Sunday, October 9, 2011

Revisiting virtualization

I dusted off some virtual machines today. They are experiments, to see how virtualization performs in different environments.

I use VMware and Oracle's Virtualbox. Both had updates; new versions of the virtual machine software.

The updates for Virtualbox ran a bit smoother than VMware. Both programs alerted me to the new version, and let me download them. Both updates were replacements and not patches. That is, both un-installed the old version and then installed the new version. And both new versions had new "inside helper" packages, utilities that install drivers inside the guest operating system to help the "outside" virtualization program.

It was the installation of helper utilities that Oracle beat VMware. With VMware, you have to run a script, answer several prompts, and let the script compile several modules for your system. (The questions have default answers, but they are sensible only to a sysadmin.) Virtualbox, on the other hand, installed the helper utilities without prompts or compiling.

For both VMware and Virtualbox, the improvements were evolutionary, not revolutionary: small improvements to the GUI, noticeable changes in support for devices.

And for VMware and Virtualbox, performance was acceptable. Virtual machines run a little slower than a "real" machine, but the difference is small.

I don't use Microsoft's Virtual PC, as it refuses to run on my hardware. My PC is a two-year-old SystemMax PC with a nice Intel processor and 4 GB RAM, but apparently it lacks something that Virtual PC needs. So I limit my tests to VMware and Virtualbox.

No comments:

Post a Comment