Cloud-grilled delights a high-tech approach to perfect BBQ

A cloud solution to perfect your BBQ skills.

1. Cloud-grilled delights a high-tech approach to perfect BBQ JIMMY DAHLQVIST | Believe In Serverless 2024-05-16

2. JIMMY DAHLQVIST Head of AWS @ Sigma Technology Cloud Founder of serverless- handbook.com AWS Community Builder | AWS Ambassador | User Group Leader § Hello, I'm

3. Agenda Background / Problem Architecture overview Cloud deep dive Summary

4. Background / Problem

5. Problem Multi-probe thermometer Wifi Connectivity High / Low thresholds Temperature trends / Stall History

6. BBQ Low and slow Low even temperature for a long time It’s an art form

7. Smoker types Ceramic – Kamado UDS – Ugly drum smoker Electric pellets Offset

8. IoT enabled smoker

9. IoT device – v0.1

10. Architecture overview

11. Components + AWS Cloud + IoT Device

12. IoT Device Software AWS IoT Greengrass 2.0 - Core Custom component Hardware Raspberry Pi 4 model B 2.5mm food probes MCP3008 – AD converter

13. IoT device software First iteration Python application Updated over SSH Connected directly to AWS IoT Core

14. Components + AWS Cloud + IoT Device

15. Cloud architecture

16. Lesson learned IoT Rules as router Small objects in S3 Hard to change data format Extending was a challenge

17. Improvements Cloud architecture Data transformation Event driven architecture Decouple services IoT Device Software development lifecycle Log handling Configuration

18. IoT Device

19. AWS Greengrass Interact with AWS services AWS provided components Log manager Custom components Publish new versions of components Support for AWS Lambda

20. Cloud architecture

21. Cloud architecture Data service Detection service Notification service Data augmentation service

22. Cloud deep dive

23. Cloud architecture Reliably capture data Managed services No direct EventBridge integration Data service Detection service Notification service Data augmentation service IoT Core and SQS

24. Reliably capture data Managed services No direct EventBridge integration IoT Core and SQS

25. Cloud architecture Data service Detection service Notification service Data augmentation service

26. Data augmentation Data transform pattern Fetch data from DynamoDB

27. Cloud architecture Data service Detection service Notification service Data augmentation service

28. Detection service Threshold breaches Temperature trends The dreadful stall

29. Cloud architecture Data service Detection service Notification service Data augmentation service

30. Serverless and event-driven Loosely coupled Scale and fail independently Cost effective Extensibility Highly available

31. Great BBQ?

32. I would say so

33. Takeaways and thoughts

34. Choreography + Orchestration Data storage service Detection service Notification service Data transform service

35. AWS StepFunctions workflow types Standard workflow Long running - 1 year 2000 starts per second Pay per state transition Express workflow Short running - 5 minutes 100k starts per second Pay for duration

36. Express workflows Express Workflows employ an at-least-once model for asynchronous invocation

37. Create subscriptions Creates coupling on the filter Complex to add more targets Changing the filter affect all Coupling is on the event Easy to add more targets Filters can change independently

38. Default bus Avoid using the default bus for custom applications!

39. Transform don’t transport Use AWS Lambda to transform data, not transport. AVOID

40. Future and summary

41. Next step Remove my own hardware ( Inkbird IBT-6XS) Make it SaaS!! Blog posts and Open Source coming! Add camera support Train an ML model for detection Alexa integration