Elements of Distributed Architectures
summary
Distributed systems hinge on three linked binaries—transactional boundary (atomic vs eventual), communication (sync REST or gRPC vs async messaging), and workflow (orchestration vs choreography)—which produce eight combinations that map to named saga patterns like Parallel Saga or the “Horror Story.” The talk compares databases (RDBMS, MongoDB, Cassandra, graph, NewSQL), warns gRPC trades coupling for ~10× latency gains, shows an order/payment/inventory example, exposes fallacies about versioning and compensating updates, and reports all eight patterns exist in production.
AI-written summary · about
chapters (12)
- 0:00Distributed architecture challenges
- 2:13Choosing styles and databases
- 4:40Migration and granularity
- 6:41Data integrity and tradeoffs
- 8:35Code reuse strategies
- 12:21Fallacies of distributed computing
- 14:27Service contracts
- 16:35Communication protocols
- 21:00Workflow coordination
- 25:46Distributed transactions
- 29:50Interconnected architectural forces
- 38:34Transactional sagas
related talks
01h 02mCommunication Patterns for Distributed Architectures
40mCAPtivating architecture: Navigating Distributed Systems and Microservices
30mData Consistency Patterns in Cloud Native Applications
30mDo Distributed Systems, they said. It'll be fun, they said
27mBuilding Microservices Architecture The Right Way based on Saga Pattern
28mPractical Approaches To Scaling Distributed Transactions
44mDistributed journey
01h 00mChallenges of distributed systems