Addressing the transaction challenge in a cloud-native world
summary
Claim: ACID two‑phase commits don’t scale well in cloud‑native systems; use Saga/BASE instead. The demo implements a holiday booking (flight, hotel, taxi) that uses MicroProfile LRA annotations (@LRA, @Complete, @Compensate) and runs on Open Liberty (Narayana coordinator) at localhost:9080/lrac. The approach uses compensating actions (cancel hotel, cancel taxi) when a flight is unavailable, accepts eventual consistency and no resource locking, and runs as an open‑source demo on GitHub (Open Liberty beta), proving the pattern is practical though tradeoffs remain.
AI-written summary · about
chapters (11)
- 0:00Introduction to cloud-native transactions
- 1:27Why cloud-native development matters
- 3:55Stateless vs stateful microservices
- 11:23Understanding transactions in apps
- 12:46ACID transactions limitations
- 21:09Introducing the Saga pattern
- 27:30Choreography vs orchestration
- 29:16Enabling transactions with MicroProfile
- 32:07MicroProfile LRA for cloud-native
- 39:08Summary and key takeaways
- 40:52Q&A session
related talks
31mAddressing the transaction challenge in a cloud-native world
39mA different flavor of the distributed transaction
50mTransactional solutions for microservices
53mA different flavor of the distributed transaction
55mTackling transactions in Microservices applications
46mLost in transaction? Strategies to deal with (in-)consistency in distributed systems
01h 01mLost in transaction Strategies to manage consistency in distributed systems
47mMicroservices Architecture and Data Consistency: So How Do We Do It for Real?