What the embedded computing industry is preparing for is the next generation of standards for modular system designs. COM-HPC will be THE focus of nearly all embedded computing vendors. Education on ‘why’ and ‘how-to’ will be top of the agenda. This is because COM-HPC is not just new but also highly complex and sometimes even misunderstood, leading to uncertainties.
Some see the new PICMG standard as a completely new platform for completely new applications. The embedded edge server faction thinks that way. Their embedded system engineers want to process massive payloads in rugged environments. They demand Server-on-Modules offering switched 10 GbE plus connectivity. They are also keen on massive memory which will be satisfied by the up to 8 banks of onboard RAM the new server-grade COM-HPC modules provide. And they greatly welcome having a real open standard for their modular rack server and box server designs. This is because the ability to simply exchange a module for performance upgrades will save them about 50% of cost compared to entirely new 1U or 3U rack mount systems.
The second fraction are the COM Express users. They do not embrace the new standard. They want to protect their existing COM Express investments and are unsure how the market will behave in the coming years. They are driven by questions like Will there still be COM Express modules available? How long will I be able to get them? Do I really need to switch to COM‑HPC now? What are the benefits for my design or, respectively, my customer? Those developers don’t glance at the large server-grade modules. For them, it’s important to know what the so called COM-HPC client modules bring to the table.
Let’s hold on for a minute. I haven’t explained the new COM-HPC standard yet. It will define 5 new form factors: two for high-performance edge servers and 3 for clients with graphics – or let’s call them Computer-on-Modules for the standard embedded computing faction as they are familiar with this term – and of course also with modular designs, which have been standard since ETX Computer-on-Modules were released early this millennium. Especially in this expertise lies the answer to nearly all of these questions above. A leading COM standard has always been the guarantor for massive long term availability and so for all existing COM Express designs. COM Express Computer-on-Modules will never ever disappear quickly as thousands of OEMs use them in their industrial-grade applications, relying on the long term availability promise of embedded processors and embedded form factor standards. There is definitely no need to switch as long as the existing interfaces are sufficient enough for the embedded platform.
New designs are only recommended if you want to benefit from the following off-the-shelf functions: PCIe Gen 4.0, more than 32 PCIe lanes for massive I/O, as well as 4 video interfaces. Also, if you know you will leverage USB 4.0 and/or 25 Gb Ethernet, COM-HPC will be your standard. But one thing is clear: In the long term, you will have to switch anyway as new technologies will arise and older ones will become obsolete. So start planning your next major design review now. You will then know exactly how to embrace the next major evolution. But there is definitely no reason to hurry. So relax and continue to rely on the well established open standards and come to our booth to learn more about COM-HPC. Ultimately, the plans we haven’t disclosed yet in terms of Reliability, Availability, Maintainability, and Safety (RAMS) may make you eager to switch faster. Considering all the new stuff that is related to digitization, IoT connectivity and the management of all those digitized devices, we can easily imagine that this will be the case. We’ll certainly be very happy to discuss all those needs with you at our embedded world booth in Hall 1-358.