For IoT/Edge, use patterns have emerged around data ingress, processing, event detection, data egress. The intent here is to share example(s) of those patterns
We usually use one or Streams for ingress. This makes it easier to connect the data to other data structures via a Gear. See Anti Patterns for additional scenarios.
RedisGears registers for events (new data) on Streams and then takes appropriate action. This could be something entirely covered by the Gear or it could be a handoff to another data structure.
Techncially you can do anything in a Gear that you can do in Python. But do you really want to write all those conditions. Something like JSONLogic makes an excellent engine for this kind of business logic.
- If all the data is time series, RedisTimeSeries effectively acts as a data/sample stream. Of course a Gear can't react to it directly though data traditional Redis events can be used as triggers or the app can invoke a Gear. Typically the app is interacting directly with one or more time series.