Data bus features for very large installations: the YDB Topics example
summary
YDB Topics is presented as a data bus solution for very large installations (thousands of servers), focusing on the specific requirements and behavior such scale imposes. The talk covers design and operational differences of data buses at that scale, using YDB Topics as the concrete example technology to illustrate those challenges and approaches.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:11Data bus requirements
- 2:15YDB Topics history at Yandex
- 4:02Hardware limits
- 6:31Encoding optimization
- 8:23Server configs
- 9:46Fault tolerance
- 11:11OSS DC and Federation setups
- 14:05Load balancing
- 16:12YDB architecture
- 19:50Cluster updates
- 20:33Resource management
- 23:23Quotas nuances
- 25:11Metrics and audit
- 27:37Why OSS fails big firms
- 28:44YDB Topics roadmap
- 30:49Q&A
related talks
30mAn approach to unite tables and persistent queues in one system
47mState machines for product data from YDB and queues
44mYDB Java SDK: The Battle for Performance
29mData Modeling for Distributed SQL
48mIntegrating Distributed SQL Stores with Java Frameworks: The YDB Example
01h 00mParallel Asynchronous Replication b/t YDB Database Instances
26mImplementing YDB CDC: The Yandex Monitoring Case
38mTransactional Work with Topics