Friday, November 13, 2009

Results of the Ubuntu Virtualization Survey

A big thanks to everyone that participated in the Ubuntu Virtualization Survey. I am pleased to share the results with you now.
I will provide a few of my own observations, but we are very interested in your own conclusions!
  • There were a total of 354 responses -- excellent feedback!
  • Nearly 2/3 of all responders use virtualization on Ubuntu every day -- wow!
  • Over 3/4 of responders have VT acceleration -- that's overwhelming, I think, and it supports our focus on KVM.
  • Still, there's 21% of responders who cannot use KVM. kqemu has been deprecated by upstream QEMU, so I think VirtualBox represents the best option at this point for non-accelerated virtualization.
  • 36.7% of responders most use VirtualBox, 22.6% most use KVM. VirtualBox is in Universe and essentially unmaintained by Canonical (though some community individuals are doing an excellent job maintaining it!). I don't know what the business opportunity is around VirtualBox. But it is clear that it's popular among Ubuntu users. People really like the interface and the usability. And we could probably really improve the experience for a large number of Ubuntu virtualization users with some dedicated Canonical effort to clean up the VirtualBox bug backlog.
  • My survey design was evidently flawed on Question #3, as a large number of people "wrote in" an "EC2" answer there. This is an interesting approach, as it diminishes the importances of having VT on the local system.
  • In terms of interfaces, virt-manager and virsh are both lagging behind kvm-from-the-command-line and VirtualBox. I don't know if this means that we should, or should not invest more in the libvirt-based tools. Is the lack of a good GUI for KVM hindering its adoption? I think this data says so...
  • Finally, the overwhelming majority suggests that better documentation is simply required for Ubuntu virtualization. I wonder how we should approach solving this? Is this something that we as engineers should be able to just crank out ourselves? Or should we tap into the Ubuntu-Documentation-Team, and attempt to rally a virt-documentation blitz from some more skilled tech writers?
In summary, I think the most important observations are that:
  • The overwhelming number of respondents have access to VT hardware.
  • VirtualBox is quite popular in the wild, despite a lack of Canonical investment.
  • The lack of a better user-interface is hindering KVM's adoption.
  • Better documentation is undoubtedly requested.
Are there other observations you'd like to share or conclusions you can draw?

We will be in Dallas next week for the Lucid Ubuntu Developer Summit, discussing the future of Virtualization on Ubuntu. Thank you so much for your feedback!

Cheers,
:-Dustin

6 comments:

  1. Hello Dustin, I missed this survey, but in my small tiny case, the lack of good libvirt based frontends is definitely hindering wider adoption.

    Virt-manager is (in my case) slow enough to be unusable (does it work well for you?). I use libvirt mostly via virsh, but a nice GUI would allow people with less virsh experience to manage the virtual machines as well.

    Some efforts in this direction would certainly be very useful, at least to me!

    ReplyDelete
  2. I tried the virt-manager interface on a 9.10 desktop last week. Did not work well. I found it easier to use virsh from the command line.

    ReplyDelete
  3. Hi,

    Thanks for taking the time to do the survey and report back the results. I regret though that you did not mention at all Linux Containers and the lxc tools. They are very potentially very good for non-accelerated virtualization and as replacements for OpenVZ and Linux-vservers.

    All of my Linux servers are Ubuntu based and I could do without the KVM overhead. I only need KVM for Windows virtualization.

    LXC really should be added to the virtualization options of Ubuntu. I hope you do.

    Best regards
    Philippe

    ReplyDelete
  4. [OFF TOPIC]

    I really want this idea to be heard by Canonical, since it’s a business plan which, if executed, will make my life a whole lot easier. So here (http://brainstorm.ubuntu.com/idea/21921/ ):

    A. Switch concepts from “Software Repositories” to “Online Product Inventories” (OPI). OPI can contain products of all types, including software, music, or grocery items.

    B. Create an online payment service called “Ubuntu Shopper Partner” (USP), which the online shopper can use to purchase products from OPI. This service will integrate with the package management software on the user’s PC so that stuff like product keys will be handled in a manner transparent to the user. USP will make money for Canonical through fees.

    C. There will be a program offering to include OPI in default Ubuntu installations in return for monetary compensation. All products in such OPI must be fully compatible with the Ubuntu release in which they are included.

    D. This obviously doesn’t come to exclude Free Software, only to add commercial options as well.

    ReplyDelete
  5. You say nothing about Xen. What's with that? Will we see xen packages in the future (perhaps when Dom0-code finally arrives in vanilla kernel)?

    ReplyDelete
  6. I'd really like to see a decent GUI for KVM too. I tried virt-manager and liked the VirtualCentre-style view of my VMs that it gave me, but I think it's lacking in individual VM control such as network options, USB, or snapshots. All done very well in VirutalBox. It's what I'll continue to use, probably.

    ReplyDelete

Please do not use blog comments for support requests! Blog comments do not scale well to this effect.

Instead, please use Launchpad for Bugs and StackExchange for Questions.
* bugs.launchpad.net
* stackexchange.com

Thanks,
:-Dustin