexity is rife oh, and cloud is no longer the thing.

These concerns are 100% linked. My hypothesis was, and remains, that the biggest problem with cloud is nothing to do with technology per se; it’s more (just like a real cloud) about its lack of hard edges. All who remember the advent of VM sprawl (and previous incarnations) now recognise it is similar to cloud. 

Cloud versus on-prem is a bit like self-publishing. The publishing industry may have been old, stuffy, hard to get into, and so on, but it acted as a filtering mechanism for readers. Take away the boundaries, and of course, you open the door to a great deal more creativity, and more of everything possible. 

There’s a place for hard boundaries: just as necessity is the mother of invention, so constraint begets thinking harder about prioritization, controls, quality and so on. Honestly, there’s a place for both — I have self-published, and it’s a great model. But it, like cloud, comes with consequences that can’t just be brushed under the carpet. 

To be clear, this isn’t an anti-cloud rant. I deliberately said, “Cloud is no longer the thing,” but it is absolutely a thing. A very important thing. The sea change probably started a year or so ago but was rubber-stamped by the cost of money. No particular bone to pick, more an observation that I have tested and honed across dialogues. 

I’m more interested in the consequences. A few days ago, I spent time with NetApp and some of their customers, talking about all things tech. What became obvious was just how profound this shift from “the thing” to “a thing” actually is, not just for providers and customers but the entire ecosystem of tech supply, deployment and operations. 

We can (and do, with abandon) throw in terms like “hybrid multi-cloud,” but what’s becoming clear is the impact this has on the bit in the middle. If you will be running AWS, Azure and GCP alongside your on-premise or hosted applications, who manages all that? How do you architect, procure, secure, run, break/fix/replace?

The answer, frankly, is unclear right now. We moved from the early days of more open cloud and SaaS architectures to more walled garden approaches (remember when you could cross-post across nascent social media sites). And the same applies to the hyperscalers, who are not seeing the meteoric growth they once were. 

If I could put my money anywhere, it would be on tech providers that work horizontally across both cloud and on-prem. This is a curate’s egg — consider the concerns from enterprises around Broadcom’s planned acquisition of VMware, or the open-source controversy about changes to Red Hat’s licensing models. 

There’s definitely going to be money to be made from helping organizations control costs across what has become a highly distributed, unconstrained technology architecture. In my view, the harvest is not yet ripe — now’s the time to weave the baskets, not pick the still-sharp fruit. But there you go. 

This also plays into systems integrators, technology resellers, managed service providers and so on. The big item that we have lost is responsibility. Back in the day, end-user companies used to look for a limited number of tech vendors to work with, limiting to a handful of (“throat to choke” – not the nicest analogy) strategic partners. 

Today, they no longer have that. AWS may be a strategic partner, but then so will everyone else. If not at the corporate level, the other players will likely be embedded within departments or geographies. That, to me, is the new vendor opportunity, but it can only be achieved by actually supporting multi-cloud and on-prem models. 

And meanwhile, the opportunities are rich for those skilled at navigating this newly accepted reality. For some workloads, microservices models will be best; for others, it’ll be (shock, horror) monolithic stacks. Sometimes, a migration may be cost-effective, and other applications might be best left to run inefficiently but still effectively. 

Historically, If I had a soapbox, it’s been about bringing back old governance skills – risk management, quality management, configuration management, process management and so on. The good news is that you get these out of the box when you think architecturally, building to last rather than breaking things as a business strategy. 

So, I’m thinking I can put that soapbox away for the time being, as well as the hybrid one I first stood upon in 2007, not because I disliked cloud models, but because I could never see that any one ring would rule them all (sorry, Sauron). We’re entering a new phase of tech, requiring new skills and realigning with the old. 

This is true for all stakeholders across end-user organizations and the companies that serve them. We’re back to the old business model formula of what you want to keep in-house against what you want to get from your suppliers. Enterprises large and small, the floor is yours.