The rationale behind edge processing is to place analytics intelligence in the same place as possible with related assets. Since edge computing and its relationship to the cloud is still an ongoing work, the definition and architecture are still very fluid. Since physical space or resources may not be able to implement dedicated edge servers, intelligence may need to be embedded into existing infrastructure, such as gateways, PLCs, industrial PCs, or various other device clouds that exist on the connected factory side.
In essence, edge computing exists at the level of a single machine control system, operating locally and complementing the work of heavyweight applications hosted in the cloud. Edge applications can perform a task as easy as getting and storing data from multiple sources and filtering data before forwarding to the cloud. More sophisticated visualizations bring analysis and even machine learning to the realm of edge computing to generate smart responses in real time. The basic components needed to implement this complex vision include data ingestion, event processing engine, connected device management, user applications, and secure data transfer to the cloud (Figure 2).
Starting from the first principle, building a complete intelligent edge processing platform is a huge challenge. Another approach is to employ a hardware-independent software framework that provides basic building blocks such as event processing engines, device management, and secure streaming using protocols such as MQTT Lightweight Messaging Protocol or Secure HTTPS. Many of these frameworks are reaching the IIoT site, such as GE's Predix, Cisco's IOx, and FogHorn Systems' lightning platform. In addition to the functional components, these packages provide a variety of software development kits (SDKs) to allow users to run their own custom applications, as well as development environments that help configure the system and define analytics. These frameworks also provide tools for managing edge devices, including monitoring, control, and diagnostics.
Lightweight, resource-friendly single-board computers like the Raspberry Pi Foundation's Raspberry Pi 3 provide the foundation for this device. In fact, General Electric has demonstrated Predix machine software for edge devices running on such platforms. On the other hand, engineers who have access to more powerful industrial PCs in gateway devices or automation systems can take advantage of additional resources and compute performance to execute more complex applications. Desktop-level performance can now be used in proven form factors, such as PC/104, on the VersaLogic Liger development board, with an optional Intel i3, i5 or i7 (Kaby Lake) dual-core processor running at frequency Up to 2.8 GHz. These boards are extremely rugged and have MIL-STD shock and vibration resistance for use in harsh environments. With up to eight local digital I/O channels, a Mini PCIe port for further expansion, and a high-speed SATA storage interface, these boards can handle complex automation tasks and handle edge processing workloads. The Gigabit Ethernet interface makes it easy to connect to the Internet and the cloud through an enterprise gateway. In addition, the board includes a Trusted Platform Module (TPM) that allows hardware password acceleration and secure key storage, making it an ideal choice for autonomous devices that need to be highly resistant to hackers.