Platform for the food and beverage industry
Capturing data in the machine and in production. It is the essence of Industry 4.0 to allow machines to communicate flawlessly with each other and generate more insight from which optimizations are distilled. Much of that data disappears to the cloud where sufficient computing power is available. But what about data that can actually be processed just as well or even better locally, for example because of their time-critical nature. They are sent to the edge of the network where a smart controller or PLC can do its thing with them. Edge computing, in other words.
Due to the increased need for flexibility in a manufacturing process, machines today are being designed to account for variations in the process and production environment. They must then weigh the impact of all the captured data in a split second to make the right decision. Although the cloud has enough computing power to make these tradeoffs, the way there and back takes time. Not minutes, but certain processes now require ultra-short response times. Perhaps the most famous example of this is the self-driving car. The faster it can react to its environment, the safer its driving will be. But there are also applications in an industrial environment, especially when artificial intelligence enters the playing field.
The alternative then is to make sure there is enough computing power on site. Machine and production line controllers are usually busy enough with their tasks to keep everything running seamlessly. Therefore, people are looking for a picture in the periphery, at the edge of the network: edge computing. This often involves IoT devices deployed to process certain data. Edge computing is useful for all applications where it can be interesting to minimize the distance between the location where the data originates and where the information that can be distilled from it must be deployed again. For example, to perfectly synchronize robots in a factory so that waiting times never occur. The main advantage of installing a separate controller or PLC at the edge of the network to control certain processes is first and foremost the time savings. The shorter the path, the faster the response can be. In addition, this also keeps the (expensive) bandwidth within the rest of the company out of the way.
Everything out of the cloud and to the edge then? Not at all. The two technologies are just particularly complementary. The results that follow from the computations of edge devices can in turn be sent to the cloud for analysis or trending or storage, for example. On one condition: the edge IoT devices must then be compatible with cloud services and operating systems. However, most industrial applications already in existence today prove that the cloud and edge are perfectly compatible. The fact that there is also particularly much hailing on industrial benches for the arrival of 5G has everything to do with this. After all, with such a network one can start connecting 1 million devices on 1 m². The low response time (latency) will also boost the speed at which data, even large quantities, can be exchanged. Real-time will then become truly real-time. And all for only a fraction of energy.
Another term often mentioned in the same breath with edge computing is fog computing. The difference is purely in the location where the data is processed. Edge computing seeks a place at the edge of the network, whereas in fog computing the computations happen in the local network and the data is then sent through a gateway. This is a decentralized computing infrastructure, where there will be a dedicated device between edge devices and the cloud. Data, computation, storage and applications are distributed in the network at the most logical and efficient place between the data source and the cloud.
