Bridge Your IoT Leaf Devices to Local Clusters with Ease Using Akri...
summary
Akri extends Kubernetes device handling to expose networked IoT leaf devices (OPC UA, ONVIF, UDev) as shareable CRD resources, implemented in Rust and released as v0.12.20 while targeting Akri v1.0 roadmap. It integrates the new Dynamic Resource Allocation (DRA) API to support resource classes, resource claims, CDI entries, and capacity-based sharing, and splits control into workload and driver controllers to avoid races. A live demo uses an MDNS discovery handler with a Raspberry Pi printing flow (printer jobs queued, resource claims allocated) that ultimately works intermittently due to flaky Wi‑Fi.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 0:30Agenda
- 0:53IoT Leaf Devices
- 1:47Device PlugIn System
- 2:40Akri
- 3:52Akri Architecture
- 5:08Dynamic Resource Allocation
- 13:09Architecture
- 16:00Roadmap
- 18:47Documentation
- 19:27Community Meetings
- 19:45Wrap Up
related talks
23mAkri: Making IoT Devices Accessible to Your Edge Kubernetes Clusters
42mLeaving no Leaf Device Behind: At the Frontier of Edge Computing
04mAkri: Using non-Kubernetes devices in Kubernetes
09mMaking IoT Simple: Leveraging Akri & Nephio in a Hybrid Cloud En...
24mFrom Sensors to Servers: Efficient Edge Computing with Akri and WebAssembly
35mCloud Native IoT: OTA Updates and Device Repurposing With K8s
27mUncharted Waters: Dynamic Resource Allocation for Networking
39mKubernetes IoT Edge Working Group: Edge Device Onboa...