On Geo-Distributed Transactions
summary
Explains limits of databases (both monolithic and distributed) that arise from physical-world constraints, and identifies application properties that let engineers work around those limits. Describes techniques and practical steps for when transactions must span geographically remote databases, i.e., approaches to implement geo-distributed transactions.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 1:08CAP theorem and system classes
- 4:43Business transaction needs
- 7:11Sync replication limits
- 8:15Algorithm core ideas
- 9:40SLOG: local transactions
- 11:26SLOG: global transactions
- 13:12Serialization and causal consistency
- 16:03Data recovery problems
- 18:13Testing SLOG algorithm
- 20:31DETOK: local transactions
- 20:51DETOK: global transactions
- 22:17Dependency graph issues
- 23:29Graph nodes and condensation
- 26:23Key movement between regions
- 27:05Comparison with 2PC
- 29:23Sberbank distributed systems
- 32:24Q&A
- 49:54Summary and gifts
related talks
28mPractical Approaches To Scaling Distributed Transactions
11mTowards a Digital Society: Investigating Integration of Distributed DBs and Distributed Processing
38mHow Geo-Distributed Apps Use Global Cloud Infra for Speed and Compliance
29mModern Database Architecture: From SQL to NoSQL & Distributed SQL
36mGeographically Distributed Prallel Databases - A Reality!
59mTransLattice Elastic Database Architecture Deep Dive
54mDistributed Graph DBMSs — The Future for Analytics on Big Data?
01mDon't Drop ACID - Transactions in Distributed NoSQL