Fog in the Factory

The fog is starting to roll in on the factory floor.

Manufacturers who have embraced the Industrial Internet of Things (IIoT) have made gains in performance and operational efficiencies, thanks to connected sensors that measure and manage equipment performance throughout the factory. Tiny sensors stream performance data to the cloud, helping to troubleshoot before equipment fails.

In these digital factories, IIoT is exposing the performance gaps of traditional cloud-only systems. IIoT requires a connected factory that is capable of providing reliable, real-time accessibility to the data being captured and analyzed by automated systems and processes. The amount of data from industrial manufacturing sites that operate drones, industrial robots and industrial control systems is now measured in petabytes, dwarfing all previous networking demands.

Given the requirements for real-time communication flows throughout – and beyond – the factory floor, it’s becoming clear that cloud-only approaches can no longer keep up with the necessary volume, latency, mobility, reliability, security, privacy and network bandwidth challenges. Streaming the massive amounts of data to the cloud and back, under the assumption of 100% availability of broadband service with less than millisecond latency, is not possible in today’s cloud-only computing models.

Ushering in the fog

Fog computing is the horizontal architecture that distributes compute, communication, control and storage closer to where the data is originated, enabling dramatically faster processing time and lowering network costs. It augments, not replaces, investments in cloud to enable an efficient, cost-effective, and constructive use of IIoT in manufacturing environments. Fog extends the traditional cloud-based computing model where implementations of the architecture can reside in multiple layers of a network’s topology. Fog scales both horizontally (fog node to fog node) and vertically (between fog layers), and can dynamically respond to network load changes and failures more effectively than a simple extension of the cloud.

While many use the term ‘fog’ and ‘edge’ interchangeably, there are key differences. Fog computing always uses edge computing, but edge does not always require fog capabilities. Fog is inclusive of cloud, whereas edge is defined by the exclusion of cloud. Fog pools the resources and data sources between devices that reside at the edge in north-south, east-west hierarchies, where edge tends to be limited to a small number of layers. Any device with computing, storage, and network connectivity can be a fog node. Examples include industrial controllers, switches, routers, embedded servers, and video surveillance cameras.

At the OpenFog Consortium, we term the advantages of fog computing as SCALE: Security; Cognition; Agility; Latency; and Efficiency:

Security –Most cloud-based security services focus on providing perimeter-based protection for threat detection to industrial control systems. New control requests are redirected to the cloud for authentication and authorization processing. Security can be further compromised by the common practice of installing hardware and software updates during a system’s scheduled down time, rather than any time a security compromise is detected. If threats breach the protections, the common response has been to manually take the system offline to be cleaned up. These processes are inadequate for today’s connected factories. Fog addresses these IIoT security challenges through its distributed architecture. Factories can protect the fog nodes using the same corporate IT policy, controls, and procedures. The fog nodes can quickly identify unusual activity and can mitigate threats or attacks before they are passed through to the system without disruption of service.

Cognition – The fog architecture can best determine where to carry out the computing, storage and control functions along the cloud-to-thing continuum. Decisions can be made on the device or via a nearby fog node, avoiding the need to transport the data solely to the cloud for decision making. Smart sensors can make autonomous decisions and trade-offs regarding manufacturing execution. Multiple connected machines can communicate within production environments and learn from their decisions, improving performance over time.

Agility – Sensors and systems generate data that is turbulent, bursty, and often created in huge volumes, and must be rapidly interpreted and turned into actionable insights and decisions. The predicted scale of the data generated by automated systems means that it is highly unlikely that humans alone can make the astute operational decisions for the benefit of the business. In addition, the architecture and connectivity must be flexible. Fog works over wireline/optical and wireless networks and also inside these networks. This makes it ideally suited for industrial elements based on wired SCADA systems, OPC-UA interfaces, Modbus, and so on, which can be connected to fog nodes.

Latency – Latency is the most cited benefit of a fog-based approach. Many industrial control systems require end-to-end latencies within a few milliseconds, falling outside the range for mainstream cloud services. While a one-second delay is simply annoying in gaming or in talking with Siri, it is an eternity to an idle piece of manufacturing. Milliseconds matter when you are trying to prevent manufacturing line shutdowns, avert accidents or restore electrical service. Analyzing IoT data close to where it is collected minimizes latency, supporting the time-critical processes of the connected factory in averting disaster or shutdowns

Efficiency – Fog pools resources to allow applications to leverage idling computing, storage and networking resources to enhance overall system efficiency and performance. The fog architecture takes an ‘immersive distribution’ approach. If a network goes down, it immediately switches to other fog resources which are available throughout the network. This enables flexibility of management and ease of integration with existing IoT environments. 

Fog computing relies on rapid, trusted and secure transmissions between devices and systems – in other words, it needs interoperability based on open standards. Developing this open reference architecture framework is the work of the OpenFog Consortium. Eight pillars, as viewed in the above figure, represent the foundational elements to the OpenFog reference architecture. Members of the Consortium from universities, startups and industry giants are collaborating on it to ensure an open, interoperable architecture that will eventually enable manufacturers to choose fog-based solutions for their smart factories from a diverse, vibrant supplier ecosystem.

For manufacturers, letting the fog into the factory is not just a promising approach – it’s a necessary one to enable the full potential of the Industrial Internet of Things.


This blog was written by Lynne Canavan, Executive Director, OpenFog Consortium.

About your Guest Blogger:

Lynne Canavan is the Executive Director of the OpenFog Consortium, the global organization formed to accelerate the adoption of fog computing in order to solve the bandwidth, latency and communications challenges associated with IoT. OpenFog was founded by ARM, Cisco, Dell, Microsoft, Intel and Princeton University Edge Computing Laboratory in November 2015, and has over 50 members today. Lynne can be reached at, on LinkedIn at or on Twitter at @lfcanavan