Hardware for IIoT Mist Computing… at the Edge

The latest industrial Intel® Atom™ processors are empowering new, small form factor systems—as IIoT hardware optimization using a bottom-up approach gathers momentum.

Much has been written recently about the difficulties with current IIoT implementations, as IIoT continues to expand into the far reaches of the industrial and commercial environment. In many of these environments (think smart grid, wind farms, oil and gas, autonomous vehicles, etc.) reliable connectivity to the cloud is plagued by intermittent connectivity, latency and security issues. Add to that the fragmented reality of trying to build a cohesive IIoT cloud solution from the vast array of legacy and modern equipment, machinery, control software, and disparate databases, and the task begins to take on monumental costs and time proportions.

To address some of these issues, recent attention has turned to pushing IIoT hardware, data storage, data analytics and communication resources nearer to the IIoT edge in close proximity to the “things” being controlled. First and foremost, this helps address intermittent connectivity and latency issues resulting in better uptimes and overall efficiency, but it also provides more optimal distribution of resources and helps limit the scope of the security task.

Social media conversation and many recent articles have centered on these new IoT/IIoT computing strategies. Extending the analogy of the IoT/IIoT cloud in the meteorological sense, this idea of moving IIoT resources closer to the “things” being controlled is often referred to as fog or mist computing. If fog computing defines IIoT resources in close proximity to things, Mist computing defines IIoT resources directly on or in things.

Table 1: Hardware requirements according to Industrial IoT (IIoT) layers.

Table 1: Hardware requirements according to Industrial IoT (IIoT) layers.

Fog Computing

Promoted by the Open Fog Consortium with founding members including Intel, ARM, Cisco, and Dell, fog computing is defined as, “…a system-level horizontal architecture that distributes resources and services of computing, storage, control and networking anywhere along the continuum from Cloud to Things.”

Fog computing addresses the needs of IIoT at a local level providing distributed data and control resources for increased efficiency and reliability. Fog computing makes use of new software-designed automation elements like software-PLC controllers and digitization of equipment and processes with sufficient detail as to be termed, “digital twins.” These virtual and digitization strategies are a key component in addressing the fragmented state of communication and control at the lowest hardware levels.


Figure 1: Microsoft Azure certified ADLE3800SEC from ADL Embedded Solutions, Inc.

Mist Computing

Extending this analogy one step further, a recent article in the Spring issue of this publication proposes the term “mist computing” in reference to those compute, communication, and storage elements integrated directly into or onto machinery and equipment thus extending IIoT computing to the hardware level. According to the above article by industry expert Angelo Corsaro, Ph.D. one of the primary objectives for mist computing is “…enabling resource harvesting by exploiting the computation, storage, and communication capabilities available on the things.”

IIoT Hardware Layers

Table 1 lists the typical hardware necessary at the various IIoT computing layers. At the “cloud” level, the hardware elements revolve around server farms, immense in some cases, and sophisticated enterprise-scale control centers designed to store and analyze truly massive amounts of data for management, control, and monitoring of the enterprise down to the factory floor.

At the fog computing level, the scale of the equipment takes on smaller proportions via server rooms and local storage supported by an array of smaller networking elements including gateways, routers, and industrial PCs with local databases enabling local data analytics, monitoring and control of things.

Mist computing completes the resource migration picture by extending key hardware elements of fog computing directly onto or into “things” … albeit in much smaller embedded form factors. Beyond providing the equipment control and monitoring function, this hardware must also support fog and mist computing sharing of resources.

Hardware for Mist Computing

The reality of close proximity or direct physical integration onto/into things is no small feat. From an environmental standpoint, the hardware must be able to survive the same environmental conditions (temperature, humidity, mechanical stress, etc.) as the things into which it is integrated. Increasingly, these things are in exposed or remote locations making the choice of mist computing hardware a critical design element.

As well, the product lifetime and quality of mist computing hardware cannot degrade in any way the overall quality and product lifetime of the machinery or equipment which it is controlling. From a vendor standpoint, this translates into a careful selection of hardware BOM components that optimizes product lifetime and quality. As well, the circuit architecture must be such that operation over all temperatures and voltage conditions is guaranteed…all while maintaining a compact form factor suitable for embedded integration.

Figure 2: Microsoft Azure certified, ADLEPC-1500 from ADL Embedded Solutions, Inc.

Figure 2: Microsoft Azure certified, ADLEPC-1500 from ADL Embedded Solutions, Inc.

From a functional standpoint, fog and mist computing hardware must support multiple cores with virtualization technology to support software-defined automation and digitization requirements. This hardware must also provide the necessary performance for on-machine data analytics, control, monitoring and communication with other elements of the mist or fog computing network.

Addressing many of these needs are new, small form factor embedded CPUs and system offerings from companies like ADL Embedded Solutions, Inc. and others, which are bringing full-featured compact CPUs and industrial embedded PC designs to market.

A good illustration of this is ADL’s new ADLE3800SEC designed with the latest industrial Intel® Atom™ processors with extended junction temperatures of -40 ºC to 110 ºC. This compact (75mm x 75mm) edge-connected SBC is capable of standalone operation or easy integration with expansion I/O boards, which helps provide a single computing board across equipment and product profiles for consistency of hardware, firmware and BIOS.

Compact solutions like the ADLE3800SEC or the derivative mini ADLEPC-1500 (Figure 2) also ease the task of IIoT deployment by maintaining compatibility with IoT development platforms like Microsoft Azure and others to help optimize security and overall stability. Their substantial functionality and performance at generally lower cost helps reduce the cost of fog and mist fabric creation—necessary for efficient distribution of data storage and data analytics for fog and mist communication, monitoring, and control.

As this bottom-up approach to IIoT hardware optimization continues to gain momentum, and the ecosystem of hardware vendors providing fog and mist computing solutions grows, we can expect to take a giant step forward in the near future toward making the promise of IIoT a reality.

JCJC Ramirez is Director of Engineering at ADL Embedded Solutions, Inc. and is the current vice-president of the PC/104 Consortium. Ramirez, BSEE, MBA, has a technical background that includes Navy nuclear plant supervision, nuclear instrumentation, semiconductor product development and embedded systems engineering.

Share and Enjoy:
  • Digg
  • Sphinn
  • del.icio.us
  • Facebook
  • Mixx
  • Google
  • TwitThis