Indemnity in the Data Layer
summary
Now in Android implements an offline first data layer that uses Room and DataStore for local storage, exposes read flows and async writes, and delegates queuing and network monitoring to WorkManager via a sync worker. Sync is push based with per-entity monotonic versioning, writes carry timestamps and resolve conflicts with last-write-wins, retries use exponential backoff, and repositories convert between network, entity, and UI models. The system runs in production and restores UI state within seconds after connectivity returns.
AI-written summary · about
chapters (10)
- 0:00Introduction to Indemnity
- 1:15What is the data layer
- 1:58Meaning of indemnity
- 3:01Offline first design basics
- 6:30Modeling data in offline apps
- 7:31Offline first reads strategy
- 10:41Offline first writes strategy
- 13:28Sync and conflict resolution
- 20:26Real world implementation
- 27:10Summary and Q&A
related talks
37mIdemnity in the Data Layer
06mCreate offline-first apps
38m"Offline" is not an error
42mAndroid Application Architecture
36m“Make your app work perfectly offl…” — Connection Interrupted
20mData Persistence In Android: There’s Room For Improvement
39mOffline Applications With Architecture Components, Retrofit, Room, Dagger, RxJava
17mFrom SQL to Flow: clean offline-first architecture with Kotlin 2.x