RTC Magazine

Innovation. The lifeblood of the technology business. Just look at the last decade. iPods, iPads, smartphones, flat panels everywhere. Innovation is always good, right? It’s perhaps hard to believe, but this has been a topic of discussion among SFF suppliers during the past year or so. And the answer is not nearly as clear cut as you might think.

Remember earlier discussions of evolutionary vs. revolutionary change? Over time, the concept of revolutionary change became one that left technology users wary. Revolutionary change implied a clear break with the past, doing perhaps totally new things in a totally new way. Evolutionary change is less drastic with the implication of bringing the past along for the ride—yet providing significant new capability in a compatible way. We’ve beaten that one to death over the past few years. There’s clearly a much higher comfort factor with evolutionary change in the SFF business among system OEMs.

Both evolutionary and revolutionary change require innovation. Evolutionary change may be as simple as providing an additional PCI Express lane on a connector. Or using a less expensive connector to keep costs down. Or recasting a connector pin definition to make for more efficient designs. Or shrinking a form factor as new high integration components become available. All good innovations of value to many people. 

And yet there is a cost for some of these evolutionary changes. In particular, the cost may be felt most by those SFF designs for which an ecosystem has evolved. The differences in the impact of evolutionary change may be best understood by comparing the impact on two of the mainstay technologies of the SFF market—stackables (such as the PC/104 form factor) and COMs (such as COM Express). COMs have no hardware ecosystem per se. Because of the difficulties in creating interchangeable, interoperable COMs, COM users understand that any new COM, even with the same pin definition and form factor, will most likely require some change to the baseboard. There are few off-the-shelf carrier boards that plug into a COM connector. Changes to a COM pin definition, connector or even form factor have little effect on users or other hardware suppliers. Hence, as much as we chastise COM suppliers for the plethora of types (pin definitions), proliferation really doesn’t hurt anybody. 

That leaves us with stackables—the ubiquitous PC/104 (90 x 96 mm) stackable modules that can provide CPUs as well as I/O expansion for PC/104, EPIC, EBX or other form factors. The strength of this architecture over the years has been the broad set of suppliers and the incredible array of perhaps a thousand or more different boards available off the shelf. It’s easy to build a system that does almost anything using these off-the-shelf parts. The introduction of PC/104-Plus (adding PCI support) was an evolutionary change done in an upward compatible way and sustained the growth of the market. As PC/104 has evolved to support new bus technologies (PCI Express), there have been some disagreements among suppliers, and there are now multiple solutions available on the market. This fragmentation causes a lot of sleepless nights for some people, but is not as bad as it sounds. The market will ultimately decide who wins this battle—helped along by the growth of the ecosystem of these technologies. 

And just as PC/104-Plus grew slowly in market share, so will the new PCI Express implementations. It will just plain take time for these to take off. In the meantime, the growth of the ecosystem will build. It’s tempting at this early stage to tweak these standards—recast the pin definitions or change a connector—to make the design more “optimized.” Evolutionary change if you will. Innovation. But incompatible changes at this stage cause those responsible for the ecosystem to take a deep breath. They say “Maybe it’s not time to invest in this new technology. Let’s wait a few years to see how this develops.” Ecosystem growth would slow or perhaps come to a complete halt. Innovation, for all the right reasons, would deny the market the very thing it needs to grow from its infancy. 

We’re all engineers. We share a common trait—the need to endlessly tweak things to make them better. Tweaking the new PCI Express stackables standards is a scary proposition for ecosystem suppliers and system OEMs alike. Let’s live with what we already have and grow this market for everybody.