Small form factors: a new embedded computing design paradigm

The notion of relatively small size, weight, and power consumption (SWaP) in aerospace and defense electronics generally suggests the ability to place more computing power in a smaller space.

Oct 1st, 2016
John Keller New

The notion of relatively small size, weight, and power consumption (SWaP) in aerospace and defense electronics generally suggests the ability to place more computing power in a smaller space. That's true, but it's only half the story. SWaP-constrained electronics technologies like small-form-factor (SFF) embedded computing does enable big power in small spaces, but the ever-shrinking size of today's electronics is helping systems designers think of space in a whole new way.

Small-form-factors for embedded computing - like COM Express, PCI/104 Express, the Smart Mobility ARChitecture (SMARC), and Mini Embedded Technology eXtended (Mini-ETX) - are shaking up the embedded computing design paradigm by helping designers consider new kinds of distributed embedded computing architectures, above and beyond traditional board-and-backplane designs.

Small form factors can help systems designers think about placing computer components where they make sense, rather than consolidating computing in board-and-backplane boxes. When it comes to digital signal processing (DSP), for example, designers often talk about placing computing components as closely to antennas and sensors as possible. Small-form-factor embedded computing may offer just the ticket.

Using a distributed embedded computing architecture based on small form factors can help designers place analog-to-digital (A/D) and digital-to-analog (D/A) converters with some field-programmable gate array (FPGA) pre-processing in a small package next to antennas and sensors, and then route this data over high-speed optical fiber interconnects to more powerful processors conveniently placed elsewhere in the system where size and weight are not critical issues. Moreover, small-form-factor embedded computing can help designers place components where they fit best, rather than worry about finding a space big enough to accommodate a traditional computer box.

Potential benefits of distributed embedded computing architectures don't end there. The whole idea of dividing computers into separate parts may offer advantages in thermal management, controlling electronic emissions, and graceful degradation when things go wrong.

A distributed computing system might tolerate overheating, power surges, or battle damage better than a centralized board-and-backplane computer box. If the worst happens and the computer box goes down, the mission might be over. Bring down a preprocessor memory packaged in small form factors and the mission might be able to continue, but with degraded performance.

Distributed architectures enable designers to spread out not only relatively small components over a larger system, but also waste heat over a larger area. Hot processors could be located near potential cooling sources like flowing air; components susceptible to electromagnetic interference (EMI) could be physically separated from EMI sources; and relatively cool-operating components like memory and controllers could go near the center of the system where heat is not a big challenge.

Serial networking becomes a central concern for distributed architectures, and confronts designers with crucial decisions on whether to use optical fiber, copper wire, or even wireless networking. Fast Ethernet and other advanced network architectures are making these decisions easier than they used to be.

Small-form-factor embedded computing is on the verge of introducing a deep new design paradigm for aerospace and defense applications. Possibilities are limited only by the imagination.

You can read more about the latest trends in small-form-factor embedded computing in the Technology Focus feature, entitled "The new frontier of small-form-factor embedded computing" on page 18.

More in Computers