WebAssembly on the Factory Floor: Efficient and Secure Processing of High Velocity Machine Data
summary
MachineMetrics implements a Rust LTTB downsampler compiled to WebAssembly and deployed on wasmCloud + WM, ingesting kilohertz (1 ms) sensor streams at the edge to avoid sending raw HF data to the cloud where Grafana and browsers choke. High-frequency traces (5,000 points reduced to ~50 with preserved shape) run inside a sandboxed actor, use NATS/JetStream key-value for distributed caching and config, and hot-swap to an in‑memory provider with near‑zero downtime. Remaining work covers secret rotation, specific WIT provider features, and debugging tooling.
AI-written summary · about
chapters (10)
- 0:00Introduction to Machinemetrics
- 1:29IoT Architecture Challenges
- 3:21High-Frequency Tool Monitoring
- 5:22Why WebAssembly and Wasmcloud
- 8:31Data Downsampling Algorithm
- 11:01Proof of Concept Architecture
- 14:24Live Demo and Visualization
- 21:09Overcoming Technical Challenges
- 24:34Future Architecture Plans
- 26:06Q&A Session
related talks
26mWebAssembly for a Reactive Internet of Things
37mAI/ML at the Extreme Edge with WebAssembly: A Path Forward
25mObservability at the Edge: Instrumenting WebAssembly with OpenTelemetry
34mWebAssembly at the IoT Edge
25mFrom Cloud to Edge computing - Unleashing the power of WebAssembly at the edge
32mWebAssembly for Front End Devs
19mWebAssembly's Journey Beyond the Browser
41mCreating High-Performance Web Apps with WebAssembly