The Sysbox runtime enables the use of containers for several use cases that extend beyond microservices.
Kubernetes-in-Docker (aka KinD) means using Docker containers as Kubernetes nodes (instead of physical hosts or VMs). A K8s cluster is a collection of Docker containers.
This is very useful for learning, testing, and CI/CD.
The Sysbox runtime is the first to enable KinD using simple Docker run commands, simple container images, and strong isolation (no privileged containers).
This gives you full control of the cluster configuration and ensures your host is properly isolated from the activity within the cluster.
You can deploy the cluster using simple Docker run commands, or more commonly with a tool such as the K8s.io "kind" tool or Nestybox's kindbox tool.
And its fast and very efficient: you can deploy a 10-node cluster in less than 2 minutes with minimal resource usage.
It's often useful for software developers to have a dedicated Docker sandbox environment, inside of which they can play around with containers in isolation.
You can provision these with VMs, but they can be pretty costly.
What if instead of spawning several cloud VMs (one for each sandbox), you could spawn one VM and "split it" into multiple Docker sandbox environments using containers?
The Sysbox container runtime enables you to do this, using simple Docker run commands, simple container images, and strong container isolation.
This not only reduces cost, but also increases agility as the container image acts as a preconfigured virtual host, one that can be deployed on your development machine or in any cloud, in seconds.
Sysbox can be used to expand the capabilities of CI/CD pipelines.
Many CI/CD frameworks use Docker containers as the unit of job execution.
By using Sysbox, these job containers can now run Docker containers (Docker-in-Docker) or Kubernetes (Kubernetes-in-Docker).
This enables you to run CI/CD jobs that build Docker containers and deploy them within ephemeral Kubernetes clusters with high speed and efficiency, and without resorting to costly Kubernetes clusters on the cloud, slower VMs, or unsecure privileged containers.