Sending every raw sensor reading to the cloud can become expensive at scale.
Imagine thousands of devices producing measurements every second.
That's a huge amount of network traffic and storage.
One approach is to process some information locally.
For example, instead of sending:
72.1
72.2
72.1
72.3
72.2
72.4
...
the edge device might calculate:
Average = 72.2
Maximum = 72.4
Minimum = 72.1
and send the summary.
You can also filter out readings that aren't meaningful.
Of course, you shouldn't discard information blindly.
Sometimes raw data is valuable for later analysis.
But intelligent filtering and aggregation can reduce:
Bandwidth
Storage
Cloud processing
This is one reason edge computing can be useful beyond latency.
It can also be an architectural and cost decision.
Top comments (0)