Semi Concurrent Deduplicated Incremental Delivery
summary
A deduplicated incremental response format lands in the graphql-js alpha, claiming one execution per field and smaller incremental payloads by grouping keys by their deferred-target sets via the new CollectFields algorithm, enabling semi-concurrent execution of deferred fragments. An implementation is merged and demoed with GraphQL Yoga (requires a small shim), and tooling suggestions include powerloader/Paraloader and resolver helpers. Caveats: deduplication must prevent re‑execution (not just skip delivery), early execution can block on shared resources, and gateways face extra complexity when subgraphs lack defer support.
AI-written summary · about
chapters (9)
- 0:00Introduction and background
- 3:35Motivating incremental delivery
- 11:13Defer syntax and advantages
- 15:46Deduplication challenges
- 18:13Deduplication theory and algorithm
- 24:38Advantages and limitations
- 27:44Early execution safety
- 29:28Demo and call to action
- 30:52Q&A
related talks
35mState of @Defer and @Stream
20mThe State of GraphQL Federation
31mEmbracing resiliency in GraphQL-Drop the cache-aside pattern and overcome the...
37mAll in client GraphQL
20mColocating & Composing Data Dependencies with Fragments
28mGraphQL without GraphQL - Backend Agnostic GraphQL inspired Queries For Colocating Data Requests With Components
25mThe Power of Strongly Coupled GraphQL Queries for Internal APIs
07mExperience-Oriented Graphs with Batching