[jacobgorm]

Also, Xen's main clame to fame was that para-virtualization allowed it to host Linux and *BSD VMs at close to zero overhead, but at time what everyone was looking for was a way to host Windows VMs, which is where all the pain points and the money was. CPUs were evolving to support this use-case, making para-virtualization less important, and Xen had to evolve quite quickly to include QEMU in the mix, leaving a bit of convoluted mess initially, and causing a lot of friction during the attempts to get merged into the Linux kernel. On top, the Xen management tool-stack had been written by happy amateurs in Python and Twisted, before any of those technologies where near ready for production use, with massive slowness and unfixable memory leaks as results. KVM provided a fresh take built with the benefits of hindsight, got merged into Linux on the first attempt, and gained the backing of Redhat, and the rest is more of less history.