LOD will be adding a new fabric, to run labs in containers. Currently data centers have lab hosts to run virtual machines, but we will be adding lab hosts to run containers.
- Docker will be configurable on lab profiles, in the same drop-down where you would select Hyper-V or vSphere.
- The user experience will be similar to Linux SSH labs, where there is no desktop UI for the student to interact with.
- Running labs in containers results in labs that can be launched much quicker than a traditional VM-based lab.
- Running labs in a Docker container will be particularly exciting for Linux labs. Linux environments will be ready for student use in a matter of seconds, instead of minutes. This is because each Linux sub-system does not need to be loaded per lab instance, as most of the files needed for the lab are already running on the Docker lab host.
- Containers enable Linux labs to perform similar to a Windows-based VM with improved mouse support.
Docker v1 fabric will include support for:
- Load balancing across hosts
- Lab save/resume support
- Multiple containers
- Split window support
- Activity based assessments (ABA)
- Life Cycle Actions (LCA)
- ABA and LCA in containers use Bash for scripting and scoring. Refer to the Activities documentation for more information about Bash scripts.
Docker v2 fabric will include support for:
- Ability to save a running container as a new image, similar to diff disk capture on VMs. (uses Docker Commit)
- Ability to provision images on the fly on Docker lab hosts.
- From Docker image repository
- From LODS image repository
- Extend lab host availability scanning to include Docker images (which hosts have all images for individual labs)
- Extend ASL/ASR to load/unload images on Docker hosts.
If possible, Docker fabric v2 will include:
- Container networking between containers
- Container to VM networking