Data becomes an issue
Online data storage is cheap and cloud computing power is widely available. This creates the situation where IoT solutions send all their data to a cloud platform before analysis starts. Why not have all this great data available? A problem rises where data streams become too big. This can be a costly adventure (paying per byte will get expensive). It is also causing high loads on the cloud end of the IoT solution. Availability of services floats towards the danger zone. In comes edge computing!
With edge computing, the local computational power is used to (pre-) process data. Local analysis is done before communicating this to the cloud. Even a shift from cloud applications to local applications is possible. Edge computing is literally computing data at the edge of the network.
Always edge computing?
The shift can be made towards the tiny end of the IoT solution (for example a sensor with local storage) but also a combination of things with a gateway does an analysis of data before sending it to the cloud platform. Take 1000 sensors in a factory transmitting messages 50 times per second and there are 50,000 messages per second. Only for one type of sensor! Edge computing can create a situation where all the data samples are averaged locally and then sent. It reduces the data stream by factors. Edge computing solutions are potentially advantageous but also have their downsides.
This webisode of IoT Friday gives insight into edge computing, pros and cons, and insight in the position within IoT solutions. To watch all the previous webisodes, go to our Youtube Channel and don’t forget to subscribe.