Illumina Innovates with Rancher and Kubernetes
Since joining Rancher Labs to head up the Australia, New Zealand, and Singapore region, my day revolves around discussing containers/Kubernetes use cases and adoption with many of the top enterprises, DevOps groups, and executives in the area. Not only is this a great learning experience and a fantastic way to meet people, it is also a huge eye opener into the many reasons why Kubernetes adoption is growing so rapidly and what the current challenges are. I want to quickly share some of my observations and make an offer for you to join us for some free hands-on training.
It doesn’t matter which event, meetup, or customer discussion I’m in — every enterprise is doing something with Kubernetes. It’s like the adoption of virtualization, only the discussion is slightly different. It’s not so much about which vendor or standard — Kubernetes is the focus. Instead, it’s about how to do Kubernetes and what are the associated best practices, scalable architectures, and security considerations.
The community and ecosystem around Kubernetes is growing every day, with strong capabilities, so there is a strong desire to stay on “native” Kubernetes and not get sucked down a branch, fork, or vendor-specific offshoot of Kubernetes. It seems that most enterprise and groups begin this way and get into production with Kubernetes. However, there is a clear point at which scale becomes an operational challenge and basic tools need to be supplemented or worked on to help manage multiple Kubernetes namespaces, multiple clusters, authentication, RBAC, policy, monitoring, and logging across many development teams.
Nobody wants to be in the business of creating Kubernetes snowflakes, or be in the business of allocating their resources to do work that adds no value. There is a learning curve for operationalizing Kubernetes, using Kubernetes, and deploying workloads into Kubernetes environments. Many enterprises are looking for ways to eliminate the learning curve or the need for specialized skills and instead just consume Kubernetes, using a Kubernetes-as-a-Service model. Much larger and faster gains can be made if consuming Kubernetes becomes the focus instead of making Kubernetes.
As enterprises grow, iterate, and merge, an ever-increasing mixture of infrastructure environments and needs emerges. The same enterprise may create Kubernetes clusters using on-premise bare metal, with OpenStack and VMware-type infrastructures, as well as out on public clouds using Amazon, Google, Azure, Alibaba, and others. The portability and rapid pace of containers lends itself to these hybrid or multi-cloud scenarios (more so than VMs) and is quite quickly sprawling in this way. There is also quite an urgent need for air-gapped Kubernetes environments.
Most enterprises are now seriously looking at the Kubernetes services offered by public cloud providers, like EKS (now available in Australia & Singapore), GKE, and AKS. These are viable options and really do support some of the notions mentioned in my other observations, like consumability. Technical discussions here become much less about the Kubernetes cluster control planes and architecture, and more about integration of these clusters into enterprise management capabilities like authentication domains, security models, deployment pipelines, and multi-cloud strategies (e.g. on-premise or multiple public clouds).
We run free, half-day training sessions called Rancher Rodeos throughout the world. Among others, this month we have Rodeos in Sydney, Melbourne, and Singapore (registration for Singapore is not open yet). During these sessions, DevOps and IT professionals can get hands-on experience with how to quickly deploy an enterprise-ready Kubernetes environment on any infrastructure or cloud provider (or multiples of these) using Rancher. We will show how Rancher helps make enterprise Kubernetes consumable and native, with rapid results for development and infrastructure teams.
Please take us up on the offer, register here, and join us!