The Cloud as a Tectonic Shift in IT: The Death of Operating Systems (as We Know Them)

Sacha Labourey

This is the third in a series of blogs on The Cloud as a Tectonic Shift in IT, authored by Sacha Labourey, CEO, CloudBees. In the series, Sacha examines the huge disruption happening across the IT industry today, looks at the effect the cloud is having on traditional IT concepts and reviews the new IT service and consumption models that have emerged as a result of the cloud. Finally, Sacha makes some predictions about where this tectonic shift will lead us in the future.

The move to the cloud represents one of the largest paradigm shifts to ever affect IT. More than just a simple technology evolution, the cloud fundamentally changes many of the cornerstones on which IT evolved. From redefining the concepts of operating systems and middleware, to revolutionizing the way IT services are built and consumed, the cloud is ushering in an era of change unlike any we have ever seen.

In Part 1: The Industrialization of IT, Sacha compared the evolution of electricity and its delivery. He compared the instant-on access and pay-as-you-go models in the power industry with the parallel evolution occurring now in IT service delivery.

In Part 2: The Irrelevance of Infrastructure as a Service, Sacha took a closer look at IaaS.

In Part 3, below, Sacha discusses The Death of Operating Systems (as We Know Them).

The Death of Operating Systems (as We Know Them)

 

software & infrastructure stack

The software & infrastructure stack

When adding new services to their IT portfolio, most companies will go through the same thought process: First, they’ll look for a SaaS solution that fits their needs. If they find one, they’ll use it. Otherwise, they’ll build a custom application using PaaS.

As you can see from that simple process, both decisions are made without any regard for the low-level software stack. When you are using SaaS, you only care about the services offered. When you are using PaaS, you care about the application runtime and services offered to you as a developer. These services include persistence, security, transactions, elasticity, failover and messaging.

By design, the underlying layers – hardware, OS, virtualization, networking, etc. – are handled and abstracted away by the PaaS or SaaS provider, making them totally invisible to the end user. So in the future, IT end users won’t have to care about the server-side OS. The PaaS and SaaS providers will.

Now, let’s dig one level deeper and look at how those PaaS and SaaS providers will behave.

What OS will PaaS Use?

When it comes to what makes the ideal OS, the majority of PaaS providers want something that is highly flexible, customizable, easy to upgrade and maintain and supported by an active community/ecosystem. At a high level, the feature set should be laser-focused, providing a hyper-robust OS that can deliver lightweight OS-level virtualization, fine-grained security and highly controllable resource allocation – with the lowest possible overhead. Ideally, this should result in a limited number of very small, well-defined OS images. These are frequently referred to as “Just Enough Operating System,” or “JeOS,” and are typically as small as 60 MB in size.

Several specialized Linux distributions match those constraints, which makes Linux the prime choice for many providers.

What OS will PaaS NOT Use?

Many may be wondering why PaaS providers don’t routinely use Windows. While Windows has a very active development community and vibrant ecosystem, and is pretty good at finely allocating resources among different tenants, it loses out as the OS for PaaS for two reasons.

First, Windows is not known to be the best OS when it comes to creating JeOS images and customizing, scripting and spreading them around. The second reason is simply cost! Why would a PaaS provider, using only a very small subset of an OS, spend a considerable amount of money per core for a full-fledged OS, when that value is not even going to be offered to the developer or end user? Windows is simply a no-go for the cloud and this will obviously have a long-term impact on Microsoft, since Windows is the main hub of their current ecosystem.

Now, does this mean Microsoft will be the only company impacted by this change? Not at all, even Linux vendors such as Red Hat will be impacted by the shift. Initially, as companies try using the cloud via the IaaS layer, Red Hat will do very well, as many organizations will choose Red Hat Enterprise Linux over Windows.

To understand the move away from today’s type of Linux distributions, it is important to realize that the unique value of these organizations lies in their ability and expertise in testing their distribution with a huge number of back-end devices and systems, and certifying an ecosystem of thousands of independent software vendors (ISVs). In PaaS and SaaS scenarios, the number of back-end devices being tested declines – AWS, for example, exposes the same hardware on its millions of servers – and the ISVs, having adopted SaaS, aren’t directly exposed to the OS anymore. In other words, the OS vendors are the ones orchestrating the ISV ecosystems. As ISVs morph into SaaS vendors, PaaS deployments – not the OS – become the new hub for that ecosystem, not the OS anymore. Consequently, as PaaS and SaaS offerings take the lead, cloud providers will use their own customized version of Linux – one that is either self-supported or managed by IaaS providers at a much lower cost.

What About the Server-Side OS?

When it comes to servers, the OS as we know it today will slowly fade away and disappear. In their place, we will see the emergence of self-supported, IaaS-supported and mini-Linux OS organizations focused on providing JeOS at a much lower cost. The time of the fully-featured, heavy OS has passed. Also, it is important to realize that this is not just a technically focused discussion: The move away from an OS-centric IT ecosystem will have important consequences for a number of well-established companies in that space.

What About the Client OS?

OS requirements will heavily change on the client side as well. Thanks to the rise of HTML5, CSS3, JavaScript and browsers will become the new unified interface. The underlying OS services required to run a browser-based environment will be very basic. As such, there will no longer be a need for the traditional fat-client OS. Obviously, HTML5 does not yet offer the complete feature set required to fully compete with native applications, so we will have to wait. But by the time, say HTML6 or 7 is released, there will likely be a device-agnostic client “universal plug” strong enough to defeat native/proprietary environments.

Unlike for the server-side OS though, the reason for this evolution won’t be so much an attempt to reduce the feature set offered – client devices will grow richer, if anything – but to converge towards the concept of the universal plug.

And as enterprises increase their usage of the cloud and move toward PaaS and SaaS solutions with an HTML5 frontend, the concept of the OS will evolve toward a much more focused alternative offering a fraction of the OS services available today. This change won’t just be a technical one, but will come with a complete reshuffling of how ecosystems are organized, and around which vendors they are centered.

Up next in the blog series: Part 4, The Death of Middleware.

To learn more:

           Sacha Labourey

Onward!

Sacha Labourey
CloudBees
www.cloudbees.com

 


The Cloud as a Tectonic Shift in IT:


 

 

Comments

Sacha, I agree that Windows might not be the first choice for PaaS providers, but I see more and more PaaS providers adding .NET support to their platform. As long as there is a .NET develoiper ecosystem and customers willing to write new .NET apps or migrate existing .NET apps to the cloud, Windows OS will need to be supported by the PaaS vendors. Let me know what you think. imtiaz.koujalgi@db.com

Imtiaz, you are correct, the existing .Net developers will be looking for a way to run their apps. Now, language ecosystems are not static sets: developers come and go. If MSFT is not as successful as they expected with Azure, Windows 8 and other offering, .Net developers are likely to go and find other language-ecosystem to monetize. So, as it stands today, the .Net ecosystem, thanks to its amazing size and inertia is not going to disappear anytime soon, but it is not clear to me that this ecosystem will grow or even stay at the same size. And remember, that the .Net developers play a relatively small role in the number of .Net servers deployed out there, a lot of them are being used to run pre-packaged applications (SQL Server, Exchange, Sharepoint, etc.), not custom applications.

Add new comment