Kubernetes for VMware admins: a gentle introduction

Struggling to get your head around Kubernetes (K8s)? Explore the core concepts and use cases in our guide.



They say that in life, only two things are certain: death and taxes. To that short list, we can add that IT professionals have a new concept to grapple with every few years.

This endlessly unspooling learning curve is both a curse and a blessing. A blessing, because it makes the cloud computing world a dynamic one to work in. And a curse, because the sheer volume of new information to absorb can make your head spin.

At the time of writing, the hottest new concept on the cloud computing block is unavoidably AI. But for up-and-coming admins, there's another concept with an equally threatening visage: Kubernetes, often shortened to "K8s" or "K8".

Kubernetes is an open-source platform that automates the deployment, scaling and management of containerised applications.

But how, exactly, does it work? And what, precisely, is it used for?

In this article, we aim to break it down gently.

What is Kubernetes?

Kubernetes is a platform for managing containers. These are digital boxes which bundle together the code, configuration and dependencies of an application. Each container has its own resources and runs independently.

Every application is "containerised" – put into either one or multiple containers. These containers are then grouped together into "pods".

A key benefit of Kubernetes is its ability to work on multiple kinds of infrastructure: bare-metal servers, virtual machines (VMs), public cloud, private cloud and hybrid cloud.

Perhaps the best way to understand it (as outlined by Hany Michael in a blog post) is by comparing it to vSphere. "It is to containers," he says, "what vSphere was for VMs to make them datacenter ready."

How does it work?

Kubernetes works in stages. First, you define how you want your application to run – the so-called "desired state". These specifications, defined in YAML files, are submitted to the Kubernetes API.

Containers are packaged in pods and assigned to the best nodes in the cluster. Kubelet then launches the containers and gives the application as a whole a stable IP address and DNS name.

To achieve the admin or developer's desired state, Kubernetes now continuously monitors the application and makes corrections as it goes. If a pod crashes, the Kubernetes controller will automatically run a replacement.

And when it's time to scale, Kubernetes continues automating. It creates and removes pods as required to meet your scaling needs.

What are the three core design principles of Kubernetes?

There are three core design principles of Kubernetes: it should be secure, user-friendly and extendable. It shares these with other open-source solutions.

Of course, all cloud-based IT systems should be secure. But Kubernetes is intended to follow best practices, not the next best thing.

User-friendly, because Kubernetes can be run anywhere with just a few simple commands. And extendable, because Kubernetes is provider-agnostic and can be easily customised.

What are the benefits of Kubernetes?

There's a reason Kubernetes has legendary status among the tech-inclined. This is because of its wide range of benefits for businesses and tinkerers alike. Here are some of the most relevant to VMware admins.

1. Operations are automated so that applications are always running the way you intend them to run – the so-called "desired state".
2. Admins can focus on applications, not the underlying environment. Compute, networking and storage are all virtualised and handled elsewhere.
3. Kubernetes can orchestrate and manage large, complex environments with plenty of containers and plenty of hosts.
4. It auto-scales according to compute, networking and storage needs.
5. It runs continuous health checks. If a container fails or stalls, Kubernetes automatically restarts it.
6. Resource utilisation is optimised by packaging containers onto nodes according to their requirements. In turn, this helps lower infrastructure costs.
7. For developers (not so much VMware admins), development cycles are accelerated. More ideas can be tried, tested, removed and sent to market.
8. Application management is simplified. Kubernetes provides a uniform approach to the deployment, updating and managing of applications.
9. The software is open-source and portable. It can be run on the public cloud, on-premises data centres and on hybrid set-ups. The same simple commands can be used in a range of environments.

Kubernetes and Docker: what are the differences?

A common misconception is that developers and admins need to choose between Docker and Kubernetes. In fact, they're often used together to manage containers – and Kubernetes can be run both with and without Docker.

Essentially, Docker puts applications into a container and Kubernetes delivers that container to where it can be used. They work together to manage containers in a wide variety of cloud environments.

What are the use cases for Kubernetes?

Kubernetes has a wide range of use cases. Here are four of the most common:

1. Developing apps faster

Because Kubernetes allows microservices to be deployed at speed, developers rely on it to accelerate development cycles.

2. Avoiding vendor lock-in

Kubernetes can be run anywhere. That means it can be run across multiple cloud providers. This is a strategy favoured by enterprises that want to optimise costs, leverage high availability and avoid the dank spectre of vendor lock-in.

3. Auto-scaling on demand

One of Kubernetes' secret weapons is its ability to automatically adjust the size of a cluster. In practice, this means it can auto-scale video encoding (to take one example) according to the number of users watching.

4. Legacy app modernisation

With Kubernetes, traditional applications can be containerised and run on any environment.

There are plenty more use cases for Kubernetes. These examples are intended to give you a flavour of its breadth of application.

Where did Kubernetes come from?

Kubernetes began life as Google's internal cluster management system, known as "Borg". It was developed at Google Cloud and released as open source in 2014.

Are you looking for cloud optimisation services from VMware experts? Get in touch with Ascend Cloud Solutions today for a quick, competitive quote or explore our range of services online.