What is ‘Edge Computing’?


Edge computing

Edge computing is a computer model in which computer tasks and operations are largely performed at the “edge” of a network on distributed ‘nodes’. A node is a device located on the network which performs actions or tasks requested of it, most commonly a server or similar device running a resource program or software package.

The ‘edge’ is the name given to the distribution of ‘nodes’ on a network. While the nodes may not necessarily be located on the literal edge of the network, the main point is that they are not centralised, as would be common within most networks before the Internet of Things (IoT). The reasoning behind locating these devices on the ‘edge’ is to provide server based resources, data analysis and artificial intelligence in closer proximity to data collection sources, and a cyber-physical system (the main example of a cyber-physical system being Industry 4.0).

Edge computing is derived from the concepts of wireless sensor networks, with both the IoT and edge computing commonly being referred to together. Edge computing is primarily concerned with computation of tasks being performed at the edge of networks, while IoT mainly focuses on data collection. Both edge computing and the IoT are integral components of the 4th industrial revolution, which is predicted to improve product design and industrial feedback by providing manufacturers with more insight into usage information. Hence, ultimately aiding in predictive analytics, resulting in product updates and fixes based directly off customer feedback. ‘Edge’, and the related Fog Computing have also been proposed as more environmentally friendly alternatives to cloud computing, by helping to reduce network electricity and machine cooling costs.

Edge computing works by pushing applications, data and computation power (known as services) away from the central area of a network. Edge computing takes advantage of micro-services, which allow portions of applications to be moved to the edge of a network for operation making them run faster. While content delivery networks move fragments of information across networks, which may be spread across a wide area, edge computing moves applications to the ‘edge’ in order to process applications.

What is an Edge Controller?

IoT is a revolutionary advancement in the modernisation of the generic factory, by incorporating networking and internet based services with factory machinery and devices. It allows the connection of a broad range of devices and machines to occur over a standard protocol, most notably Ethernet. An edge controller is a device connected to the network of devices and machines which allows them to continue to operate even in the event their connection to a server or servers and the network is lost.

For example: You are monitoring the operating temperatures of a piece of machinery.
When the machine is running it is constantly monitoring its own temperature, and when this temperature is within a certain boundary (often referred to as the ‘Operating Temperature’ which for this example is 50-80ºC), the machine knows it is safe to keep operating. However, when the machine senses that the temperature has exceeded this boundary, say at 85ºC, it knows that a series of events need to take place following this. So, the machine senses that it is working too hard and needs to cool down immediately to preserve parts, or the entire machine.

It can trigger an event which temporarily slows down or stops the machine allowing it to cool sufficiently. The logical series of events which take place are saved locally to the machine, this means that regardless of internet or network connectivity, it always has access to these safeguarding measures.

This is the purpose of the edge controller: it stores these events and triggers them accordingly.