xen_run

Create a new set of page tables which will allow access to any physical memory address;
Create a set of signatures which can be used to hunt for _EPROCESS structures in kernel memory;
Find the _EPROCESS memory address for our executing process, and for the System process; and
Replace the token of our executing process with that of System, elevating us to NT AUTHORITY\System.

A fully supported #Xen solution on a great hardware platform for #embedded development. If you plan to develop an e…

What is XCP-ng?

XCP (Xen Cloud Platform) was the free/libre equivalent of XenServer before this last one was open sourced. It was logically shutdown then.

But since XenServer 7.3 removed a lot of free features, we need a community solution again: XCP-ng is now the logical successor.

Remus provides transparent, comprehensive high availability to ordinary virtual machines running on the Xen virtual machine monitor. It does this by maintaining a completely up-to-date copy of a running VM on a backup server, which automatically activates if the primary server fails.

Hardware virtualization for cloud computing has come a long way, improving performance using technologies such as VT-x, SR-IOV, VT-d, NVMe, and APICv. At Netflix, we've been using these technologies as they've been made available for instance types in the AWS EC2 cloud. The latest AWS hypervisor, Nitro, uses everything to provide a new hardware-assisted hypervisor that is easy to use and has near bare-metal performance. It's an exciting development in cloud computing: hardware virtualization is now fast.

I've summarized hypervisor developments in EC2 with the above diagram. The columns show dimensions of instance performance, ordered in importance for typical workloads at Netflix (CPU-bound being the most important). The rows of this table are virtualization types, and the cells show the type of virtualization and are colored by the expected performance.

You may first notice that, over time, green is creeping in from the left. That's deliberate engineering: optimizing the most important workloads first. Each dimension has progressed through these stages: