IBM's power architecture gets KVM virtualization

Even before the end of the current quarter, the KVM virtualization for the power architecture is ready. IBM has been working on the Porting PowerKVM since 2011. The Linux-integrated hypervisor is so far only compatible with Intel and AMD processors.

IBM has completed the development of a version of the Linux-integrated hypervisor Kernel-based Virtual Machine (KVM) for its own power architecture. A power system version of KVM called PowerKVM will be released later this quarter. This is written in a blog by Jim Wasko, director of IBM's Linux Technology Center . PowerKVM will be released for the next generation of IBM's linux-based power system servers.

Since kernel version 2.6.20 of 2007 KVM is part of Linux. This facilitates virtualization by using virtual machines (VMs) on Intel or AMD processors with the virtualization extensions Intel VT and AMD-V, respectively. KVM does not work on other chips.

IBM has been working since 2011 on a port for its own power architecture. Red Hat and SUSE plan to integrate PowerKVM into the enterprise Linux distributions Red Hat Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SLES).

According to Wasko, IBM does not only want to enlarge its own distribution in the open-source sector with porting. "First, Linux users want a 'familiar' look and feel for virtualization, and second, cloud solutions demand KVM's flexibility, performance and OpenStack integration," said the IBM manager. "We also find that for those who prefer a Linux-only environment, working with KVM is highly desirable."

The mention of OpenStack is also not to be underestimated. PowerKVM should help to set up OpenStack clouds based on power servers.

Wasko explains, "Like Linux, KVM for Power leverages the underlying hardware including multi-threading, large-memory support, and I / O. It also supports Kimchi - a graphical open-source tool for easy virtualization management in simple configurations. More complex configurations, such as clouds, can be managed with OpenStack-based tools. "