Lessons from the Field: Running Kubernetes at the Wellsite
summary
Case study of running Kubernetes on ~1,000 oilfield trucks that process a terabyte per day from a legacy Windows probe into a local k3s single-node clusters shows what actually works: an EVE-based edge OS with A/B partitions, embedded hypervisor and measured boot, GPUs on Xeon boxes for on-truck CV, and strategies for flaky networks, physical security and long ten-year lifecycles. Surprising scale problem: thousands of single-node clusters break classic kube control planes, so teams shard Rancher or use Helm-to-edge or alternatives like Open Horizon instead.
AI-written summary · about
chapters (13)
- 0:00Kubernetes at the wellsite
- 1:06Oil well edge use cases
- 3:35Real-time well analysis
- 6:45Introducing EVE
- 7:48Scaling edge Kubernetes
- 8:28Challenge: distribution and features
- 9:52Challenge: unreliable network
- 12:00Challenge: security threats
- 13:26Challenge: hardware diversity
- 15:11Challenge: people costs
- 16:36Challenge: scaling clusters
- 19:07Future of edge Kubernetes
- 20:10Q&A
related talks
06mEdge Kubernetes in Action: From Connected Vessels to Industrial Au...
01h 16mIntro + Deep Dive: Kubernetes IoT Edge WG
36mEdge Computing in Kubernetes: Best Practices for the Future
10mDevelop once, deploy anywhere: flexible infrastructure extends to the edge
39mShould You Bring Kubernetes on Your Edge Roadtrip?
17mKeynote: Extending Kubernetes: Application Orchestration to 1000’s of Clusters in...
32mRunning Kubernetes in a Manufacturing Line – What Could Possibly...
24mThe Cluster Is Not the Platform: Operating Industrial Edge Kubernetes as a Fleet