Presentation: Cloud Native Infrastructure With Kubicorn
What You’ll Learn
- Find out about Kubicorn and what infrastructure problems it can solve.
- Learn about the infrastructure reconciler pattern and what are some of the perils of managing infrastructure.
- Hear about writing infrastructure software that can solve real problems.
Abstract
Kubicorn is an open source Go project that is aimed an solving the Kubernetes infrastructure problem. The day after the project was open-sourced it had already climbed to the number 1 fastest growing Go project on GitHub. This talk will dicuss the reasoning behind the project and how the core of the tool is developed strictly for cloud native application. We will clearly identify the problem space with existing infrastructure tooling and discuss how Kubicorn satisfies new patterns that can easily be vendored into control loops and operators.
Managing infrastructure via a cloud native application is a ground breaking idea and the project is a shining example of how we can begin to reason about infrastructure in this exciting new cloud native world we are living in. We explore the bootstrap problem of needing infrastructure in place to run the infrastructure management application and look at the importance of atomic infrastructure changes.
We will learn how the tool is a lovely implementation of the infrastructure reconciler pattern defined in my book "Cloud Native Infrastructure" and learn the dangers of managing infrastructure in other ways. The lessons discussed are battle tested and have been proven over time. The audience will gain a rich understanding of what it means to run a cloud native application that manages underlying cloud native infrastructure through concrete examples from the Kubicorn project.
What is the focus of the work that you do today?
I'm working today on making the infrastructure layer of the stack better, faster and stronger, particularly in regards to Kubernetes, and how we handle all this noise known as infrastructure bits.
What's the motivation for this talk?
This is almost a life goal of mine. I wrote a book on it. I have contributed to Kubernetes for quite some time, and I open sourced this new tool called Kubicorn. This is all work that I've done trying to use software to manage infrastructure. As a DevOps engineer I started my career as a systems administrator, and I've done clicking buttons, writing scripts, infrastructure as code, and I decided to write software that does all of that for me because I'm done dealing with it any other way.
What's the meat of this talk, infrastructure automation, or a tool to deal with the infrastructure noise?
Some of my best friends, founders of Kubernetes, said "We won't deal with infrastructure; that is something you get on your own." Ending up with a fragmented community, I decided to write this tool called Kubicorn. That would be the core of the talk. One of the beautiful things about this tool is that it gives you a starting place but it doesn't necessarily give you all of the answers to all of the questions. It has some opinions, it allows you to bring your own implementation along the way. There are eight or nine places where you can plug in out of the box, but if you want to pull out one of those components and replace it with your own, please, by all means, do it.
Is there an alternative to Kubicorn in this space?
There is and isn't. There are tools that attempt to solve the same problem as Kubicorn but in a very different way. Kubicorn has this idea of residually running software over time, that you don't really see in the infrastructure land. You might see it in application land. That's one of the reasons Kubernetes is so successful because you have robotics control that reconciles applications state over time, but for infrastructure it's very much a one time throw the ball over the fence, that kind of mentality. Kubicorn has this while loop that runs over and over again to reconcile infrastructure.
Are there takeaways for someone who might not be able to use Kubicorn?
The pattern is what really gets me excited. It is this idea of having a loop that runs over and over again. I think it's a lot easier for people to learn about this pattern with a concrete implementation like Kubicorn. And my last slide is going to be all caps: START WRITING INFRASTRUCTURE APPS. Go, do it on your own. I want you to go do this. I want you to learn these lessons, but here's how I did it.
Who is the persona that you're talking to?
In my mind there's this fictional grumpy DevOps guy complaining he's been doing this work for 20 years. I want to meet this person and this is who I'm talking to. And I have a little bit of interest in becoming a software engineer because a lot of what I'm prescribing here is this idea that after writing Bash scripts we can take it to the next level and start writing enterprise level applications to solve these problems that we've been working in the past.
Similar Talks
Linux Foundation's Project EVE: A Cloud-Native Edge Computing Platform
Co-founder, VP Product and Strategy @ZededaEdge & Member Board Of Directors for LF Edge @linuxfoundation
Roman Shaposhnik
License Compliance for Your Container Supply Chain
Open Source Engineer @VMware
Nisha Kumar
Observability in the SSC: Seeing Into Your Build System
Engineer @honeycombio
Ben Hartshorne
Evolution of Edge @Netflix
Engineering Leader @Netflix
Vasily Vlasov
Mistakes and Discoveries While Cultivating Ownership
Engineering Manager @Netflix in Cloud Infrastructure
Aaron Blohowiak
Optimizing Yourself: Neurodiversity in Tech
Consultant @Microsoft
Elizabeth Schneider
Monitoring and Tracing @Netflix Streaming Data Infrastructure
Architect & Engineer in Real Time Data Infrastructure Team @Netflix
Allen Wang
Future of Data Engineering
Distinguished Engineer @WePay
Chris Riccomini
Coding without Complexity
CEO/Cofounder @darklang