YTsaurus Dintables — yet another DBMS from Yandex
summary
Overview of YTsaurus dynamic tables (dyntables): how Yandex uses them as a new DBMS feature for dynamic, high-throughput workloads. Covers YTsaurus implementation details and notable features of dynatables, plus concrete usage scenarios inside Yandex.
AI-written summary · about
chapters (24)
- 0:00YTsaurus Dintables intro
- 1:13Origins and early challenges
- 3:26Distributed WAL and OLTP support
- 4:31Aggregating counters with functions
- 5:56OLTP vs MapReduce workloads
- 7:41MapReduce for recovery and debugging
- 8:47Yandex Search and antifraud integration
- 10:38Load balancing and interference
- 12:03Random and pushdown reads
- 13:48Antifraud dedicated clusters
- 15:12Bulk inserts and large transactions
- 15:54Multi-tenant isolation
- 16:57Metadata masters scaling
- 18:01Real-time ad counting demo
- 19:46Exactly-once distributed transactions
- 21:53In-memory data for low latency
- 23:00SSD and NVMe optimizations
- 24:46LSM-tree storage and versioning
- 27:16Columnar storage and execution
- 29:23Data center failure handling
- 31:30Cross-DC replication
- 32:35Developer experience and integration
- 33:39Community and support
- 34:01Q&A
related talks
48mYTsaurus — The Future of DWH, Already Here in Yandex Market
39mTransactional Replication in YTsaurus
51mYandex Database: Distributed Queries in the Clouds
01h 27mYugabyteDB YSQL Development
59mServerless nature of
40mDatabases in the Cloud
52mImplementing OLAP: how we built columnar storage in YDB
42mPetabytes in YDB over HDD in Yandex.Metrica Processing