Falcor "One Model Everywhere"
summary
Claim: Falcor exposes many cloud sources (Netflix uses over 80 microservices) as a single virtual JSON model so clients request only fragments, cached on-device. Implementation details: uses JsonGraph references to dedupe entities, an HTTP data source and express middleware to serve model.json, and a Falcor router that lazily builds path/value responses from databases. Demo moves an in-memory model to a Node server and back, showing on-demand fetching, client caching, and a working end-to-end pipeline; compares tradeoffs with GraphQL (schema/query vs. lightweight JSON paths).
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:26Why do we need one
- 1:28Falcor solution
- 2:15Falkor vs GraphQL
- 4:14Design your domain
- 4:49Simplify your model
- 5:44Falkor
- 9:24GraphQL
- 9:41Jason Graph
- 12:01Data Source Interface
- 14:43Router
- 17:43GraphQL vs Falcor
- 20:05Type systems
- 21:38Query syntax
- 23:46MapReduce
- 25:59Architectures
related talks
20mFalcor: Simplifying Your Data
20mBinding to the Cloud with Falcor
53mGraphQL vs Falcor
38mBFF vs GraphQL: A critical perspective
43mGraphQL in the Era of Server Components
37mGraphQL and Microservices or How we learned to be big and small at the same time
16mUnify Data Sources with GraphQL at the Edge
29mBig Graphs, Tiny Contexts: Dev Tools for Agents