Software-Defined Anything: The Engine for Change
Do you hear that sound? Do you feel something pushing you forward? It’s the loud, strong winds of change, disrupting everything you’ve known before and reworking the landscape in ways you maybe dreamed of, but assumed would never be possible. And what are the winds of change bringing in? Software defined anything (or software defined everything), often abbreviates as SDx.
When software becomes capable of doing what was only possible with hardware before, it changes the scope of what you can do and even how you think. Features and functions expand exponentially. The market is disrupted, changing the way established players operate but also enabling unexpected new entrants. Costs come down and solutions have increased value-add.
A Repeating Pattern
There was an excellent wide-scale example of the impact of software augmenting hardware in the 1970’s with automobile engines. Emissions regulations on automobile engines led to hardware solutions, like the carburetor, that killed fuel efficiency. Then fuel prices skyrocketed, making that emissions control versus fuel efficiency a terrible trade off.
But then carburetors gave way to electronic engine controls. This was great because software could factor in many more sensors and have much more adaptable behavior to engine conditions. That allowed car manufacturers to recover the lost fuel economy and even add new features. Eventually, software-based engine controls led to the ability to stop/start engines at stop lights and driver-selectable “efficient”, “sport,” and “ludacris” performance modes.
Implications for Buyers
So what does this shift towards SDx mean to infrastructure buyers? The improvements in hardware from generation to generation are slowing down. We’re getting more compute resources, but it’s harder to take advantage of them. But, by using those resources for new features and SDx, we enable the explosive power of software to make possible what was only a dream before.
You’ve probably already heard of software-defined storage (SDS), which uses policy-based provisioning and management to increase the flexibility of data centers. But there is so much that’s possible now when you apply software to storage problems.
How Vendors Respond
In our case, for example, the Penguin Computing®️ FrostByte®️ line of storage solutions not only uses SDS, it also includes other features that would otherwise require dramatically more expensive and proprietary hardware but still offer fewer features. But, by combining this underlying software technology with additional file system innovations, we can deliver effective, scalable file systems at reasonable costs.
But, on top of the baseline value of adding software to help gain more value from hardware, the technology market has one major advantage that automobiles and many other markets don’t — open systems. Software-defined networking (SDN) is gaining traction because of the flexibility it brings. Now think about what more you can get from SDN if its not vendor locked.
Expanding the Power of SDX with Open Technology
Software stacks based on Linux and open source technologies mean the features and functionality explosion gets even bigger and costs go down even further. It’s in the name. Open SOURCE. It’s software by definition. And OPEN software that you both contribute to and take advantage of will allow you to ride that wavefront of expanding functionality and drive it even faster as you and your peers think of new things that are possible that you could only dream of before.
For example, the Penguin Computing®️ Arctica®️ line of network switches lets customer choose their own software to manage the switch. Customers can choose to make the switches just an extension of the plethora of open tools they may already use to manage their server infrastructure at scale. No additional cost for integration, even more flexibility. Thanks to open technology, network switches become just another server type and automation gets dramatically simpler.
Some infrastructure companies, like Penguin Computing, are adapting more easily to the SDx world. We have a long history of working with software (recall that we acquired Scyld Software almost 15 years ago and have a robust collection of cluster management and other software offerings). But, for others, integrating SDx may be a challenge because of the very disruption we talked about before and how it’s impacting infrastructure buyers. User needs are changing and we’ve found that you have to know the full infrastructure ecosystem with a high degree of expertise to see where SDx can add value rather than just be an unnecessary add-on. To paraphrase Avatar, with SDx its learn fast or die.
Implications Going Forward
I’m really looking forward to are all the things that SDx spawns that we can’t even imagine. An SDx world will put a lot of pressure on the infrastructure that has to support all the use cases that will emerge. Software cycles will shrink and even hardware that’s optimized for software cycles will evolve faster. All this, both the pressure and the increased speed of innovation, will no doubt combine to generate solutions we can’t even imagine today. And I can’t wait!