Microservices for the Masses – Integrating a Distributed OS with Kubernetes
summary
An open-source runtime called Amino runtime integrates with Kubernetes and a DRA vm to provide transparent microservice replication, sharding, caching and live state migration by interposing on RPCs via configurable deployment managers. A prototype runs on Nexus phones and cloud nodes, migrates a physics/face‑recognition microservice to improve latency and reaches about 11 FPS in one demo. The system works end‑to‑end but remains experimental with language permutation rough edges and external state migration limitations.
AI-written summary · about
chapters (23)
- 0:00Intro
- 0:11Microservices for the Masses
- 0:27Overview
- 1:36Once upon a time, applications were..
- 5:15Turns out, it's still really, really difficult... Developers still have to write the
- 8:14Specialization...
- 11:30Amino OS Distributed Cloud-native Application Programming Platform
- 13:34Amino.Run Goals
- 15:32Amino.Run Architecture
- 17:28A brief word about multi-language and GraalVM
- 19:44Deployment Kernel
- 21:04Deployment Management Layer
- 23:28Deployment Manager Library
- 24:02Outline
- 24:55Deployment Manager Architecture
- 27:36Replicating a Microservice
- 29:38Caching Microservice State
- 31:40A Brief Note Regarding Composition of Deployment Managers
- 32:21Experimental Setup
- 34:12Code-offloading for Physics Engine
- 35:22Summary
- 36:03Next Steps?
- 37:17Get Involved
related talks
36mWhy is Cloud-Native Application Development Still So Hard?
01h 04mDeploying and Scaling Microservices
33mA Monolith, a Microservices, and a Function Walk into a Bar
49mDeveloping microservices applications on the Kubernetes platform
39mMicroservices & Kubernetes in a High-demand Environment: The Experience
51mDeploying And Scaling Microservices
34mDistributed applications and Kubernetes
53mNext Generation Microservices