Modelling Durable IoT Workflows with Cloud-managed Finite State Machines
summary
Claim: cloud‑managed finite state machines orchestrate durable IoT workflows so upgrades survive reboots, intermittent connectivity and scale. Implementation uses AWS IoT Core, IoT Device Management jobs, Step Functions as the orchestrator, Lambda + SNS for events, a Dockerized TR‑369 agent simulating devices. Demo spins up ~100 containers, about 77 successful executions while others wait to connect, showing per‑device state, retries and rollbacks are observable. Design proves externalized orchestration and event‑driven state tracking work at fleet scale.
AI-written summary · about
chapters (12)
- 0:00Welcome
- 3:17IoT workflows as distributed systems
- 8:10IoT workflow characteristics
- 12:20Expectations for IoT systems
- 16:32Design approaches
- 23:33State machines for workflows
- 26:03Firmware upgrade example
- 33:09Implementation architecture
- 37:20Live firmware upgrade demo
- 50:24Scaling to 100 devices
- 54:36Open source alternatives
- 55:59Q&A
related talks
41mBuilding Durable Serverless IoT Device Workflows using Cloud
41mGetting started with Orchestration and Choreography in Distributed Applications
35mFail Fast IoT: accelerated innovation with cloud-powered hardware
28mThe Art of the State: Coordinating Services Using AWS Step Functions
46mBuild, Scale, & Create IoT and Mobile connected applications with the Cloud
28mTransform Tomorrow's Industries With AWS IoT
49mDesign patterns for distributed client server systems via IoT
31mDevSecOps with IoT - Is it a myth?