Interview: Ubuntu Takes on Hyperscale Using Open Source Technology

Print Friendly, PDF & Email

UbuntuLogo

With the massive amount of data created daily by users, many companies must  quickly scale out their resources to handle huge workloads. We caught up with Christian Reis, VP of Hyperscale for Ubuntu, to discuss these challenges.

insideBIGDATA:  What is it about Ubuntu that makes it so popular as an OS?

kikoChristian Reis: Ubuntu makes it easy to deploy and use the widest range of applications possible: from nginx to memached. We focus on making Ubuntu the best quality we can – everything should work beautifully. Not only do we deliver the widest range of applications, but we deliver them like clockwork every six months. This means that anybody who needs the most recent version of an application can get it easily. Another quality that users like is that Ubuntu is available everywhere – both on the desktop and on the server – making it an ideal OS for developers.

Ubuntu has become today’s fastest-growing server platform on the back of a core set of characteristics:

  • Workloads included: Ubuntu is the way the majority of applications and workloads are delivered to the end-user: from nginx to memcached to MongoDB, we include everything. With Ubuntu you don’t need to look anywhere else for your server applications.
  • Quality: we’ve put in place process, tools and infrastructure that naturally builds in and improves the quality of Ubuntu
  • Economics: in delivering Ubuntu free of charge to end-users, we remove a major barrier to adoption and, as is particularly desirable in the server space, to scaling out to an entire datacenter. But differently from other freely available distributions, Ubuntu has unwavering backing from Canonical, which is committed to the evolution of the platform, and to the delivery of the highest quality commercial support both for applications and operating system. That is a unique combination.
  • Relentless Technological Progress: not only do we deliver applications and infrastructure, we do this in the form of a validated and maintained release, like clockwork, every six months. This means that anybody who needs the most recent version of an application, library or kernel can rely on it being delivered on our release date in a stable form.  And we always meet the release date.

insideBIGDATA:  Please tell me a little bit about hyperscale, what does that mean exactly?

Christian Reis: Hyperscale is how Canonical defines next-generation server platforms, beyond standard rack-width servers and even modern blade-based systems. We see servers evolving to a model of systems assembled from a combination of compute, storage and network units. In practice, this means different, smaller form-factors, a variety of suppliers for each category, and configurable fabrics to interconnect them. HP Moonshot is a great example of a hyperscale system, and we are encouraged by HP’s commitment to break new ground.

insideBIGDATA : Where and how does Ubuntu fit in with hyperscale?

Christian Reis: If you keep in mind the core set of properties outlined above, as a server operating system alone Ubuntu is a perfect proposition for hyperscale: cutting edge, free to deploy at scale, and packing any workload you would want to run on that class of hardware. And in hyperscale we bring an additional advantage, which is the unlocking of smart differentiation on the hardware side: through flexible workload deployment provided by Juju and our Cloud Infrastructure, we can optimize, tune and configure specialized hardware while selecting exactly the right workload to run on it.

insideBIGDATA: Let me ask you about Big Data. Is it, in your opinion, hype or here to stay?

Christian Reis: The term and current marketing around “Big Data” may be over-spun, but it’s clear that the ability to store and process lots of data will unlock many interesting use cases that were simply not even considered to be requirements in the past. There are certainly going to be an increasing number of companies and sites that focus on analyzing the data they are generating and collecting with an eye to generating valuable insights – answers to questions like “in what geographical location do my products sell the most?” and “what demographic buys my highest margin products?”

insideBIGDATA: What role do you see Ubuntu playing in solving Big Data problems presently?

Christian Reis: If you consider what people doing big data analytics want from a platform – the latest software infrastructure, no per-unit licensing friction, great hardware compatibility, a reasonable and controllable commercial support bill – Ubuntu is really an obvious choice.

insideBIGDATA: What about in the future?

Christian Reis: Ubuntu’s core ability is to deliver the most up-to-date applications on any substrate at the highest quality and lowest cost of any operating environment. Big Data is an area unlocked by recent technological evolution, but in half a decade that technology will have evolved to be widely accessible and cheap, and there will be new approaches to solve another set of problems – ones we aren’t even considering at the moment. And that’s where Ubuntu is a perfect match: we will be here to deliver the applications of tomorrow, on top of system-level infrastructure which just works, and at a lower cost than any other platform.

Speak Your Mind

*