Serverless with Kubeless
summary
An internal extensibility platform called Kima runs serverless workloads by wiring Kubeless on Kubernetes with Istio for mesh and the Open Service Broker bindings so functions become namespaced deployments. Kubeless creates a Deployment and an always-on pod per function (cold-start avoided but higher base cost); suspend support is pending. OpenWhisk and Fission were rejected for multi-namespace and integration reasons. Demo creates a Kubeless “lambda” that becomes a Deployment, supports CloudEvents, and exposes service credentials via broker bindings.
AI-written summary · about
chapters (18)
- 0:00Introduction and expectations
- 1:30About the speaker and SAP Hybris
- 3:57What is serverless architecture?
- 5:45Why build a serverless platform?
- 7:11Introducing Kima platform
- 8:56Kima architecture overview
- 10:43Evaluating FaaS platforms
- 12:58OpenWhisk vs Fission vs Kubeless
- 14:21Why choose Kubeless?
- 16:50Kubeless integration with Istio
- 17:34Open Service Broker API explained
- 20:46Service broker in Kima
- 22:17Live demo introduction
- 23:01Kima environment and service catalog
- 25:31Creating a Kubeless function
- 26:37Function code and CloudEvents
- 27:54Kima wrap-up and benefits
- 29:17Q&A session
related talks
39mServerless on Kubernetes with Kubeless [A]
21mServerless computing using Knative Open Source
01h 00mServerless Madness on Kubernetes
33mTowards Serverless As Commodity: A Case of Knative
21mKeynote: Kubernetes Serverless Present and Serverless Future
23mServerless with Knative
20mKnative: A Kubernetes Framework to manage Serverless Workloads
31mBeyond Traditional DevOps: Dive into Serverless with K8s