Data Consistency Patterns in Cloud Native Applications
summary
Compares concrete patterns: Distributed Saga (AP) requires idempotent workers and compensating actions and often leaves unsafe intermediate states; Akka clustering pins sessions and uses CRDTs for conflict-free counters but lacks transaction isolation; Fauna (Calvin protocol) gives deterministic single-phase global ordering, temporal snapshots and a demo loading Northwinds then performing a read-then-insert as one serializable transaction. Calls out trade-offs: sagas trade isolation for availability, global-transaction databases pay higher latency and complexity for serializability.
AI-written summary · about
chapters (8)
- 0:00Introduction to data consistency
- 3:10Consistency models overview
- 7:09Actor clustering pattern
- 9:14AP consistency with sagas
- 12:35CP consistency at scale
- 17:45Deterministic databases
- 23:52FaunaDB demo and query language
- 29:03Q&A session
related talks
58mElements of Distributed Architectures
01h 01mLost in transaction Strategies to manage consistency in distributed systems
52mArchitectural Patterns for Distributed and Decentralized Data
46mLost in transaction? Strategies to deal with (in-)consistency in distributed systems
17mSaga Pattern for Distributed Transactions
30mCONSCIOUS CONSISTENCY WITH AKKA CLUSTER, CRDTS AND DISTRIBUTED DATA
27mData Consistency in Microservices Architecture
36mLost In Transaction