Despite Amazon’s embrace of Kubernetes, the open source container orchestration engine keeps dreams of a multicloud world alive.
No matter how fervently a CIO may intone that she is “all in” on this or that cloud, or any particular vendor, the messy reality of enterprise IT is that every enterprise is hybrid by happenstance, not choice. Yes, it’s easy to find overly optimistic opinions on “the benefits of a multi-cloud strategy” (authored by vendors that crave your cash), but “hybrid” and “multi-cloud” tend not to be strategies at all.
They’re simply what happens in a world driven by developers.
For organizations that find themselves hopelessly hybrid (which is to say, every company on earth), take heart. There is a way to make this mess look like (and serve as) a very real strategy. It’s called Kubernetes, derived from the Greek word for “helmsman,” and it offers a way to turn hybrid computing into a smart business strategy, even if that’s not how it started.
AWS vs. Kubernetes
For nearly every enterprise, “hybrid” actually started with AWS. That is, companies had “standardized” on Windows or whatever for their data centers and woke up to find developers were adopting AWS en masse. Even today, as much as a CIO may pledge allegiance to Microsoft Azure, someone, somewhere within her organization is running AWS (and probably lots and lots of AWS).
Indeed, seemingly overnight (but in reality it has taken over a decade), AWS is everywhere. The reason is simple: AWS makes life simple for developers, much as Windows did before it. The two have a lot in common, as one developer has noted:
LIKE WINDOWS, AWS IS A PRODUCT. IT’S NOT FLEXIBLE, ITS BEHAVIOUR IS RELIABLE. THE APIS ARE WELL DEFINED, THE KPIS ARE GOOD ENOUGH TO BE USEFUL FOR MOST ‘REAL’ WORKLOADS. THERE ARE LIMITS ON ALL SORTS OF RESOURCES THAT HELP DEFINE WHAT YOU CAN AND CAN’T ACHIEVE. MOST PEOPLE WANT THIS, LIKE MOST PEOPLE WANT A CAR THAT RUNS AND DOESN’T NEED TO BE FIXED OFTEN.
AWS has consistently won by enabling developer choice, largely by making those choices easier to make and to digest. Which brings us to Kubernetes or, as Redmonk analyst James Governor labeled it in 2016, “the Anyone but Amazon club.”
Navigating (away?) from AWS
AWS took a long (long) time to embrace Kubernetes, despite how popular it had become with developers and the enterprises that employ them. From the AWS angle, they already had ECS, which made it easy to manage containers at scale…provided you stayed within the walled garden of AWS. More pertinently, however, Kubernetes was an existential threat to AWS: It makes it relatively straightforward to move workloads between compute environments.
Here (on AWS) today, gone tomorrow?
Reality is a bit messier, of course. The majority of Kubernetes workloads run on AWS, and despite its initial foot-dragging on Kubernetes, AWS now has its own Kubernetes service (EKS). While EKS still suggests Amazon’s heart may not be in it (Google’s Kelsey Hightower, for example, called out how slow it is, concluding “this is not what I expect from a managed Kubernetes offering”), it may be enough to convince Kubernetes fans to stick with AWS just long enough to eventually move on to serverless offerings like Fargate. As the zwischenzugs author has stated, “EKS (like all AWS services) is heavily integrated with AWS IAM. As most people know, IAM is the true source of AWS lock-in (and Lambda is the lock-in technology par excellence. You can’t move a server if there are none you can see).”
And yet there remain enterprises (read: all of them) that are stuck needing to manage workloads across AWS, yes, but also other clouds and private data centers, with a mix of virtualization technologies, OpenStack, bare metal, and every other permutation of computing. For these, Kubernetes gives them a way to package and manage applications in a consistent way across inconsistent environments.
As such, the two companies to watch are Red Hat and Google, the two primary contributors to Kubernetes (Google accounts for 39% of all contributions, while Red Hat is 16%. VMware is third with 3%). Of the two, Red Hat is the safer, smarter bet with its Kubernetes-based OpenShift, both because it has a long history of making open source work for the enterprise, and because it doesn’t have a public cloud “horse” in the race. It also helps that so many enterprises already depend upon Red Hat Enterprise Linux, as Red Hat’s Daniel Riek has posited, “OpenShift, as an extension and superset of traditional RHEL, is the future cloud-native operating system.”
Does this mean Red Hat is AWS’ biggest competitor? No. But it does suggest that Kubernetes will likely prove AWS’ biggest threat, and those vendors that make Kubernetes easiest to consume will be best positioned to capitalize on a truly hybrid (multicloud, public and private) world.