Streaming analytics on edge with Kuiper and KubeEdge
summary
Kuiper runs as a tiny edge stream processor (about 8 MB install with 10 MB memory overhead) that executes SQL-like ETL, joins, tumbling/hopping/sliding/session windows, ~80 built-in functions and binary image ops. It supports native plugins or RPC/REST calls for ML inference, geohash utilities, and sinks to MQTT, HTTP or TSDB. Integrated with KubeEdge it lets cloud-side rule deployment and edge orchestration, and reports benchmarks like 12,000 tps on Raspberry Pi 3 B+ with ~70% CPU and ~20 MB RAM.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 0:23About me
- 1:13What is streaming analytics
- 6:20Kuiper overview
- 9:39Steps to use Kuiper
- 11:18Kuiper SQL analytics
- 13:40KubeEdge advanced analytics
- 16:10Extension and plugin
- 18:26Other scenarios
- 20:14KubeEdge integration
- 22:01Customer case
- 22:45Next steps
related talks
34mStream Processing at the Edge with Apache Kafka
39mAdvanced Stream Processing on the Edge
25mAccelerating Data Processing and Consumption for Building Modern Apps at the Edge
40mStream Processing and New Approaches: Edge Processing
29mRealtime Analytics Unleashed: Unlocking the Potential of Data at Scale
07mData Gravity and Kubernetes: Managing Large-Scale Data Ingest With Minimal Latency
35mReal-time analytics at IoT scale
38mStreaming analytics and CEP ... #bbuzz