From Sensors to Servers: Efficient Edge Computing with Akri and WebAssembly
Talkat Kubernetes on Edge Day North America 2024 · on Tuesday, 2024-11-12· at 16:30· Salt Palace | Level 2 | 250 A·
summary
Akri discovers IoT devices and registers them as Kubernetes extended resources, supporting MQTT, OPC UA and udev and injecting broker credentials into Pods; SpinKube runs serverless WebAssembly as Pods by loading a containerd shim (runwasi) and a Spin operator. Wasm artifacts are tiny and start in sub‑millisecond times, packaged as OCI images. Demo: four Arduino sound sensors publish 0–1600 to an MQTT broker on AKS, Akri creates instances, a Spin Wasm app persists values to SQLite and the front end shows live time series — it deploys and works on a two‑node AKS cluster. Akri discovers devices as extended Kubernetes resources SpinKube runs serverless WebAssembly via a containerd shim
AI-written summary · about
chapters (14)
- 0:00Edge Computing challenges
- 1:06Akri and Spin for Edge workloads
- 2:10Traditional containers pain points
- 5:43WebAssembly benefits
- 7:48Spin: Serverless WebAssembly apps
- 9:13SpinKube: WebAssembly on Kubernetes
- 10:37Live demo setup overview
- 12:21Local Spin and MQTT dev
- 15:52WebAssembly as OCI artifacts
- 16:35Deploying with SpinKube
- 17:17Discovering devices with Akri
- 19:01Akri with Spin integration
- 21:07Live demo results synergy
- 22:10Q&A: WebAssembly vs containers
related talks
29mLiving on the Edge: Using IoT Devices on Kubernetes WebAssembly Applications
23mAkri: Making IoT Devices Accessible to Your Edge Kubernetes Clusters
21mConnecting to devices at the edge with minimal footprint using AKS Edge Essentials and Akri on WASMs
25mOrchestrating Machine Learning on Edge Devices Using Akri and WebAssembly
24mWeb Assembly, Docker, Containerd, and Kubernetes
39mRunning WebAssembly on Kubernetes
31mBuilding Event-Driven WebAssembly on Kubernetes: Runtime, Observability, and…
04mAkri: Using non-Kubernetes devices in Kubernetes