I must admit that I wasn’t on the Kubernetes bandwagon from beginning. However, I’ve seen it mature rapidly over the last 4 years with an expanding ecosystem and high rate of adoption that has given me confidence that there is a future with Kubernetes and now is the time to invest in learning it. Here are the 3 things I observed over the last 12-18 months to solidify that now is the time for me to focus on becoming a Kubernetes expert:
Cloud Native Landscape Explosion
The Cloud Native Computing Foundation (CNCF) began publishing the CNCF Cloud Native Landscape back in 2016 to illustrate the projects and enterprise software that are part of the cloud native ecosystem and community. The Cloud Native Landscape graphic originally contained three projects: Kubernetes, Prometheus, and Opentracing. Today, it includes over 48 (not including CNCF member/non-member products and projects)! Observing this type of growth signals heavy investment by organizations and individuals to improving core functionality of Kubernetes or filling gaps in the ecosystem.
Kubernetes is Mainstream
Managed Kubernetes services seem to be the way many organizations are beginning to adopt Kubernetes. Services such as Google Kubernetes Engine (GKE), AWS Elastic Container Service for Kubernetes (EKS), and Azure Kubernetes Service (AKS) have accelerated Kubernetes deployments as it offloads the responsibility of infrastructure and cluster maintenance to the service provider while organizations can simply deploy pods for their applications. The whole point of the containerization effort is to streamline the deployment of applications. Developers don’t want to deal with the hardware and software resources to run their code. Frankly, they’re not paid for that either! Infrastructure teams, on the other hand, must support how developers are delivering code to the infrastructure so it’s important they have a streamlined process to support developers and new applications without the huge learning curve to design, build, and maintain Kubernetes clusters. These hosted options enable organizations to quickly adopt Kubernetes without having to re-train IT to support this new architecture.
Education and training resources have impressed me for such a young ecosystem. CNCF has been an exceptional steward of Kubernetes and the cloud native ecosystem by continuing to foster growth and evolution of the ecosystem by promoting the technologies with events, education, and certifications. KubeCon + CloudNativeCon attendance has continued to boom year-over-year since the first conference in 2015. That event had 550 attendees while the three conferences in 2019 saw over 23,000!
CNCF also created individual and partner certifications such as the Certified Kubernetes Administrator (CKA), Certified Kubernetes Application Developer (CKAD), and Kubernetes Certified Service Provider (KCSP) in 2017-2018 to help professionals get trained ahead of the adoption curve. The availability of official technical enablement is always a requirement if I’m going to invest in a technology or recommend it to someone else.
VMware’s all in with Kubernetes
Disclaimer: I work for VMware. I left this for last but it’s the most important and relevant to me. Don’t roll your eyes yet, hear me out! Three years before I joined, I felt that VMware was in a difficult spot — virtualization revolutionized the world 15 years ago but the infrastructure requirements to deploy applications have rapidly changed. As developers made investments in platforms that weren’t infrastructure-based, their awareness of or dependence on a virtualized platform was minuscule or non-existent. I waited to see how VMware was going to respond as re-innovating to take on this new challenge was necessary, or they risked becoming the new legacy.
VMware has progressively made small in-roads to join the cloud native space as early as 2017 when they joined the CNCF but 2019 got my attention when they turned this marginal venture into a big bet with the acquisitions of Pivotal, Heptio, and Avi.These acquisitions, along with VMware Cloud Foundation (VCF) and Tanzu Kubernetes Grid (TKG), deliver what I felt have been missing for mass adoption of Kubernetes to occur: the creation of a foundational, standardized platform that enables organizations to run new, container-based applications alongside VMs on private or public clouds. By integrating Kubernetes into vSphere, VMware enables organizations to have a unified infrastructure that allows developers to continue creating new applications in containers which can be run on software infrastructure which is already supporting legacy VMs and is familiar to operations admins. This integration provides businesses the technical and financial agility they require to innovate on any cloud while continuing to support older, critical systems. In my opinion, as most businesses’ applications are already on premises and the rising trend of cloud repatriation due to cost, most Kubernetes clusters will be deployed on premises and VCF/TKG provides a simple, efficient platform for the future.
Why Didn’t I Pivot to Kubernetes Earlier?
As I previously mentioned, I’ve been skeptical about Kubernetes’ success since I started watching it in 2015. A few of the turn offs were:
I’m a simple person — where’s the easy button to install this?! If you’ve experienced Kubernetes the Hard Way, you know firsthand how intricate and tedious setting up Kubernetes can be. The process entails multiple installations of components and subsequent configuration and management of individual components. There’s value in disaggregated, distributed systems but I saw a highly complex system with a huge learning curve that was maturing and promising but still wasn’t ready for the enterprise. Granted, it’s called Kubernetes the Hard Way for a reason and not necessarily intended to be the path to production, but it does highlight how challenging Kubernetes can be.
Is this another OpenStack? From the outside, I questioned if Kubernetes was setting a path to become the next OpenStack. OpenStack received a lot of attention by enterprise infrastructure administrators and architects anticipating free, full stack enterprise datacenter software as an alternative to VMware. Following Kubernetes the Hard Way, I observed many of the same complexities in Kubernetes that riddled OpenStack and challenged many organizations to defer deployment or silo for niche projects. Most importantly, the complexity overshadows the value it can deliver to the business. The problem with OpenStack was the surrounded hype of free, open source, enterprise datacenter software that was highly customizable and widely anticipated to replace VMware as the datacenter platform. That same hype turned OpenStack into a dystopia due to lack of individuals skilled in operating the stack which prevented organizations from having successful implementations.
It was a Google project — that can be a blessing or a curse. I’ve been a Gmail user since it was in beta and loved the different perspective on email since the beginning, but Google has a bad habit of releasing minimally viable products (MVPs) then walking away from them or killing them. Google’s name carries a lot of prowess but personally I’ve been gun shy to immediately jump on a new Google product or tool because of this track record.
But I missed what Google was trying to accomplish by open sourcing and releasing Kubernetes — a gateway to sell GCP. By productizing Google Kubernetes Engine (GKE), Google has brought more attention to GCP as a destination for Kubernetes workloads as opposed to AWS or Azure. They wanted and needed it to be successful so it was never going to be a whimsical product release that could easily die on the vine. As it turns out, according to the 2019 CNCF Survey, AWS is still the preferred platform to run Kubernetes.
Are you learning Kubernetes in 2020? Find me on Twitter, and share your thoughts! I hope you’ll continue to follow along my journey from Kubernetes novice to expert with Kubernetes 0 to 7!