AWS IoT Greengrass
Run cloud processing at the edge
A runtime on edge devices (gateways, industrial machines) to run Lambda, containers, and ML models locally, do local messaging, and sync with IoT Core when connected; components are deployed from the cloud.
- Cost$ Low
- Difficulty●●●●●
- Costs money when idle?A little
- Managed by AWS
Think of it asLike a branch office that handles day-to-day work on its own and only sends reports to headquarters when needed.
When to use it
✓ Use it when
You need processing close to devices because of low latency, unreliable networks, or to reduce data sent to the cloud.
✗ Avoid it when
Devices always have good connectivity and only send raw data → IoT Core; small microcontrollers → FreeRTOS.
Common uses
- Filter and aggregate sensor data before sending
- Run ML inference in factories
- Keep operating when the internet connection drops
What does it cost?
- Pricing model: Billed by the number of active core devices each month; other AWS services called are billed separately.
