Function as a Service in private cloud
summary
Alibaba private cloud implements an internal FaaS platform built in Kotlin to cut developer overhead, keep functions small (ideally ~80 lines) and provide a consistent SDK for HTTP/RPC calls, tracing and metrics. Platform groups functions into multitenant containers to avoid cold starts, supports coroutines so one JVM runs millions of lightweight tasks, and records event payloads at about ~1 KB per event. Production footprint is under <10k functions, observes Spring startup around 1 s (optimizable toward 40 ms with AOT), and enforces per-function specs, versioned routing and end-to-end monitoring.
AI-written summary · about
chapters (5)
- 0:00Introduction
- 10:57Microservice architecture problems
- 25:15FaaS platform in private cloud
- 36:13Summary
- 49:23Q&A
related talks
12mFaaSNet: Scalable and Fast Provisioning of Custom Serverless Container Runtimes...
35mAchieving Faas + Dapr + K8S technology integration, and build a new Alibaba Cloud Se...
39mFaaS: A Contemporary, Comparative Study
18mIntroduction of FaaS (Function as a Service) in starlingX
37mGo FaaSter: Cold Start Optimization in a Serverless Platform
37mGo FaaSter: Cold Start Optimization in a Serverless Platform
28mOperating a Global-Scale FaaS on Top of Kubernetes
39mFunction as a Service in the Cloud — Why You'll Still Use It
