Many of the more than 26 million software developers in the world (and growing), are focused on making small, semi-autonomous devices (upon which our modern lives depend) work better. These can range from controllers that sense when industrial paint sprayers are “missing a bit,” to connected lawnmowers that know just how you like your lawn cut.
What all these IoT connected devices, or things, have in common is that they can communicate with online platforms, and the data that they produce is valuable in some way.
In many industries such as IIoT in manufacturing, transportation and logistics , these devices are located far from where the data is processed. Soil analysis devices, that determine and control when fertilizer should be applied to crops, can be miles away from the farm’s headquarters. Your lawnmower could be thousands of miles from the manufacturer’s nearest data center. So, any data that comes from these remote devices is typically delayed and patchy.
The solution? Edge computing. It enables real-time usage data processing, so data can be analyzed closer to the source, reducing the physical distance it must travel and thereby reducing latency. According to IoT testing and benchmarking analystMachNation, edge platforms are beginning to play a “vital role in the IoT ecosystem.”
MachNation research suggests that most IoT solutions in verticals such as manufacturing, commercial transportation, building automation, and logistics “will be required to leverage edge computing in order to meet business and operational requirements,” according to the MachNation 2021 IoT Edge Scorecard report.
All end-users of relatively small (processing-power wise) semi-autonomous devices – like medical technicians using centrifuges, farmers tending to their crops, or facilities managers ensuring optimized pump operation – need edge computing. They need to be able to extract even more value from their products through increased uptime, additional operational intelligence and improved user experience.
But there are many flavors of edge along the continuum, with all coexisting to provide intelligence on individual remote assets or groups of co-located assets as needed.
Thin edge or thick?
So, how do you know what to use to do the job? Developers are currently debating this. Does your device need thick edge, with highly analytical processing power? Or thin edge with lighter-weight computing?
What is the difference?
Typically, classical thick edge devices are single industrial PC versions of the same IoT technology used in the cloud. They oversee the operation of a range of assets in a single site, in a fully autonomous manner.
Think of the boxes on building control systems, water processing plants, or the computers within self-driving cars and trucks. Or they can be high performance ruggedized tablets for tough environments, or industrial PCs on the factory floor.
Thick edge devices can perform extremely intensive analysis for an entire environment of connected assets. Take, for example the oil and gas industry; computers within the facility collate data from hundreds of sensors to monitor and analyze oil drilling, refinery or storage facility conditions, so that they can get automatic alerts to any maintenance and operational issues.
Thick edge devices are perfect for situations that desire fully autonomous, high-performance analysis, and operational visualization and where size, power and the addition of a new physical component is not an issue.
But there are many situations in which size, cost and power are major constraints.
Thin edge deployments are typically much smaller than thick edge in terms of physical size, product cost, power requirements and installation complexity. They operate with minimal resources in small lightweight compute devices – like those in modern network routers, set-top-boxes or display units.
Thin edge devices can perform less-intensive analysis on a single or small number of connected assets but will do it in a semi-autonomous manner. If there is a problem, the device would “call for help,” alerting online platforms that there are vibrations or humidity measurements, for example, that are outside locally manageable thresholds.
Recently the push for more intelligence to be embedded into smaller devices has been driven by AI-powered video cameras, which continuously share derived metrics, like car number plates, but only stream the video on request.
However, historically enabling smaller devices to become thin edge devices has been a challenge.
While an increasing number of IoT use cases demand a greater degree of local edge processing, thin edge developers are still grappling with the challenges of secure connectivity and app management. This is where the open source thin-edge.io software framework comes into play. It offers simple, secure and reliable cloud connectivity to a range of cloud platforms.
thin-edge.io software capabilities are small enough that they can be deployed “within” modern industrial products, protocol gateways and control boards with minimal impact on current performance – essentially embedded within the current infrastructure.
Software AG is a leading contributor of the open source thin-edge.io project. It is standards-based, container compatible and supported by a global developer community. thin-edge.io solves the IoT platform connectivity and app management issues inherent in the development of all lightweight IoT devices, and it does it in a way that maintains the developers’ freedom of choice of IoT platform, app programming language and hardware.
thin-edge.io is being assessed by customers and partners for a range of applications in industrial solutions, providing embedded system developers with the tools to make their increasingly sophisticated solutions faster, more secure and more robust with zero hardware, language or cloud lock-in.
Click here to learn more about thin-edge.io.
And learn why Software AG is rated as a leading vendor in MachNation’s 2021 IoT Edge Scorecard.