Military Embedded Systems

Making a case for small form factor embedded computing for military applications


May 27, 2020

John Reis


Defense funding for unmanned and artificial intelligence (AI) systems is increasing by leaps and bounds, driving the need for computing systems to be smaller, faster, and more compatible. The demand for compatibility and commonality follows the convergence of open architecture initiatives like the Sensor Open Systems Architecture (SOSA), C4ISR/EW [electronic warfare] Modular Open Suite of Standards (CMOSS), and Modular Open Radio Frequency Architecture (MORA), just to name a few.

The demand for AI is driving much of this trend toward small form factor (SFF) computing. Solutions that enable AI like Jetson and Movidious make it an easy transition into the smaller form factor realm by utilizing carrier boards and I/O solutions that will promote an advantage in reduced size, weight, and power - cost (SWaP-C). Intel and Nvidia are making this possible for the defense and aerospace market by supporting their chips sets on the embedded roadmap for 10 plus years.

While SFF computing has its strong advantages like SWaP-C, there are still going to be some challenges. Cooling, power, and processing power can be an issue in some applications. But on the flip side, meeting reduced SWaP-C requirements will be a challenge for the traditional 3U/6U embedded computing systems.

The SOSA Consortium has formed an internal working group to address SFF computing and and how it will eventually fit under the SOSA Technical Standard. Collaboration is key and we need to see a bipartisan relationship between VITA and PICMG to have a cross-functional mindset when looking at SOSA. SOSA can bridge the gap between these two working groups, enabling them to work together in a symbiotic way which has not happened to date.

In my opinion, we (vendors) need to show the benefits of a convergence of these specifications to both VITA and PICMG. This will not only generate significant interest for military electronics applications, but benefit both standards bodies to welcome new suppliers to their memberships.

Pressure on embedded computing suppliers to enable AI in defense applications is only going to increase - and this is a good thing. We can be 99.9% certain that AI is not going away - from your household devices for everyday living to unmanned air and ground systems, missiles, radar systems, electronic warfare applications, etc. We are living in a world of a technological advances like no other. AI will drive next generation computing platforms and small form factor computing is a critical part of that next generation platform.

Uncovering these applications and driving this initiative will be a first step in convincing the end user that they - and the program - benefit not only from reduced SWaP, but from the lower cost systems.

Another challenge that the ecosystem faces is our need to prove to users and to the application that we can satisfy the specifications and environment. We need to continue to change and re-align the defense customer's mindset to not look immediately at 3U/6U technology to satisfy their application. How do we do this? By working together in a cooperative yet competitive nature. Having standards like SOSA, CMOSS, and others to drive us to align and be compatible will be a good start. But not the only one.

The convergence of specifications such as CMOSS and SOSA are imperative to enable industry to have a unified family of products to satisfy requirements. Having too many standards to align and comply with is just too many cooks in the kitchen. We need a primary specification that has all the factions of and needs for all of the branches of the military.

In my opinion, this is the only way to make this recipe a blue ribbon winner, working as a team as well as separately to get the message to the masses. The best analogy now is what the current world landscape has provided us with: "TOGETHER APART."