VIRTUAL MACHINES MADE MANIFEST

THESE DAYS it’s one of the less-crazy conspiracy theories that the world isn’t real and that all of this—peers around the decaying Maximum PC office—is an elaborate virtualized simulation run by some bored alien teen on their Apple IIq quantum computer. Brushing aside simpler questions such as, “Why?” and, “Could they bump up my pay check please?” it shouldn’t mask that the modern world is in fact now largely run in a virtual world.

This text is being run as part of a Java Virtual Machine inside a browser. It’s stored in the “cloud” as a managed virtual storage pool that’ll be maintained as part of a Kubernetes orchestrated virtual containerized system. The webpages we’ve read to research this will have been served off high-availability load-balanced virtual server clusters, and the music-streaming service entertaining us while we write will be from a similar server farm. The list goes on.

This is very much the commercial end of our virtual world, where complexity is abstracted away behind automated software layers for easier management. As virtual technology advances into the GPU realm it’s now orchestrating game-streaming services and enterprise-accelerated 3D desktops. But even on your personal machines this same abstraction can help you manage your own suite of virtual machines for running and testing software from Linux servers to old Windows builds, DOS games and more. So let’s look at the tools being used, how you can run virtualized machines, and how the virtual technology extends to GPUs.

THE KEY TO VIRTUALIZATION success is ultimately financial, and by financial we mean laziness. Banks of virtual servers can be spun up on demand and dismissed when no longer required. This makes system-administration jobs easier but means that services can scale easily—up and down—depending on demand. This minimizes overheads and can ultimately be made an automatic process.

While this technology might be first aimed at 22-core Xeon