Industrial IoT devices are proliferating, as is the data they generate. Edge computing can boost the usefulness of analytics in these environments.
As 2019 commences, it will be virtually impossible to ignore the emerging partnership between the Industrial Internet of Things and edge computing.
The Industrial Internet of Things (IIoT) uses IoT technologies to enhance manufacturing and industrial processes. Increasingly, these processes need connected devices to conduct tasks efficiently.
Edge computing architecture moves compute processing closer to the users and devices that need it rather than having processing occur centrally in an on-premises data center or a public cloud. The edge has become critical for industrial and manufacturing processes that use vast amounts of data, that require rapid reaction times and that need rigorous security attached to them.
The sheer volume of data is more than enough to justify the acceleration, because IoT-generated data is growing exponentially faster than the traditional cloud environments where data is stored. Additionally, cloud as destination introduces data transport issues – latency and bandwidth – making the speed of the journey the central issue. The edge is needed as a solution to the inefficiency of IIoT-to-cloud architecture.
Industrial IoT devices and edge computing have grown at impressive rates. Accenture predicts the IIoT market will reach $500 billion by 2020; and IIoT already generates 400 zetabytes a year. Gartner estimates that IoT currently generates about 10% of enterprise data; by 2022, Gartner has predicted this will increase to 50%.
According to IDC, IT’s annual investment on edge infrastructure will hit 18% of total IoT spending; and per last year’s Forrester Analytics Global Business Technographics Mobility Survey, 27% of global telecom decision-makers say their companies will either implement or expand edge computing this year.
When Industrial IoT devices and edge computing work together, digital information becomes more powerful., particularly in contexts like smart cities where data may need to be gathered in classic edge contexts, such as at a parking meter or from a connected trash bin on a residential street. It is becoming commonplace for smart city engineers to install Internet-enabled sensors in trash receptacles throughout metropolitan areas. The receptacles can then be monitored remotely through sensors; when full, a city sanitation department receives a notification and an order can be logged for a receptacle to be emptied.
The sanitation system has become more efficient. But, more important, as data is gathered more closely to these devices, systems can learn over time. In this example, the system comes to realize which receptacles become filled to capacity on which days, and it can plan to empty those bins and optimize collection routes accordingly.
By putting processing resources into the IIoT environment to make that processing possible, traditional IT problems in industry are now being addressed.
Manufacturing equipment and software are often woefully mismatched in terms of their lifecycle, creating security vulnerabilities for legacy manufacturing equipment that isn’t patched regularly.
Edge computing places gateways between these old connections and the resources that crunch their data. If a device is accessed by a malicious attacker, the breach ends there. The attacker gets the device’s data, but does not gain access to a network or other devices. Limiting the transfer of data to the edge, rather than making a roundtrip to a centralized data center or the cloud, eliminates an easily exploitable threat.
By bringing artificial intelligence to devices themselves, edge computing can also make decision making more context-driven and rapid at the edge.
Emergency medical technicians, for example, must rapidly determine the vital signs of a person in distress and gather information from those in the area. If a factory worker collapses on the job while wearing a smartwatch that monitors its wearer's physical condition, EMTs have immediate access to vital signs in the moment and prior to the collapse. They would also see information about the environment at the time of the incident: temperature, noise level, the presence of fumes, and any other data that might provide context for rapid and effective treatment.
Moving from cloud-centric processing and eliminating the time of a cloud-bound transaction, doesn’t just mean faster systems – it is the next step in information-driven decision making.
The raw data can always be moved to the cloud after it has done its local work, to contribute to more global AI processes, or to be archived there.
With Industrial IoT devices combined with the edge, high-fidelity analytics and a smaller footprint will become the AI norm in 2019.
Perhaps the greatest product of the partnership between Industrial IoT devices and edge computing will be the boost it provides to emerging AI functionality out in the world.
Industrial AI will be more than just mechanical – it will be “artistic,” including new areas like natural language processing and computer vision – two areas also getting a huge boost from edge computing.
Most of the cameras in the world are already on local networks. Connecting these cameras to clouds is by definition a major data-transport burden. It is far easier to install resources for vision processing where cameras reside rather than retaining centralized connections. The same goes for natural language processing; voice communication happens in the field, not the server room.
During an entertainment event at a civic arena, cameras can identify additional seating and reroute incoming attendees efficiently and comfortably, based on real-time analysis in the edge of the traffic being measured on-camera.
As edge networks get built out in the IIoT, both of these technologies will rapidly evolve, and that accelerated development will have benefits across all of IT.
Seldom in the history of digital technology has there been such perfect timing, when it comes to innovation and resources meeting emergent needs.
Scott Robinson is an enterprise architect and AI consultant with a 25-year history in business intelligence, analytics, and content management in the healthcare and logistics industries. He is currently CIO of the GlenMill Group, a research consortium providing new AI technology and infrastructure for enterprise applications and services.