Edge-to-Core Machine Learning and Data Pipeline on k8s
summary
A runnable edge to core pipeline on OpenShift/Kubernetes that deploys LPR and OCR services, local and central Kafka clusters with Kafka MirrorMaker to move events, Ceph object storage for long term retention, and dashboards via Apache Superset and Grafana. Ansible automation brings up ~36 pods and the full stack in about 15 minutes, simulates cameras with a load generator, archives messages to S3, and shows generated results like ~45,000 vehicle events and £215,000 toll/penalty totals, all working end to end.
AI-written summary · about
chapters (6)
- 0:00Smart city ML pipeline intro
- 4:58Solution design and data flow
- 11:55Ansible for cluster setup
- 17:35Live system demo
- 21:48Real-time Kafka analysis
- 24:16Summary
related talks
24mDemo of Kubernetes at the Edge for AI Model Deployments
06mEdge Kubernetes in Action: From Connected Vessels to Industrial Au...
21mMastering the DevOps Machine Learning Pipeline
22mBuild and Monitor Machine Learning Services in Kubernetes
23mMigrating 5G Core to the Edge, Wisely and Confidently with Kubernetes
31mIntro to KubeEdge: Kubernetes Native Edge Computing Framework
33mCompliant Data Management and Machine Learning on Kubernetes
36mEdge Computing in Kubernetes: Best Practices for the Future