Multi Cloud is Dependent on Knowing Infrastructure
I had two thoughts while on the treadmill yesterday. One to get multi-cloud right, you need to know infrastructure. Two, Kubernetes is all about infrastructure.
I have a newly formed theory that I’m beginning to test. A successful multi-cloud deployment requires a deeper understanding of infrastructure. Let’s zoom out and look at single cloud vs. multi-cloud. I think think of multi-cloud as a scale-out solution vs. a single cloud as a scale-up solution.
When inside a single cloud, you can use that control plane to scale the number of applications and the overall system's capacity. Multi-cloud is triggered when a technical or functional requirement isn’t fulfilled with a single cloud.
We enter a multiprocessing computer science problem. We have to understand the control and data planes for processing I/O with any multiprocessing problem. I don’t see a way around it. The I/O challenge is especially the case if you select Kubernetes as the abstraction between public clouds. K8s is all about creating a contract between the application layer in the infrastructure.
Disclaimer: CTO Shorts are brain dumps of my thoughts. Feel free to criticize them with that thought in mind. If you want to be thoroughly irritated with my thoughts, visit https://thectoadvisor.com/blog