When computing is performed at the edge, data is collected at the source from sensors and then some amount of lower-level processing takes place. Processing data at the edge enables real-time decision making by eliminating the need to send sensor data to the cloud for processing, then waiting for the results. Artificial intelligence (AI) will, in most cases, occur at the edge where the Internet of Things (IoT) devices tend to reside and where decisions need to be made in real-time. However, there will still be some cases where AI activities will need to also take place in the cloud. This is generally those that are more compute intensive.
Cloud servers, for example, are used to store data, perform calculations on a massive scale, and possess the necessary bandwidth for long runtimes. Cloud-based computers also handle AI model training and the building of complex AI algorithms.
With this massive computing taking place in the cloud, lower-level Intel Atom and Arm processors can be deployed at the edge. This allows analytics, AI, and the Industrial IoT (IIoT) to converge for better decision-making and productivity, while putting compute-intensive training and heavy bandwidth requirements where they belong, in the cloud.
WINSYSTEMS’ industrial embedded SBCs and data-acquisition modules are examples of solutions that provide the necessary gateways for AI data to flow within and between the various computing environments. For example, the WINSYSTEMS’ small-footprint SBC35-C427, based on Intel’s Atom E3900 series processors, includes three simultaneous display outputs, dual DisplayPort, and LVDS with digital backlight and touch support for a host of IIoT applications. The board also includes two 10/100/1000 Mbit/s Ethernet ports and a host of USB and serial interfaces. The operating temperature ranges from -40°C to +85°C to address rugged environments.
The SBC provides edge computing with multiple configurations and expansion options for added functionality. It enables more processing to take place at or near the data sensor, providing up to 8 GB of ECC DDR3 RAM.
In comparison, the WINSYSTEMS’ SBC35-C398Q quad-core single board computer delivers high performance and rugged durability while reducing power consumption. This comes thanks to the SBC’s Arm-based microprocessor. Also providing -40°C to +85°C operational temperatures and excellent efficiency, it’s well suited for demanding graphical applications. The series delivers an IO60 expansion connector for additional functionality supporting I2C, SPI, TTL-UART, and PWM signals for stackable expansion through off-the-shelf or application-specific designed modules.
From the Edge to the Cloud and all points in between, data security is a must. WINSYSTEMS also helps clients to secure the data integrity between their AI at the Edge and cloud-based servers through encryption technology employing NXP Layerscape processor architecture. This occurs on next-generation networks with up to 100 Gbit/s performance and enhanced packet processing capabilities. Layerscape network processors allow users to fully exploit the hardware for maximum optimization and to easily adapt to network changes for real-time control over the network.