The Ubuntu Developer Summit, starting today in Florida, is a gathering of Canonical employees, industry partners and Ubuntu community members to “define the focus and plans for [the] up-coming version of Ubuntu”. That version, 12.04 codenamed “Precise Pangolin”, will be released in April of 2012 and will be the next Long Term Support (LTS) release of the distribution. The changes scheduled for 12.04 are interesting, and simultaneously represent the current state of the art of the Ubuntu distribution as well as represent the foundation on which future developments will be built. I spoke with Mark Shuttleworth, Ubuntu’s Benevolent Dictator For Life, about what to expect in Ubuntu 12.04 and beyond.
First and foremost, Shuttleworth pointed out that the support schedule for 12.04 has been extended from three to five years on the desktop. Ubuntu LTS releases have historically provided three years of support on the desktop (five on the server), with new LTS versions coming out every two years. With 12.04, desktop users will enjoy support through 2017, which is a pretty long time to offer support for a desktop operating system.
I asked Shuttleworth what the motivation was for extending the 12.04 support offering on the desktop and he immediately responded “corporate deployments.” There are a number of companies standardizing on Ubuntu for their desktop computing needs, and the three year support from prior LTS releases was proving inadequate. Large corporations don’t adopt the latest and greatest technologies quickly, and they often have long QA processes to ensure that new desktop operating systems will properly function with their fleet of legacy applications. (Just look at how many companies are still in the transition process from Windows XP to Windows 7!)
Shuttleworth pointed out that there’s a growing change in attitude toward corporate computing. As more and more applications move “to the cloud”, IT departments are seeing that there’s less and less need for expensive desktop lock-in. Something like Ubuntu provides sufficient computing resources for a great many environments, and is free from licensing fees.
Ubuntu 12.04 will be the conclusion of the current two year LTS cycle, which will include four total releases: Maverick Meerkat (10.10), Natty Narwhal (11.04), Oneiric Ocelot (11.10). After the release of Precise Panglon, Shuttleworth and Canonical plan to get cracking on 14.04, the next LTS version (not yet codenamed). The work building up to 14.04 will, according to Shuttleworth, challenge the Ubuntu community and Canonical’s partners to exceed the status quo.
In the meantime, the conversations at the Ubuntu Developer Summit this week will be the beginning of public conversations around the notion of “Ubuntu on devices”. There’s already been a lot of internal conversation on this subject, and some engagement with industry partners (witness the recently launched Ubuntu webbook in South Africa), but UDS will really open these conversations up for community participation.
Shuttleworth reminded me that the Ubuntu tagline has always been “Linux for human beings.” He went on to note, though, that human beings computing — and will continue to compute — using a lot more than the legacy PC. Webbooks are just one small subset of devices on which Shuttleworth has his eye. There’s also phones, tablets, and “smart screens” — intelligent devices with limited input mechanisms like televisions, in-dash automotive displays, and more.
Unity, the user interface rolled out in Ubuntu 10.10 Netbook Edition and officially adopted as the default desktop interface in Ubuntu 11.04, lays the foundation for future development of Ubuntu on devices. One of the major thrusts of Unity was to embrace alternate computing form factors while still presenting a reasonably consistent user interface. Unity was refined in Ubuntu 11.10, and will continue to be refined in 12.04 and beyond. The lens framework will grow, and presumably third-party lens contributions will make the Unity experience smoother for all Ubuntu users.
It’s almost impossible to talk about Unity without hearing some vocal body of people complain about it. As a long-time Ubuntu user, I admit I found the transition a little challenging and I still experience some frustrations with it today. I asked Shuttleworth about this, specifically looking to learn what Canonical might know about user feedback that doesn’t get heavy publicity.
Shuttleworth was obviously prepared for this kind of question because he started his response by saying “We don’t know exactly how many people are using Ubuntu, but we have a pretty good estimate based on download numbers and watching browser user agent strings. I can tell you that Ubuntu 11.04, the first to offer Unity, was the fastest adopted version of Ubuntu to date.” Whether this was simply a natural growth in the Ubuntu user base, or people being curious about the new interface, is hard to tell.
Shuttleworth went on to describe the user testing that Canonical performs on each version of Ubuntu. They do real-world performance testing using randomly selected focus groups: people who may or may not be comfortable doing any number of computing tasks. Canonical asks these folks to perform a variety of real-world things (get photos off a digital camera and posted to Facebook, for example) and watches their performance. With each successive iteration of Ubuntu, they’re able to document how much easier these tasks are (or are not) and adjust their development priorities accordingly.
All key design issues are evaluated in the context of these tests. According to Shuttleworth, “people get much more done, from scratch, when presented with 11.10 versus 11.04 or any prior release of Ubuntu.” The obvious goal is to continue that trend into 12.04 and beyond.
“Ease of use is not incompatible with power users,” he went on, noting that some of the most vocal complaints against Unity come from the most technical of users: people who consider “computing” to be having a dozen terminal windows open across multiple virtual desktops. Shuttleworth acknowledged that Unity in its current form might not be ideal for those kinds of users, but noted that it’s not impossible to reconcile the two. As an example, he pointed out that a great many Linux developers are today using Macs running OSX. Clearly OSX has staked its livelihood on “ease of use”, and has been remarkably successful. And yet power users are using it more and more. Shuttleworth is sure that the Linux desktop can strike a similar balance. Expect to hear more about this from UDS this week.
One of the big features introduced in 11.10 was the Ubuntu Software Center. Of particular interest with this is the proliferation of proprietary, closed source software. I asked Shuttleworth about this. “We celebrate freedom!” was his immediate response. Users are free to get stuff done, and free to get it done using the software they want to run. According to Shuttleworth, it is not Canonical’s place to enforce their software morality onto their users. Sure, they’d love to see everyone being successful with exclusively free software, but on the whole they’re pragmatic enough to know that their users may want — indeed, demand — something else. According to Shuttleworth this stance is the result of many heated discussions inside Canonical, and not everyone is happy with it. Nonetheless, Shuttleworth is very proud of the fact that many commercial developers are working to run their software atop Ubuntu, and Canonical wants to support them as much as possible, with the Software Center and the developer.ubuntu.com portal.
After talking about desktops for a bit, I shifted the conversation to servers, virtualization and cloud computing. I noted that Ubuntu and Canonical don’t seem to be as aggressively pursuing virtualization as Red Hat, who is making a big push with their Red Hat Enterprise Virtualization product and the oVirt project. Shuttleworth pointed out that Ubuntu was the first major distribution to make KVM a core part of their distribution. With that said, however, he acknowledged that they’re largely favoring cloud over virtualization.
Shuttleworth observed that Ubuntu is the most popular guest OS across all public cloud environments. Canonical wants to extend that to be the most popular base OS for cloud servers as well. HP’s recent decision to use Ubuntu as their preferred host distribution is one small step toward that goal.
He identifies Amazon Web Services as the de facto standard for cloud solutions: anyone looking to deploy a cloud offering today must be API compatible with Amazon. AWS controls much of the conversation simply by enjoying “first mover” status. Despite much effort by other proprietary companies like Oracle and VMware, Shuttleworth doesn’t think any of these others have any chance of significantly influencing the cloud conversation.
Canonical is pushing “cloud by default” by providing all the necessary features to run a cloud infrastructure with their core server offering. Canonical has thrown their weight behind OpenStack, which Shuttleworth identifies as the clear leader in open source cloud solutions. OpenStack was distributed as a component of Ubuntu 11.10 and according to Shuttleworth you should be able to deploy an N-node cloud in 15 minutes using just the core Ubuntu Server download.
Cloud is just a flattening of infrastructure, says Shuttleworth: “it’s just compute, storage, network and credentials.” In his mind, “building 50 servers is pretty much the same as building a 50-node cloud.” That is to say, it’s not really hard — or all that time consuming — today to provision 50 physical servers. Rolling out a cloud infrastructure should be as easy. He expects to see this made even easier with Ubuntu 12.04.
This raises some questions, though, as to how Canonical will reconcile the rapid advancements in cloud technologies with the five year support cycle of Ubuntu 12.04. LTS versions of Ubuntu get support for the software distributed with it, but don’t usually get new feature additions. As cloud technologies advance, 12.04 users may be left behind. Shuttleworth observed two possible solutions. The first is that folks interested in more closely tracking cloud (or other technology) advances could easily switch to the non-LTS versions of Ubuntu, updated every six months. Another possible solution would be to make available accelerated cloud feature additions to 12.04 users via an optional software repository. There’s some precedent for this within Ubuntu for other fast-moving infrastructure components, so this might well be something to expect with Ubuntu 12.04.
Ubuntu has been a remarkably successful Linux distribution. It started out exclusively as an easy-to-use desktop distribution, but has since added advanced server functionality, pioneering cloud computing integration, and sponsored an awful lot of terrific open source software development. This week’s UDS will likely produce some interesting developments.
Mark Shuttleworth, as Benevolent Dictator For Life of Ubuntu, told me “change is a fact of life, usually for the better.” Keep an eye on Canonical and the Ubuntu community to see how they aim to make Ubuntu better.