Virtualization Technology News and Information
VMblog's Expert Interviews: Canonical Talks Big Software


CIOs have always had to deal with the challenge of scale, but that challenge has itself scaled beyond the capabilities of traditional methods and practices.  The unfamiliar and daunting challenge of Big Software is on CIOs' doorsteps, and they need a new approach to their enterprise technology.  Canonical seems to be anchoring its development of cloud deployment services and technologies on solving the issues of Big Software. 

To find out more, I spoke with Anand Krishnan, EVP and General Manager for Cloud at Canonical, to get an explanation of the challenges and the potential solutions for CIOs in the age of Big Software.

VMblog:  How do you define Big Software?

Anand Krishnan:  CIOs have always had to deal with the challenge of scale, but that challenge has itself scaled beyond the capabilities of traditional methods and practices. They are now facing the prospect of a gargantuan network of fluid, flexible, ever-changing infrastructure driven by demands that change hourly.

Software in this era is different. The cutting-edge workloads of our time - big data, machine learning, software-defined infrastructure - are all composed of multiple moving parts.

This is Big Software. It's the battle that CIOs are facing in a cloud-driven age: running multiple apps across multiple platforms while answering questions of scalability and cost, public or private cloud, security and interoperability. And, it's a challenge that, for many CIOs, is unfamiliar and daunting.

VMblog:  What are some of the pain points that legacy enterprise IT systems face as a result of Big Software?  Why are they so ill-equipped to handle this industry shift?

Krishnan:  It's not that the enterprise is ill-equipped, so much as that this is a different class of problem that needs a different class of tools. Enterprise IT has become very good at  dealing with monolithic applications, chosen from ‘best-of-breed' providers and installed on a relatively small number of servers. That is where they have drawn much of their success before the age of the cloud really took over.

The new Big Software era asks different questions of them, ones that they may not have been asked before. These are questions of scale, cost, security and interoperability and having to choose between public and private cloud, or both.

And, on top of it all, these questions are being asked at a time when environments and demands are changing at an exponential pace.

The good news is, there are answers out there to tackle Big Software. That's why Canonical is so focused on bringing these problems to light, and showing IT that there are tools and practices they can adopt to take on the challenge.

VMblog:  What are the cost benefits of a software-defined approach to running IT operations?

Krishnan:  The benefits are two-fold - one, relying on software tooling means you can now scale your infrastructure without needing to scale your headcount alongside it. That's crucial at a time of near-full employment in IT. 

And, two, you've freed up your team's time and resources to focus on higher-level projects that drive greater benefit to the company. Both of these can reap significant rewards, in reducing costs and increasing revenue.

VMblog:  How does Canonical's cloud-native tooling such as MAAS and Juju help remove the complexities of Big Software?

Krishnan:  Our cloud-native tooling removes the complexities of Big Software by automating the processes that would ordinarily need to take place, to get infrastructure running smoothly for its intended purpose.

Let's take the example of manufacturing. A robot arm can be programed to build a cardboard box, and can then be left to its devices to construct hundreds of boxes an hour, with the assurance that 99 percent of the boxes it constructs will be to the specified standard.

The vast majority of processes in the cloud are the IT equivalent of cardboard boxes, and our tooling automates those processes to save time and resources.

MAAS, for example, spins up bare metal into fully-operational servers with just a few clicks, as opposed to hours of manual repetitive setup. Juju offers IT the choice of hundreds of Charms, which are the software equivalent of those robot arms, on their own and in configurations that we call ‘bundles,' to spin up individual and interdependent ecosystems of applications for everything from data analysis to large scale database management or even container deployments. 

VMblog:  What is the role of containers in a Big Software era?

Krishnan:  Containers are just another tool in the toolbox for developers; and, as they flock to micro-services-based architecture, the use of containerized applications will only grow. The implication is that this trend will drive all of the infrastructure characteristics of Big Software, such as large numbers of nodes, elasticity, running applications composed of multiple, inter-changeable services.

Running container-driven infrastructure at scale is the classic Big Software use case, and we believe this will only be successful with the adoption of model-driven automation.

VMblog:  How does Canonical support the transition of legacy IT systems to the cloud?

Krishnan:  Our role really comes into play when we consult with IT teams on a larger scale, and show them the tools that are available to build their new cloud environments to meet their needs. We offer them support throughout their cloud deployment journey and help them transition to the new environments, and continue this support beyond the initial transition to tackle challenges as they arise and assist with upgrades in the future.

What's more, by building tools like MAAS and Juju, we support the transition by making it possible to build cloud environments that are manageable, upgradeable and proven to work with Big Software applications.

Finally, for companies that want the entire private cloud to be built and maintained for them, we offer exactly that with our BootStack product. We use our expertise in OpenStack, and our model-driven approach to software, to create private clouds that are manageable and upgradeable without compromising the developer environment or long-run TCO.

VMblog:  How do telcos differ from enterprises when it comes to adapting to the age of Big Software?  What lessons can the enterprise learn from telcos?

Krishnan:  Telcos aren't just meeting the Big Software challenge, they're meeting the urgent, ever-increasing network capacity needs of today's population. These are companies that are being driven by a strong business case for NFV: every telco we know is jumping headlong into virtualized networks and their internal teams are forced to adapt. Failure to do so will mean an inability to function in a way their customers demand.

The complexity of telco infrastructure, and the NFV projects they are undertaking, stretch the bounds of the software stack and the tooling in use. This is pioneering work that every enterprise can leverage, both in terms of the tooling they develop, and the practices they adopt to succeed. 

VMblog:  In general, how should executives respond to ‘phase shifts' in IT like Big Software?

Krishnan:  Don't hold an organization back with the phrase, "that's not how we do things." Be prepared for significant, uncomfortable changes in the short-term, because they will allow IT to break free of its legacy shackles and scale at will.

IT is used to setting standards for tools and products, limiting the ability of their developers to experiment. In the age of Big Software, executives need to listen to their developers. Most often, they are testing cutting-edge tools and methods that could be adopted on a much wider scale to benefit the organization. Cultivate an environment where they have the ability to safely test these new ideas and demonstrate their potential to their teams.


Published Tuesday, September 19, 2017 7:53 AM by David Marshall
Filed under: ,
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<September 2017>