Canonical CEO: Ubuntu wants to power everything from smartphones to the cloud

CEO Interview: Canonical's Jane Silber dishes on going your own way, the perils of crowdsourcing and her bold vision for the future of Ubuntu

Jane Silber

Jane Silber

While you may not recognize the name Canonical, chances are you've heard of its Debian-based Linux OS called Ubuntu. We spoke to Jane Silber, CEO of the privately held UK-based company, about her transition to CEO, the company's past, and its plans for spreading Ubuntu everywhere from the cloud to tablets to smartphones.

What is your professional background and what made you want to first join Canonical?

My background is in software development and management of technology teams, programs and companies. I started my career as a C/C++ developer and have worked in areas as diverse as health promotion, defense, and artificial intelligence. I joined Canonical in 2004 because I immediately believed in the vision of a collaboratively developed, free-to-use, open source operating system, backed and supported by a company like Canonical. The fast-paced, agile nature of the community and the start-up opportunity also appealed to me.

With Ubuntu 13.10 “Saucy Salamander” having shipped in October of 2013, work on the next Long Term Support release, version 14.04 code-named “Trusty Tahr”, is now well underway. What is the most important change that this new release will bring?

On the client side, the most important element is the continued evolution of the phone/tablet/PC convergence. It's really exciting to see things like the seamless linkage between phone apps and the tablet side stage, and the app developer story across form factors.

+ ALSO ON NETWORK WORLD 10 things we love and hate about Ubuntu Oneiric Ocelot +

On the cloud side, the defining characteristic of the Trusty release is that it's an LTS and that means that our enterprise customers will deploy and depend on it for years. Given how quickly cloud technology is evolving, we test, integrate and make available the latest versions of key technology in our cloud archive so that telcos and enterprise users can have a stable OS baseline and still leverage the incredible advances in things like OpenStack and Juju. Particularly on the server and cloud, the 14.04 LTS represents a perfect storm of sorts for quality, freshness and interoperability.

In addition to your company’s well-known Ubuntu OS, Canonical works on a number of other open source software projects including Bazaar, JuJu, Upstart and Launchpad. Which is the most interesting to you and why?

Juju is an incredibly interesting project. The original concept for it started as early as 2009 when we recognized that the cloud world was going to require an entirely different approach to service deployment and orchestration. We felt the traditional view of configuration management wasn't sufficient for the cloud. From a management of technology and product development perspective, the evolution of the project has been fascinating as we were solving a problem that users didn't yet realize they had. There are others who have now recognized the service orchestration challenge, but I believe Juju remains the most advanced, well-conceived, most interoperable option to configure, deploy, manage and scale services on any cloud.

Does Canonical/Ubuntu give back to the development community as much as say Red Hat, SUSE, Intel or IBM?

Yes, certainly. Canonical contributes significantly to the development community, including providing the best free platform for people to develop on and the most efficient, lowest-barrier distribution path for getting developers' code to users. We contribute extensive interoperability testing, drive demand for other companies to support Linux, had a revolutionary impact on open source design and user experience, and contribute best-in-class tooling and practices that have been followed by numerous projects from Linaro to OpenStack. And of course we also contribute significant amounts of code, particularly for a company of 600 people in comparison to those 10x-30x our size. I am proud of our contributions.

Announced in March of 2013, Mir is Canonical’s new computer display server for Linux designed to replace the X Window System for Ubuntu. What makes Mir special and why not use something such as Wayland instead?

Fundamentally, we felt that there was no existing option which met, or was on a path to meet, our requirements. The technical pros and cons of various options have been hashed out elsewhere. Naturally we're developing Mir openly, and I refer anyone interested in more detail to for an update on motivation, specs, roadmaps, and pointers for getting involved.

Does Canonical going its own way on many projects limit its ability to interact with the wider community? For example, Mir.

There is a natural tension in the open source world between starting something new and contributing to an existing project, and there will be critics regardless of the path taken. We have never shied away from making opinionated decisions in Ubuntu, tracing back to our early controversial decisions to limit the distribution to a single CD, to include only one browser and one email client, etc. We take decisions to go our own way very seriously -- there is always a significant financial investment as well as the technical factors involved. But at the same time, our goals are disruptive and revolutionary, the tech world is moving incredibly quickly and we can't be afraid to forge our own path when necessary. Some will choose to join us on that path, some will choose another path; I believe that the open source community can respect and embrace different choices.

While Canonical is mostly known for its Ubuntu Linux distribution in the open source world, the company relies on enterprise services and solutions for much of its revenue. What enterprise and/or engineering service plays the most important role for the company now? Over the course of the next three to five years, do you see this changing, and if so what will help drive the company forward?

Our cloud related services for telcos and enterprises play the most important role in this area. We offer tooling and professional services to help organizations build, manage and maintain both public clouds, private clouds and the workloads on top of those clouds. In the coming years I think this area will continue to grow, as more organizations realize that the agility they need can only come from a cloud.

Does Canonical earn more revenue from its OEM partnerships or from enterprise services and solutions?

It's approximately even.

Ubuntu was initially a leader in OpenStack development, but looking at the recent statistics Red Hat has taken over that lead. What is the plan going forward with OpenStack? Does Canonical have a plan to take on Red Hat?

Ubuntu remains the de facto OS for OpenStack deployments, and a lead contributor in many categories. We are focused on improving the OpenStack experience for everyone, and delivering the best, purest and most advanced OpenStack experience to our customers in the configuration they desire. OpenStack isn't a one-size-fits-all proposition, and we recognize that customers will have different preferences for various hardware and software components. Our OpenStack interoperability lab and tooling which can help provide any workload on any cloud are all aimed at delivering the most reliable, up-to-date technology to our customers - at cloud-speed and cloud-scale.

While Canonical has a development stack (Juju), it does not have a full management offering for clouds. What, if any, are the company’s plans to fill in that gap?

We do have a full management offering for clouds. Our suite of cloud tools includes Landscape, a management tool that allows you to manage your OpenStack cloud, as well as instances running on any cloud or directly on metal. A full Ubuntu OpenStack solution from Canonical includes Landscape, a provisioning service (MaaS, or Metal as a Service) and a service orchestration service (Juju), as well as professional support and consulting services.

Our goals are disruptive and revolutionary, the tech world is moving incredibly quickly and we can't be afraid to forge our own path when necessary.

— Jane Silber

In recent years, more work on mobile versions of Ubuntu for smartphones and tablets has taken place at Canonical, from a demonstration of the Motorola Atrix smartphones dual-booting both Android and Ubuntu, to an Ubuntu installer for Nexus 7 tablets and later the Ubuntu Touch beta. While these have all been aimed at developers, Ubuntu still has yet to ship on a commercially available device to average consumers. Why is this and when do you expect this to change?

When we announced the phone and tablet in early 2013, we said that they would ship from a manufacturer in 2014. We're still on schedule to do that. You're seeing the interim progress and developer versions now because of our commitment to developing Ubuntu transparently and in the open rather than because there has been a change in plans.

In late July of 2013, Canonical launched a crowdfunding campaign on Indiegogo for a proposed high-end smartphone running Ubuntu at its core and designed by Canonical. While the campaign managed to garner a lot of attention in the media, it only raised just over $12.8 million of its $32 million goal. Given the incredibly high goal, did Canonical believe this to be an achievable figure or was the primary purpose of the campaign to gather interest from users and potential OEM customers?

We knew it was a very ambitious goal - there's no other way to describe a goal that is 3x the previous record. We explored different models (e.g., lower goal with a higher per unit pledge level, a higher goal with a lower per unit pledge level, etc.) but felt that this was the level that gave us the best shot at success. We absolutely wanted the campaign to succeed and I and others in the team were gutted that we didn't have the opportunity to make that phone. It is true that the interest and publicity associated with the campaign has had a positive effect in terms of demonstrating demand, generating interest from OEM customers, and ratifying the Ubuntu vision of convergence. But we also really wanted to make the Edge.

How would you respond to those in the community and media who say that Mark Shuttleworth could still easily finance Ubuntu Edge without a crowdfunding campaign?

That would defeat the purpose and spirit of a crowdfunding campaign. This wasn't about a Canonical investment decision and we have no intention to enter the hardware business. We said we would do the work to create a beautiful, high-powered phone if enough of the "crowd" wanted us to do so. While there was great excitement from the crowd in response to the offer, unfortunately we didn't meet the goal.

What do you think will happen to Ubuntu if your planned phone, tablet and TV products don’t gain any additional OEMs or market share?

Like any business, if something isn't working we will need to change or stop it. There may be alternative business models to explore, but fundamentally if there is no place for an Ubuntu phone or tablet in the market then there is no reason for us to invest in its creation. However, I don't think that's the case. Every indication is that we have an excellent product, that our vision for convergence is the right one, that the industry desperately needs an open alternative, and that consumers want choice beyond the beautifully landscaped but increasingly walled gardens.

What would happen to Canonical / Ubuntu if Mark Shuttleworth went away as a sponsor?

While he has some philanthropic motives, Mark's financial involvement in Canonical is as an investor rather than a sponsor. And he is a savvy investor. As CEO, it's my role to ensure that Canonical has sufficient funds to operate. Right now, between our revenue and existing investor, that's not a problem. If we were to need additional funds beyond what Mark is willing to invest, we have the same options available to us as any company.

If you were given an opportunity to go back and change one thing during your time at Canonical, what would it be and why would you change it?

There are some early decisions we took that now carry a different set of trade offs. For example, we made a decision early on to have different names for the project Ubuntu and the company Canonical. Our intentions were to avoid confusion but in retrospect the different brands have created another kind of confusion.

Another example is the impact of our distributed workforce. We have always been committed to hiring the best people, and our "distributed by design" mantra has allowed us to gather some of the best people in the world in Canonical. It's a real strength for the company, it helps ensure that we retain the openness Ubuntu is known for, and the ability to work from home and with such an incredibly talented and diverse workforce are some of the benefits that our team enjoys.

However, after having undergone several years of high growth and some periods of rapid change, we can now see the impact that distributed nature has in a company of 600+ on communication, team cohesiveness, and ability to embrace change. If we were going back in time, I wouldn't abandon the strategy of being fundamentally distributed, but would consider some guide rails or lightweight structuring around it which wasn't really needed in the early years but which be useful now.

In one sentence, how would you describe Canonical to someone who has never heard of it?

1 2 Page 1
Page 1 of 2
The 10 most powerful companies in enterprise networking 2022