Edge computing has emerged as a pivotal element in smart manufacturing and industrial IoT. Not part of the original vision of either Industry 4.0 or Internet, edge computing has nevertheless grown to fulfill a vital role in both the future factory and in remote industrial installations, such as mines, oil rigs and industrial infrastructure.

Bringing analytics, AI, and smart manufacturing control processing directly onto the industrial site, edge computing addresses the latency and throughput issues that can affect cloud based industrial analytics and data processing. For certain types of manufacturing, low latency is vital. Complex manufacturing automation can also require very high throughput as well as low latency.

A range of experts lined up to give their perspectives and advice on edge computing and its application to Industrial IoT and smart manufacturing in the recent series of IoT Days held by the IMC.

Speakers included Brandon Hart, Director of Technical Business Development at NimbeLink and David Smith, Vice President of IoT solutions, GetWireless.They provided the keynote speech for the session, which addressed ensuring successful deployment of edge devices.

A survey of the growing role of edge computing in Industrial IoT was the subject of the presentation give by Bill Dykas, Product Manager, Telit IoT Platforms.

Adopting a hierarchical approach to implementing edge computing, as the subject of Brandon Cannaday’s presentation, who described it as a “practical edge architecture for IIoT”.

Finally, Randall Kerr and Brandon Moser of Digi International spoke on the subject of edge computing as a remote gateway to critical infractructure.

In their keynote, NimbeLink’s Brandon Hart and GetWireless’s David Smith advocated a six step approach to implementing an industrial IoT infrastructure. This comprised assessing the state of IoT readiness for the organisation, including not only IT and OT systems, but also business processes, organisational structure and key stakeholder buy-in, as well as creating a business model of the IoT solution, architecting it, implementation and deployment.

Telit’s Bill Dykas spoke of the rising role of edge computing in Industrial IoT, and described a framework for achieving smart manufacturing through edge computing that integrates with existing industrial and enterprise infrastructure.

IIoT edge systems need to integrate with both industrial technologies and enterprise IT environments, Dykas said. In addition, they need to manage real time data processing and also provide facilities for creating and executing business rules and data transformation — which in certain contexts would require high processing capabilities.

The scope of edge computing does not stop with local infrastructure, according to Dykas. Not only does it encompass physical and logical networking infrastructure, from Level 0 and 1, up through SCADA, execution systems such as MES and PMS, and enterprise computing such as ERP, but it should also incorporate and interface with cloud based functionality and data.

Dykas spoke of the potential of the Factory IIoT Common Architecture, and the company’s deviceWISE approach in integrating across OT, edge and platform, and enterprise IT environments.

Losant co-founder Brandon Cannaday spoke of hierarchical topologies as an approach to building industrial edge computing environments.

Such a hierarchy would comprise embedded and peripheral devices, including PLCs, Modbus and serial protocol devices, at the bottom of the edge infrastructure, rising through industrial gateway systems and fog computing, that provide data sampling averaging, and machine learning inference functions, to network edge MEC gateways at the apex of the edge infrastructure, providing cloud forwarding, buffering, and notifications.

Randall Kerr and Brandon Moser of Digi spoke of the connectivity options available for industrial edge infrastructure. They made the case for cellular connectivity over alternatives such as wire, satellite and private radio. Cellular, they argued is widely available, inexpensive, reliable and fast, in contrast to alternatives which were often either frequently unavailable, expensive, or slow and expensive.