Since our site rename, we’ve often been asked “Why did you rename SDNCentral? And what is this SDx thing?” To help explain the reasoning behind our site rename and why we think SDx (Software-defined everything) is a big deal and the future of infrastructure, we’ve put together an eight-part series that takes you through our thought process. We hope you’ll enjoy reading it as much as we did writing the series.
When IT people think about disruption, we think about innovations that displace earlier technologies and take things to a new level in our field. In their turns, software-defined networking (SDN), cloud networking, open networking, network functions virtualization (NFV), and the shift to merchant silicon/white box have all appeared poised to be the next disruptive technology that would revolutionize the networking industry. Indeed, early adopters of these software-driven technologies have reaped incredible efficiencies and economies of scale unimaginable a decade ago.
But more than just networking is being transformed by software. Physical items and functions from our daily life are also becoming software-defined, and this value shift to software holds implications far beyond the technology industry. As software-defined networking is eclipsed by software-defined everything, or “SDx,” what does that mean for the future of IT infrastructure?
We have a perspective as the founders and operators of SDxCentral (formerly SDNCentral) and as trusted advisors to C-level executives at Fortune 1000 companies who are developing next-generation, software-driven products and services. Through our conversations with infrastructure consumers and providers, we’ve learned that the shift to software-defined everything is disrupting traditional business models in just about every industry.
This shift is causing infrastructure companies to face an unexpected identity crisis: who are we, and what and where do we deliver value? We have spent months discussing these questions with end customers, service providers, and technology vendors, and we have developed a viewpoint around what SDx means for the future of IT infrastructure. Today, we’re kicking off an eight-part series where we’ll share our insights, early predictions for our industry, and tips on how to keep your organization competitive in a software-defined everything world.
The SDx Infrastructure Identity Crisis
We define “SDx” as any physical item or function that can be performed as or automated by software. Outside of IT, this includes apps such as Uber and Airbnb, apps on mobile devices, and Internet of Everything devices such as GoPro cameras, Nest thermostats, Phantom drones, and self-driving cars. Software-defined infrastructure (SDxI) is the next generation of infrastructure required to connect all of these software-defined devices and applications to their networks, each other, and ultimately to end users.
The SDx infrastructure includes not just software-defined networking, but also software-defined compute, storage, security, data center, perimeter, WAN…you name it. Over the last three years, SDN, cloud networking, open networking, and NFV have each received their share of excitement and attention for their potential to create new market opportunities or even save the networking industry from irrelevancy.
But lately, we see that perhaps SDx infrastructure companies have lost some of that early mojo. We’ve talked to network operators; seen disappointing industry earning reports from companies including Citrix, Extreme and HP; and heard from network vendors telling us they’re “waiting” for new revenue streams to kick in. The SDx world is evolving rapidly, and networking and infrastructure companies are trying to carve their places in it. On one hand, many early SDN and NFV adopters have reaped promised benefits in efficiency and flexibility. On the other, the proliferation of software-defined everything has blurred traditional boundaries across the board and sown confusion around who adds what value in the SDx ecosystem.
For example, when we speak with companies like GE or Coca-Cola, we see their infrastructure architectures look a lot more like a cloud service provider’s than they do traditional enterprise. Meanwhile, product companies like Brocade, Cisco, Ericsson, HP, and Juniper are rushing to add services to their offerings, while services companies like FishNet Security (soon to be called Optiv Security after its recent merger with Accuvant) are developing offerings based on open-source software on white box hardware.
With boundaries now blurred between who is an enterprise and who is a service provider, even the seemingly simple task of market segmentation becomes an existential exercise. Traditionally understood roles of customer, technology provider, and service provider are all up in the air. Before, every company was clear on which single role it filled. Now, a company may play multiple roles and find itself both competing against and partnering with the same organization depending on the use case or offering. This places friction on a still-developing ecosystem and adds to confusion and disruption.
Only now it’s not a specific technology that’s disrupting the networking world – it’s the value shift to software that’s disrupting traditional business models in just about every industry.
Does This Value Shift Really Affect Me?
This isn’t an identity problem for just a handful of networking companies. Unlike disruptive technologies that only affect companies in a few industry segments, the value shift to software affects essentially any organization that leverages technology. The implications on the infrastructure side deeply involve compute, network, storage, security, and everyone else in the industry. To remain competitive in an increasingly software-defined world, infrastructure companies must understand three important transformations underway:
- The business and technical problems customers are trying to solve are changing, which changes the use cases and the form factors/delivery models required to meet these new requirements.
- The customer for SDx infrastructure is different. It’s no longer just the compute, network, or storage guy at an enterprise, or head of network operations at a service provider. Today’s customer is just as likely to be a software development team or cloud architect, whether it’s at a new company like Uber or Vox Media, or an “old” company like Ford, Apple, or Coca-Cola.
- Old infrastructure models, frameworks, and operating processes will not be effective for long with new customers and new problems.
Let’s take a look at this last and most important point, which is really the inspiration for our series. With few exceptions, the industry as whole is still attempting to apply old models against the new SDx world. Under the old model, networking innovations meant doing the same things faster or cheaper. So for organizations following that model, the shift to software is literally just that: a change from hardware that stops at the form factor. In other words, a router that turned into a virtual router to run on a server instead of on custom FPGAs or ASICs still performs the same function.
This approach may be very helpful for cost reduction in legacy environments, but it wouldn’t be enough to effectively solve the routing requirements for an application like SnapChat or solve the virtual network requirement for IoT, such as connecting 1 million Coke soda fountains to Coke’s IT infrastructure. The result is a mismatch between the products being delivered by today’s infrastructure providers and the use cases and needs emerging from customers. That’s why traditional customers of IT are now becoming their own providers. Google has built proprietary routing and controller applications and switch hardware (while also becoming a mobile operator); Facebook is architecting and developing its own switches via the Open Compute Project, AT&T is developing its own SDN controller and CPE hardware; and Verizon is experimenting with white box switches.
How to Stay Competitive in an SDx World
We’ve discussed the value shift to software with numerous network operators, service providers, and technology vendors to help us understand what it means for infrastructure consumers and providers. With such fundamental changes occurring, we’ve decided to share our findings and current thinking in an eight-part series of articles for the SDxCentral community.
Our objective is not to define a fourth platform or set an industry-wide vision. Our aim is to share the perspective we have gleaned by working with leading cloud architects, IOT product managers, software developers, and network operators – and hopefully spark new ideas on how to keep your organization relevant and competitive in an SDx future. We’ll identify the broader macro trends affecting infrastructure, outline the key elements of the infrastructure of the future, look at who the new customers and providers will be, define key use cases, and walk though form factors.
We don’t claim to have the all answers, and we know our thinking will evolve as we learn more. We hope the series will kindle discussion about how our industry can best meet the needs of an SDx world, and we encourage you to share your thoughts along the way.
Read other posts in the series:
Read other posts in the series: