Scaling Vector Search in Distributed Databases: The YDB Journey
Session Description
This presentation will explore the evolution of YDB, a distributed database, as it developed vector search capabilities. We'll discuss our journey from implementing a basic exact search to creating a global index system, designed to scale seamlessly across thousands of nodes and efficiently store and manage billions of vectors.
Key Takeaways
Understand the stages of developing vector search capabilities in a distributed database, from initial implementations to sophisticated search indices.
Gain insights into the technical challenges encountered in scaling vector search across thousands of nodes and what solutions were employed to overcome these barriers.
Learn about the innovative indexing techniques used to support efficient and effective vector search, capable of handling billions of data points.
Discover strategies and methods used to optimize the performance of vector search in distributed database environments.
References
Session Categories
Speakers
Alexander Zevaykin
Reviews
The abstract could have been more descriptive about what were the exact goals achieved from using YDB. While, its a promising SQL database with promises of ACID guarantees, an explanation of the same could have strengthened the abstract.
The description is lacking details about the structure of the talk, or the concepts discussed.
The description is too vague to make decision.
Interesting topic but the proposal is too vague to meaningfully evaluate this and make a decision.
The reviewers felt that the proposal was too vague and lacked sufficient detail for them to make a decision. The feedback noted that the description did not clearly outline the structure of the talk, the specific technical challenges you faced, or the exact goals you achieved with YDB. For future submissions, we highly recommend providing a more descriptive abstract with concrete examples and a clearer structure of your presentation.