Mark Shuttleworth is the founder and former CEO of Canonical, the commercial company behind the Ubuntu Linux distribution. Today he holds the position “Lead Product Design”, a role in which he shapes desktop and cloud product strategy. I spoke with him recently by phone about the increasing role of Linux in the enterprise, and the shift from traditional enterprise computing to cloud computing.
Canonical and Ubuntu made a big splash early on by intensely focusing on a usable Linux desktop experience. They pared down the dizzying number of packages available in Debian and selected a few best-of-breed applications to install by default. The installation process was streamlined to be as easy and as intuitive as possible. Ubuntu was a huge success and quickly gained a passionate following.
Since its debut in 2004, Ubuntu has gone beyond just being an easy-to-install variant of Debian, and Canonical has worked to extend Ubuntu’s reach beyond the traditional desktop. Most recently Canonical has been pronouncing Ubuntu as the most popular OS for cloud computing environments; and they’ve also been trying to establish success in the enterprise data center.
Shuttleworth opened our conversation with a quick overview of what Canonical and Ubuntu have been doing of late. He articulates clearly that Canonical’s focus on quality is the same in desktop and enterprise markets. Just as consumer-oriented businesses are extremely sensitive to product flaws and issues that lead to customer dissatisfaction, so too are enterprise oriented businesses who focus on mission critical operations.
Canonical employs many practices that build a quality baseline that are good for both consumers and enterprises, asserts Shuttleworth. More specifically, any efforts Canonical might apply toward enterprise customers are not done at a cost to desktop success.
We’re in scale-out deployments
Ubuntu is now available as a supported operating system from Dell, HP, and other OEMs, which makes it much more viable for enterprise customers. But desktop and server certification is only part of the story: SAN systems, database servers and more all need to be fully supported before they can be used in an enterprise. I asked Shuttleworth whether Ubuntu was pursuing certifications from the likes of Oracle in order to gain more enterprise traction. “I wouldn’t use anything other than Oracle Linux if I was running an Oracle database,” was his response.
He told me that Oracle has made it clear to them that Ubuntu will not be a certified platform for Oracle databases. This didn’t bother Shuttleworth at all. “We’re in scale-out deployments, like Hadoop, OpenStack, nginx, Condor, etc,” he said. Shuttleworth believes that all of these technologies should be on every CIO’s roadmap for the next five to ten years, and that Oracle really isn’t relevant to this market.
Shuttleworth went on to say that virtualizing computing power is getting to middle age. In his opinion, there are good proprietary and open source solutions for compute virtualization. “That scene is settling down,” he said. But, according to him, storage and network virtualization are just getting going. Cloud solutions don’t often rely on SAN storage, but rather use Hadoop, Swift, Ceph, and the like. The commodity hardware underneath open source infrastructure — compute, storage, and network — is going to be key topic in next 5 years, and Ubuntu is right in the middle of this.
Ubuntu is the best place to consume these resources, thinks Shuttleworth — it offers frequent stable releases of these and other new technologies. Red Hat, Shuttleworth concedes, is still relevant for mission critical single server solutions. But he believes that Ubuntu is better for scale-out deployments. Similarly, in Shuttleworth’s estimation SUSE has a strong mainframe and POWER architecture relationship with IBM, and they remain relevant in those sorts of environments “but we don’t see them in cloud or scale-out conversations much.”
As popular as “the cloud” is, the reality is that many organizations aren’t yet embracing it fully (if at all). With this in mind, I was curious about Shuttleworth’s opinion of the value proposition for Ubuntu versus other, more established enterprise Linux distributions like Red Hat and SUSE. He acknowledged that traditional workloads were more likely to be deployed on those other distributions, but insisted that companies building internal cloud infrastructure are more likely to do so on Ubuntu.
And Ubuntu is still extremely popular for traditional web server roles, as well as a platform for handling big data and quick scalability. Shuttleworth mentioned Instagram’s use of Ubuntu with obvious pride.
One of the reasons Ubuntu is so popular as a cloud guest is that it is completely free. Red Hat doesn’t provide a free distribution, and CentOS doesn’t provide any support for their compiled version of the Red Hat sources. As such, Ubuntu offers the best of both worlds: free to deploy en masse, but with a paid support option available when it’s needed. This begs the question: if Ubuntu “wins” the cloud guest OS competition, how does that affect Canonical’s revenue stream?
Shuttleworth claims that as deployments grow, so too do paid support subscriptions. Without offering hard numbers, he said that Canonical has seen a very satisfying acceleration of paid customers. This is completely typical — first Linux gets deployed for internal development purposes, then it sneaks into skunkworks applications, and is finally recognized as a first class offering. At that time, support becomes necessary.
Ubuntu sits at the intersection of free software and users
My day job uses mostly Red Hat Enterprise Linux, and as such I track a number of upstream projects in which Red Hat participates and which might land in future versions of that distribution. I don’t track as closely the things that Canonical is doing. With that caveat, it seemed to me that a number of recent Ubuntu initiatives — Juju, Metal-as-a-Service, and AWESOME — had a decidedly Ubuntu-only feel to them.
I asked Shuttleworth about this, and what I perceived as the contrast between Red Hat’s upstream-first development policy. The Canonical founder got fired up in his response. I clearly struck a nerve with Shuttleworth. “There’s nothing Ubuntu-specific in any of these,” he told me somewhat curtly.
He went on to articulate that Ubuntu sits at the intersection of free software and users, and that they act on what they see — whether that’s fixing bugs or building new tools. Shuttleworth highlighted Canonical’s long-running support for GNOME, KDE, and XFCE, and observed that several patches were landed in Unity to specifically benefit other Linux distributions.
“I respect Red Hat, they’ve played an important role bringing commercial software to the mainstream,” Shuttleworth told me. But he took exception with the notion that Red Hat was somehow more “upstream first” than Ubuntu. He pointed out that Juju has been ported to Mac OSX, there’s nothing Ubuntu-specific in MaaS, and that AWESOME is simply a Python daemon not tied to any particular platform or distribution. Saying that any of these projects are Ubuntu specific is “like saying ‘DeltaCloud is Red Hat specific’,” Shuttleworth said.
He expanded on the issue of “contribution” by pointing out that it involves a lot more than just lines of code. Indeed, that alone is a poor metric for measuring contributions. There’s also design, quality, ease of use, leadership and other harder-to-track but vitally important contributions, all of which Ubuntu provides to different projects in different ways.
We are strengthened by diversity
I next asked what Canonical is doing, if anything, to encourage diversity in open source communities? Shuttleworth told me that Ubuntu had just recently updated their community Code of Conduct. Ubuntu, according to Shuttleworth, has led the use of codes of conduct in open source communities. This was an intentional decision based upon founding members experiences with vitriol, personal skirmishing in mailing lists, and other less-than-welcoming behaviour.
The Ubunutu community decided collectively to take a strong stand against this kind of behaviour. They wanted a community that was pleasant and focused on a shared view of bringing goodness to people, rather than one based solely on personal interests. “We are strengthed by diversity,” Shuttleworth said.
“Because we explicitly frown on flaming and hostility,” Shuttleworth said, “we have retained good people for a longer period of time.” According to him, it’s hard to participate long term in any open source project because of so much change: it’s hard to keep up. “If people are unpleasant to one another, the motivation to stick around diminishes greatly.”
He railed against what he called the “bad culture of ‘bro-gramers’,” where participants insult one another. Worse yet, according to him, was hostility between competing open source projects and companies. “If Microsoft said some of what Red Hat says about Ubuntu, the community would be outraged!” Shuttleworth exclaimed.
Society is not benefitted by software patents
I switched topics in our conversation, and next asked Shuttleworth how he felt about software patents, and how Canonical as a company felt about them? This was another topic about which he got fired up. He told me that he’s long been interested in the intersection of society, technology, and economics. The history of patents, he said, is grounded in the question “what will accelerate human progress?”
Patents were designed to get people to talk about their secrets, Shuttleworth opined. Industrial progress used to be all about keeping secrets — sometimes for generations at a time — but in Shuttleworth’s opinion science and society move faster if we can encourage disclosure. When one inventor talks about her insights, another inventor can build upon those insights in novel ways for the betterment of everyone. “You should only be able to patent those things you could keep secret,” Shuttleworth said.
“People have become confused,” Shuttleworth lamented, “and think that a patent is incentive to create at all.” No one invents just to get a patent, though — people invent in order to solve problems. According to him, patents should incentivize disclosure. Software is not something you can really keep secret, and as such Shuttleworth’s determination is that “society is not benefited by software patents at all.”
Software patents, he said, are a bad deal for society. The remedy is to shorten the duration of patents, and reduce the areas people are allowed to patent. “We’re entering a third world war of patents,” Shuttleworth said emphatically. “You can’t do anything without tripping over a patent!” One cannot possibly check all possible patents for your invention, and the patent arms race is not about creation at all.
“The challenge,” Shuttleworth continued, “is how to open up a legislative discussion not dominated by companies that have been successful in the past.” It’s no secret that the majority of funding and lobbying comes from people with a strong interest in blocking new entrants. The voice of the people — and the voice of the individual inventor — is simply not heard.
Canonical, Shuttleworth told me, is a paid-up member of Open Invention Network, but according to him this is “really quite distasteful.” “It’s like saying ‘I have friends with big guns’,” he remarked. According to Shuttleworth, Canonical does not file patents defensively or offensively; and although it would be straightforward for them to patent their work they don’t. They feel it would be actively harmful to what consumers want: ever improving products at ever lower prices.
I pressed the issue and asked about the Google vs Oracle lawsuit. I was specifically curious about the question of whether APIs were copyrightable. “As far as we’re concerned, this is a settled matter,” Shuttleworth stated. He said there are prior cases of people trying to copyright some kind of interface — mechanical, software, etc — and that these had all been resolved. “To countenance that would be to throw a spanner in the works of progress in general!” Shuttleworth exclaimed. “Technology should be easy to consume, and widely available. Innovation should respond to what customers what and need,” not to what established businesses feel they need to protect.
It’s fantastic up there!
No interview with Shuttleworth could be complete without the obligatory question: Would you like to return to outer space? His emphatic response: “Of course! It’s fantastic up there!” He proceeded to tell me how it was the experience of a lifetime. According to him, it wasn’t just the trip, but the entire experience of being immersed in an industry dedicated to exploration. “The folks I met were all wonderful, amazing people.” Despite his enthusiasm, he has no specific plans to leave the planet again any time soon.