This article is more than 1 year old
The cloud is not new. What we are doing with it is
Sysadmin blog In the 10 years since the modern form of public cloud computing went mainstream, it has changed the entire industry's approach to IT. In response, IT's top vendors have had to change as well. Like any technology, however, the public cloud has adapted, evolved, and become something much different than was ever originally envisioned.
The public cloud is now a part of the fabric the IT universe; inseparable from reality. This doesn't prevent people from denying reality – humanity seems pretty good at that, all things considered – but the combination of near-instant provisioning, self service, scriptability and ease of use is the new normal for IT.
What's worth noting is that public cloud adoption wasn't driven by technology. In many ways public cloud offerings are shockingly inferior to those that can be delivered by on-premises teams.
Public cloud adoption was driven by human factors. In turn, customer demands transformed the public cloud from what was planned into what it actually became.
The public cloud of today
The public cloud of today represents the commoditisation of IT. Commoditisation is an often-overused word, but here I mean commoditisation in its truest sense.
As an analogy, let's take food. If you relied on game hunting for your food, you would have to know a fair amount. Where and how to hunt your prey; how to butcher the meat, including things like skinning, gland location and caution regarding various musk glands and so forth. Unless you only managed to catch exactly what you needed as you needed it, you would also need to know about preservation, storage, cooking, detecting spoilage and much, much more.
Our ancestors eventually split up these tasks. Different people did different jobs. The generalist systems administrator became the network administrator, the storage administrator and so forth.
What changed everything was not the division of labour among primitive hunter-gatherer tribes. It took millions of years for our ancestors to evolve into an agrarian society and the blink of an eye to go from agriculture to boots on the Moon.
What changed? Commerce. I am not talking here merely of agriculture. Agriculture alone does not an agrarian society make. Agriculture enabled higher population densities. Higher population densities meant there were now more people than were required for the business of simply surviving. So they got creative.
Early humans took hundreds of thousands of years to notably change the kind of spear tip they used. This makes sense; staying alive was a lot of work back then and when they did come up with something new disseminating it probably wasn't easy.
But as soon as we had people who didn't have to hunt or grow their own food, but could instead simply wander down to the market and buy whatever they needed, well, we were off to the races. Now the idly curious could invent better ploughs, aqueducts, rockets and clouds of computers.
This is what the public cloud is today. Individuals and organisations don't have to hunt and kill a server room in order to light up a Peoplesoft instance. They select a size of instance, feed it a credit card and go.
If you want 10,000 instances of computery something you can get that too. Feed your credit card into your account and then write a script. The public cloud is both retailer and wholesaler, but at it's core it's a marketplace: visitors are concerned not with how to get or make the things they purchase, but how they're going to transform the things they purchase.
They're off inventing a better plough.
What we were promised
What we were promised regarding the public cloud was very different, and it's apparent even in the language. Regular readers will note that I am very careful about using the term "public cloud". In 2016 this is something of an anachronism; everyone else simply says "the cloud".
We were promised private clouds. We were promised hybrid clouds. We were promised that turnkey operation and ease of use inspired by the large public cloud providers would find their way into our data centres.
Most importantly, the bulk of the hype about the public cloud was not around the consumption of compute by the milled masses, but about the ability of systems administrators to move workloads back and forth from our on-premises server rooms and data centres up to the cloud and back.
It's been 10 years, and we're still waiting.
All the things we were promised are possible. You can build a private or hybrid cloud. Sort of. It's difficult, it's miserable, the software is terrible, the experience is awful, the support is worse, and nobody wants to use the wretched things.
More importantly, the APIs that power the large public clouds are computing today. While systems administrators jealously guarded their empires and lacklustre vendors offered one feeble, laughably pathetic "cloud solution" after another, people with ideas and things to do turned to the public cloud and drove real innovation.
While IT teams and the vendors that orbit them demanded praise for making it easier to find a deer to hunt or for publishing a diagram with the location of musk glands, budget owners were turning to public cloud providers; buying cured meat, vegetables and potatoes in bulk; and opening restaurants.
We were promised better, cheaper IT infrastructure. What society wanted was a revolution in how IT was provisioned in the first place.
The analogy used above, however tortured, still holds up. If the world eventually had only three or four corporations to buy goods from things would be demonstrably very bad for the average citizen. Monopolies and oligopolies are bad enough when they are confined to one sector; the idea of a stranglehold on all goods is terrifying.
In a similar vein, it's not to anyone's benefit for IT teams to simply throw in the towel and concede defeat to the public cloud providers. The last thing society needs is all access to IT to be gated by three or four large corporations.
Convenience. This is what the public cloud offers. It's what drives revolutions of innovation. Vendors, IT teams, public cloud providers and consumers of technology haven't worked out an optimal arrangement about what workloads live on-premises, which live in the public cloud, and – above all – how to make the on-premises IT portion acceptably easy to use.
Collectively, we need to find our path here, lest we cede what may be the most important revolution since our development into an agrarian society to a mere handful of companies. We need to be able to achieve efficiency by running our own gear but retain the convenience required for innovation.
To get there, IT teams will have to change... and so will the vendors they buy from. ®