QEMU revs its open source virtualization platform

Survey 100 people across the virtualization community and ask them to name a virtualization technology, and it’s a safe bet the vast majority will name a VMware technology like Workstation, Fusion, or vSphere or perhaps even Microsoft Hyper-V or some flavor of Xen. Ask this same question of people in more nonmainstream areas, and you’ll probably get answers like KVM, Parallels, or VirtualBox instead.

But it’s a safe bet none will mention an unsung hero in virtualization, namely QEMU (Quick EMUlator). This open source project, started by Fabrice Bellard in 2003, is still around and going strong (believe it or not), and it deserves some recognition. Less than two weeks ago, right around the Thanksgiving holidays, QEMU developers announced a new stable release of their code, version 1.7.0. It seems an appropriate time to pay homage to this open source project.

[ Also on InfoWorld: New Fling installs VMware Tools on nested ESXi virtual machines | Startup announces SaaS solution to manage Amazon Web Services resources | Track the latest trends in virtualization in InfoWorld’s Virtualization Report newsletter. ]

While today’s virtualization community is a loud and proud army of IT administrators, things weren’t always that way. In the early part of 2000, the community was small, the virtualization geeks were considered tinkerers, and virtualization giant VMware wasn’t, well, so giant. Fourteen years ago, x86 virtualization was considered a toy technology by mainstreamers. Who’s laughing now that virtualization has become a commodity and is used in data centers all over the world?

Vrtualization elders might recall the names of some of the fallen like BOCHS, Virtual Iron, Win4Lin, and SVISTA, to name just a few. Many others came and went as well, but QEMU still stands tall.

Why is the QEMU project still going strong 10 years later? Perhaps because it has tentacles within many of today’s hypervisors.

Modern virtualization technologies like Xen and KVM have leveraged the work of several QEMU projects. By leveraging QEMU, Xen and KVM have been able to implement numerous performance optimizations, giving those platforms a jump start.

The QEMU project was started as an open source emulator that provided the ability to run on a wide variety of platforms by doing dynamic translation of native CPU instructions. This dynamic translation enabled QEMU to provide fast and efficient emulation by breaking down the processor instructions into small blocks of atomic instructions. Those instructions were then processed, cached in a buffer, and then the responses were returned as if the native processor had acted upon them.

According to the QEMU website:

QEMU is a generic and open source machine emulator and virtualizer.

When used as a machine emulator, QEMU can run OSes and programs made for one machine (e.g. an ARM board) on a different machine (e.g. your own PC). By using dynamic translation, it achieves very good performance.

When used as a virtualizer, QEMU achieves near native performances by executing the guest code directly on the host CPU. QEMU supports virtualization when executing under the Xen hypervisor or using the KVM kernel module in Linux. When using KVM, QEMU can virtualize x86, server and embedded PowerPC, and S390 guests.

Fast-forward to QEMU’s latest 1.7.0 release, which adds a number of new features to the project. These new features may not seem as sexy as some it has made in the past, such as live migration for guests, support for SSD and thin provisioning, adding support for Large Physical Address Extensions (LPAE) ARM technology on platforms with Cortex CPUs, or GPU pass-through from host to guest. But these updates are no less important — even if they only appeal to the geekier virtualization users like me.

Some of the new highlights worth mentioning in this latest stable release include:

• Several enhancements for audio support within guest virtual machines
• Additional device assignment control using Virtual Function I/O (VFIO), a user space driver framework designed to pass through PCI and PCIe devices to guests, allowing them to access these devices at low latency and high data throughput levels without any risk to the host
• Performance and functionality improvements for USB 3.0
• Virtual disk image updates, including driver support for VMware’s VMDK files produced by ESX; the qcow2 driver can now detect some cases of corruption and prevent writing to a corrupted image; and support has been added for creating and writing VHDX images from Microsoft Hyper-V
• The “filesystem freeze/thaw” commands are now supported on Windows where they will invoke the native Volume Shadow Copy Service (VSS)

In addition, you’ll find important x86 updates:

• Guests can access the ACPI tables of the host servers directly, which in the future will enable new features without modifications of all firmware components (SeaBIOS, OVMF, CoreBoot).
• Initial support for more than 1TB of RAM.
• Xen HVM domains can now resume from suspend-to-RAM (S3) state.

An overview of these and other new features can be found in the change log. QEMU 1.7.0 is available for download from the project’s website where documentation is also provided.

Again, I’m thankful for virtualization. It has completely changed my work life for the better in so many different ways.

Thankfully, QEMU has been innovating for the past 10 years, helping to improve virtualization as well as cloud computing.

This article, “QEMU revs its open source virtualization platform,” was originally published at InfoWorld.com. Follow the latest developments in virtualization at InfoWorld.com.