Server/cluster
As noted throughout the book, not all edge devices are created equal, especially when it comes to compute and storage. When IoT-type devices such as sensors, gauges, and actuators that do not have any compute or storage are part of an edge computing solution, an edge server or edge cluster is inserted and takes on the role of aggregating the data. In some cases, it is not viable to run a large ML application on the edge device, in which case that application would have to be deployed on the nearest edge server/cluster.
AD # |
Requirement |
Factors affecting decision |
S001 |
Deployment of edge servers/clusters |
The use case will determine whether an edge server/cluster is required, and if so, based on the application to be hosted and the data generated, the server specification will be determined. ... |