Compute at the Edge: Mechanical and Thermal Design Considerations

摘要

Edge computing extends applications, data, and compute from centralized points to the logical extremes of the network, thereby shifting processing and analytics closer to the source and consumption of data. A distributed computing model is driven by requirements for latency, bandwidth cost and availability, security and network connectivity. We are seeing significant growth in internet of things (IoT) markets worldwide, driven in part by a dramatic increase in vision applications, particularly those leveraging artificial intelligence (AI). These imaging and video use cases span nearly every IoT segment. They include finding product defects on assembly lines, managing inventory in retail, identifying equipment maintenance needs in remote locations and enabling public safety in cities and airports. They all leverage high-resolution cameras and create extraordinary amounts of data that needs to be aggregated and analyzed at the edge.This paper will review the different IoT market segments and form factors that are most common within them. Enabling edge compute has direct implications for designing systems to operate and survive in challenging environments and demands that a system level approach to design with thermal and mechanical considerations becomes more crucial. One of the more challenging Edge applications is the Outdoor Edge Device that must endure a wide operating ambient temperature range, employ sealed IP-rated housings, operate in a fanless mode, are exposed to solar loads and have a long service life. The paper will present data collected from two different solar load studies that will reveal how impactful planning for solar loading can be in edge compute. Lastly this paper will discuss what can be done from the silicon and packaging design perspective to help enable these more stringent use cases.