Google Kubernetes; An Introduction

One of the advantages of Kubernetes is that users can precisely tune the allocation of cluster resources to each container. This especially useful when applications are designed as a set of microservice. For example, a microservice might need only a small fraction of a CPU. In that case, a container running that microservice can be allocated only amount of CPU needed. This allows for more efficient use of compute resources.

Alternatively, if the microservice were deployed on a VM with a full vCPU, the vCPU would be underutilized. This leaves developers with the choice of either tolerating inefficient utilization or running additional microservices on the same virtual machine. This approach has its own disadvantages. Unless all microservice have the same lifecycle, you may find that performing maintenance on one microservice can disrupt the operations of the other microservices.

By separating application components into microservices

By separating application components into microservices and running those microservices in their own containers, you can improve resource utilization and possibly make the application easier to maintain and update.

Another advantage of Kubernetes is that it can run in multiple environments, including other cloud providers and in on-premises data centers. Kubernetes can detect when containers within pods are not functioning correctly and replace them. Kubernetes can also scale up and down the number of pods to accommodate changing workloads.

Perhaps the most significant constraint of Kubernetes is that applications must be containerized to run on Kubernetes. If existing applications running on premises in VMs are migrating to the cloud, a lift-and-shift migration, in which the system architecture is not changed, to Compute Engine is likely the fastest way to get to the cloud. Once there, you can containerize your application.