There are a number of relatively new phenomena in the server world that Microsoft has been rather slow to catch up on. Server virtualization is one of them. Where companies like VMWare and Sun (now Oracle) had pretty much already built the defining server virtualization software, with a robust hypervisor (software that lets you run multiple virtual machines on a single physical server) in ESXi, and a great general-purpose software-based virtual machine in VirtualBox, Microsoft made their own hypervisor.
And in traditional Microsoft style, their server virtualization implementation required modifying the Linux kernel to get it to play nice. Rather than emulating the system hardware in such a way that the Hypervisor does all the heavy lifting, they chose to use OS-level drivers to “get the most out of” the hypervisor’s features.
This isn’t generally a bad decision, honestly. VMWare requires guest OS tools to be installed in order to do some stuff too. Microsoft’s actual failing, in this case, was in employing juvenile dudebro programmers who submitted kernel code that included a constant for the upper limit for virtual server guest IDs defined as 0xB16B00B5.