State machines for product data from YDB and queues
summary
Describes redesigning a product-data management system built on YDB and queueing to handle over a billion items, explaining why prior evolutionary scaling failed and how they addressed horizontal scaling limits. Covers concrete technologies: Yandex YDB for storage and message queues for processing, and presents the architectural and implementation changes made to improve scalability and data handling.
AI-written summary · about
chapters (5)
- 0:00Introduction
- 6:09Old architecture problems
- 11:57New architecture with YDB
- 24:01Implementation and testing
- 33:22Summary
related talks
50mHow We Built a Distributed Queue Service at Yandex
30mAn approach to unite tables and persistent queues in one system
40mData bus features for very large installations: the YDB Topics example
46mYDB Topic Service: How We Improved Message Queue Performance
27mData Modeling: Master This Critical Step to Adopting a Distributed Database
26mImplementing YDB CDC: The Yandex Monitoring Case
29mScalable Product Mappings
42mPetabytes in YDB over HDD in Yandex.Metrica Processing