Bringing Cloud-Native Applications to the Edge of the Network
summary
Shows a prototype edge management platform called OptiCon that uses Kubernetes clusters per site and a centralized manager to track clusters and push apps. Uses cluster registry plus a Helm client and a custom CoreDNS plugin to do anycast DNS and route devices to the closest instance. Demo deploys Prometheus to three Copenhagen sites and an nginx instance, with one client resolving locally and another redirected to the neighboring cluster. Targets autonomy at scale (hundreds to thousands of microdatacenters).
AI-written summary · about
chapters (18)
- 0:00Start
- 0:44On Deck
- 1:12The Bandwidth Problem
- 1:46The Latency Problem
- 2:56What is Edge Computing?
- 3:26Edge Computing...
- 4:53Benefits of Edge Computing
- 6:09Edge Computing Use Cases
- 6:51Edge Use Case: Enterprise ML
- 7:55Edge Use Case: Healthcare
- 9:51Edge-Ready Applications
- 12:17Workflow: Edge Kubernetes
- 13:58Edge Platform Architecture
- 17:16Optikon = Edge App Management
- 18:28Optikon Demo - Setup
- 24:39Optikon - What's Next?
- 25:12Key Takeaways
- 27:03Questions?
related talks
09mKeynote: Build an Open Source Distributed Cloud Native World
27mBuilding an Edge Computing Platform for Network Services Using Cloud Native Technology [I]
24mDemo of Kubernetes at the Edge for AI Model Deployments
34mManaging Edge Nodes With “KubeEdge”
39mIoT Application Running on KubeEdge + ARM Platform
39mIoT Application Running on KubeEdge + ARM Platform
18mA Cloud-Native Tool for the mass
32mEdge-Native: The New Paradigm For Operating And Developing Edge Apps