BFF vs GraphQL: A critical perspective
summary
Claims that a client-specific middleware layer (BFF) cuts chattiness and fat payloads by aggregating multiple backend calls (Netflix did this to handle roughly 800 services) and that a declarative query layer (GraphQL) lets UIs request exactly the fields they need in a single round trip. Practical tradeoffs: more deployable services and potential duplicate logic, a new single point of failure for BFFs, and resolver-side N+1 and caching pitfalls for GraphQL. A pragmatic hybrid often uses a BFF that exposes a GraphQL schema resolving against existing REST services.
AI-written summary · about
chapters (8)
- 0:00Introduction and speaker background
- 1:07Distributed team challenges
- 8:15Traditional API pitfalls
- 14:36BFF pattern explained
- 23:48BFF concerns and trade-offs
- 24:09GraphQL as an alternative
- 35:09Mindset shift for developers
- 36:54Closing thoughts and Q&A
related talks
22mBFF - GraphQL Backend-for-Frontend with React UI
27mBackends for frontends with GraphQL
49mBFF (Backend For Frontend)
21mEnhance and improve user experience with BFF
23mKiling BFFs with GraphQL and Next.js
18mIs a GraphQL BFF Necessary in a Server Side React World (RSC, SAs)?
40mBFF4EAE: Your Backend's New Best Friend Forever
54mGiving the Frontend Only What It Needs, or Why You Need a BFF