Ultimately, I wanted to access a containerized Windows machine using RDP, and enjoy the full remote desktop access to that machine.
Even though a container with a VM in it takes more disk space than other containers, the efficiency in saving disk space when we have many containerized VMs is still high compared to running un-containerized VMs. However, some of those containers run a hypervisor, and on top of which there is a Windows VM. From the perspective of Docker daemon, all containers are Linux-based. After some research, the solution that presented itself as the most viable was to install each Windows machine as a virtual machine inside a single Linux container. Natively, one cannot run Linux and Windows containers simultaneously on the same Docker daemon. However, that presented other challenges.Įach of Windows and Linux machines have their own containerization requirements - as will be discussed in the next section.
It became evident by that point that a technology like Docker containerization is the best way to go. While the web-based access to machines can easily be implemented using a platform like Guacamole, the greater challenge was how to deploy the machines in a way that utilizes system resources - particularly, CPU, RAM, and HDD - efficiently and with speed. I wanted to make the access to all machines uniform over HTTP/HTTPS (browser-based). Recently, I was tasked with developing a PoC of a lab environment where students can access their pre-installed and pre-configured machines - Linux and Windows - to do necessary training exercises.
0 kommentar(er)
