The Open Source Era: A Q&A With Canonical CEO Jane Silber

Canonical, a 750-person company with employees in more than 42 countries around the world, is the driving force behind Ubuntu open-source software. Although Canonical and Ubuntu are well-known and well-respected among hardcore technologists, most consumers have probably never heard of either.

This is an unfortunate reality of open-source software. Products and projects dedicated to democratizing technology by making computer use free and fair for everyone often fly under the radar. Whether Canonical and Ubuntu become synonymous with the general consumer is largely dependent on whether or not consumers move away from traditional device usage. Can Canonical's vision for a converged computing experience across a spectrum devices make the Canonical name as synonymous with desktop users as it is with users of its enterprise cloud and application performance management (APM) solutions?

I chatted with Canonical CEO Jane Silber, a remarkable executive with a rich technological background, over email about the challenges Canonical faces in consumer computing and even television, as well as how the company plans to maintain its status in the enterprise cloud and software markets.

[Editor's Note: Even though Silber was born in the US, Canonical is based in the UK, so please forgive her for Britishisms such as "favourite," "customising," "customisation," "customise," and "organisations."]

PCMag: Ubuntu advocates are rabid evangelists. They take to Facebook, Twitter, even the PCMag article comments section, to vocalize their love for your software. It's free, it's open, it's easy enough to use. However, it's still relatively niche as compared to other operating systems (OSes) (third behind Mac OS and Windows).

Why do you think this is? What will it take for Ubuntu to reach the next level of industry adoption? Do you envision these increases happening with the consumer base or with the enterprise adopter?

Jane Silbur (JS): Ubuntu has industry adoption that is both broad and deep. Companies such as Walmart, Netflix, and eBay build their infrastructure on Ubuntu. Telcos such as Deutsche Telekom, AT&T, and NTT build their next-generation telecom capabilities on Ubuntu. Ubuntu is the most popular operating system on public clouds such as Amazon Web Services, and companies such as Google and Intel use Ubuntu on their developer workstations. Further, Internet of Things (IoT) device manufacturers of gateways, networking devices, robots, and drones all use Ubuntu at scale. It's hard to think of that niche usage!

The place where Ubuntu does lag behind in usage is in traditional desktop computing, although it remains the leading Linux and the favourite of developers. We believe there will be a significant reshaping of this segment in the future as the notion of "personal computing" changes with innovations in hardware, software, connectivity, etc. Our vision is that, rather than discrete interaction points focused on screen size (e.g., phone, tablet, laptop), there will be a spectrum of devices and interaction models in our digital lives. We're building a converged Ubuntu Personal experience for that scenario, and expect changes in how people interact with their computing experience to occur in both consumer and enterprise spaces.

PCMag: As companies become expert in managing cloud infrastructure, the ability to customize to their specific requirements and preferences presents an opportunity for Canonical. As an OpenStack partner, what should businesses that have worked with traditional vendors know about open-source cloud deployments? How much open source better service their specific needs?

JS: We have worked with an incredibly large range of telcos, large enterprises and small to midsize businesses (SMBs) building Ubuntu OpenStack clouds. One of the workshops we often do with customers is to really understand and map their needs. Sometimes customers have solid business or technical rationale for customising a cloud architecture to fit their needs; sometimes the flexibility and freedom of open-source solutions leads to unnecessary and unhelpful differentiation.

We think it's important to help businesses understand their need for customisation and weigh that up against the benefits of standardisation at that level. We certainly support customisation and have built the largest OpenStack Interoperability Lab (OIL) that I'm aware of. In that lab, we work with hardware and software partners to build and test thousands of clouds per month, testing the interoperability and performance of various permutations of vendor software and hardware in an OpenStack environment. Because, certainly if you are going to customise your cloud, you want to know that the components you're selecting are going to work well together.

Customers who know that their business value lies in what happens on top of the cloud are often interested in a much more standard offering. In that case, we apply all our experience and knowledge about building and running clouds into a managed service offering called BootStack in which we will build and run an OpenStack cloud for customers, until the point when they are prepared to bring operations in-house.

PCMag: I'm not sure many people realize Ubuntu is also built as a Smart TV OS. Can you tell us a little bit about where the open source market is for smart TVs? Are manufacturers working with you to produce this environment? How will an open-source TV setup differ from the proprietary setups we've seen on other devices?

Also, I realize you've said you have no intention to manufacture Ubuntu TV hardware but I have to ask again. Any intention to manufacture Ubuntu TV hardware?

JS: We have no intention to make TV hardware (or any other hardware). Instead, we work with device manufacturers of all sizes to bring Ubuntu-based products to market. Smart TVs are at the intersection of two areas which we find very interesting: smart-connected IoT devices and a converged view of personal computing. You can think of a smart TV as an IoT device because it isn't necessarily that different from a smart home hub. And, other than the amount of space it takes up on your wall, smart TVs are increasingly like tablets in terms of media consumption.

In the market, you can see TV manufacturers are taking both of these approaches, positioning their devices as one of many devices in your home and digital lives or positioning them as the center of your home computing experience. Through Ubuntu Core and Ubuntu Personal, we're focused on working with hardware manufacturers to bring a secure, reliably updated, app-enabled experience to this kind of device amongst many others—that's what gets us excited. And we're excited to work with leading hardware partners to do so.

PCMag: Open-source security: The lingering notion in the industry is that open-source products are less safe than proprietary tools. What is Canonical's stance on this? Why are these notions incorrect?

JS: It's an outdated notion that open-source products are less safe than proprietary tools. There are, in fact, reasons why they can be more safe (e.g., more widespread ability to review and fix code), but open source alone doesn't guarantee a safer or more secure experience. Security isn't produced simply by a licensing model or development approach. Security is also dependent on factors such as architecture, frequency of software updates, etc.

In fact, that's one of the areas we've really been focused on in the last year. We've revamped how applications and the core OS interact on Ubuntu-based IoT devices and personal devices, such as the phone. As a result of these efforts, we've introduced snaps, a packaging format which enables application isolation and transactional updates, both of which are key to a secure, reliable experience. And, of course, one of the benefits of open source is that innovation such as snap can spread and that is exactly what's happened: Snaps have become a universal package format for Linux and now work on a wide range of Linux distributions including OpenSUSE, Debian, Yocto, and Fedora.

PCMag: Canonical appears to have coined the phrase "Big Software." Can you talk about what this is and how it will impact the industry? Some of the early responses to this phrase have been negative, in the sense that it's too theoretical, impractical, and won't actually provide immediate business value. What do you think?

JS: "Big software" is a phrase to describe a change in IT that is already happening. The term doesn't refer to a new tool purporting to [be] an IT panacea; rather, it refers to an evolving phenomena in IT that industry needs to come to terms with. Fundamentally, there is increasing pressure for businesses to operate IT at scale and at speed, with complex architectures of interconnected services, all of which are evolving at a different pace.

Historically, companies managed a few applications or solutions, rolled out on a couple machines. Now they must deal with many applications and services across tens of thousands of nodes and multiple physical and virtual substrates. This is the era of Big Software. Business value in the age of Big Software will accrue to those companies who are able to leverage the change, without succumbing to the risks inherent in this new world. And the key to that is understanding the implications for deploying, managing, and operating these complex, interconnected architectures.

This new class of Big Software requires a model-driven approach to effectively configure, deploy, operate, and manage it, and often, specific expertise in the component services. Juju is a tool which automates model-driven operations for Big Software. It encapsulates experts' knowledge in Charms, which can be re-used in multiple models across organisational boundaries, with multiple operating systems and on multiple compute substrates. It is how many organisations model, deploy, and operate their Big Data solutions, their OpenStack clouds, or even their Kubernetes cluster. All of these are examples of Big Software, already present on the changing landscape of enterprise IT.

This article originally appeared on PCMag.com.