Updating the Edge
summary
Devices number 20.4 billion and updating them is fragile so the demo builds a CI/CD pipeline for embedded Linux that actually works: a Raspberry Pi hooked to an RC car accepts live deployments from JFrog Artifactory and Pipelines with X-ray scanning, K3S and Helm push app updates in 35 seconds deploy without interrupting use, while Yakto builds minimal images and Mender performs 5–10 minute firmware update using a dual A/B rollback strategy to auto-recover failed updates and caching cuts build time by about 50%.
AI-written summary · about
chapters (20)
- 0:00Introduction to edge updates
- 1:28Scale of edge devices today
- 2:51Why updates matter now
- 4:17Barriers to fixing updates
- 6:29One size fits lots approach
- 7:10Proof of concept demo setup
- 8:33Challenges in car updates
- 10:45Software update workflow
- 11:26Tools for software updates
- 13:56K3S for edge Kubernetes
- 15:22Helm for deployment
- 16:47Silent updates without disruption
- 17:10Firmware update workflow
- 17:52Mender for OTA updates
- 19:18Dual A/B partition strategy
- 19:38Yakto for custom Linux builds
- 21:04Build cache with Artifactory
- 21:45Firmware update speed results
- 22:29Recap and call to action
- 23:14Q&A and resources
related talks
31mUpdating the Edge
22mUpdate Strategies for the Edge
45mSoftware Updates for Connected Devices: Key Considerations
50mThe ultimate guide to software updates on embedded Linux devices
45mSoftware Update for IoT: The Current State of Play
40mContainers on the Edge: Deploying Embedded Linux Systems With Modern D...
36mContinuous Deployment of IoT Sensors
17mAccelerating Development and Deployment of Updatable Linux-based Edge AI devices in the CRA era