How Data Delivery Evolved for Yandex’s Ad Engine
summary
A Yandex team describes changing the data format and introducing a new index to avoid constant field reshuffling and to enable fast delivery of updates to their advertising engine. The talk covers the data-format evolution and the indexing approach used to speed update propagation to ad-serving components.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 3:41Ad engine operation
- 7:12Old system problems
- 10:21FL buffers format
- 16:56Codegen and compatibility
- 19:02PlaniUm and partial indexes
- 24:12Data delivery acceleration
- 27:16Caching and cold data
- 29:24Final architecture
- 34:47Q&A
related talks
04mHighlights - Data and Business
41mReliability of Data Delivery in Stream Processing
21mPreparing Yandex search data — the library and processes we built
41mEfficiently Updating Database State from Stream Processing Services
29mSpeeding Up Image Preview Rendering in Yandex.Disk
26mImplementing YDB CDC: The Yandex Monitoring Case
51mSuperspeed. Yandex unified express delivery platform
41mHow to Extract Anything from the Entire Internet
