Watch more from .local San Francisco → https://www.youtube.com/playlist?list=PL4RCxklHWZ9s7IrElTzddaZ2w5uupd6TQ Subscribe to MongoDB YouTube→ https://mdb.link/subscribe MongoDB Search and Vector Search bring full-text, vector, and hybrid retrieval into a single, managed platform, integrating the database, search engine, and sync mechanism so developers can index and query directly from familiar MongoDB APIs. In this talk, we peel back the layers of the architecture: how mongod and mongot cooperate to manage index definitions, build and maintain indexes, and execute $vectorSearch queries end-to-end. We’ll walk the lifecycle of an index, then trace the life of a query through the system. Finally, we’ll zoom out to scaling characteristics in replica sets and sharded clusters, and close with the next chapter: opening mongot to the community. 00:00:00 Introduction to the Builder Stage 00:00:59 Overview: Search & Vector Search Capabilities 00:02:00 Three Design Principles for Search Architecture 00:03:45 Leveraging Apache Lucene for Search 00:05:16 Index Lifecycle Management & Sequence Diagrams 00:06:50 Replication: Initial Sync vs. Steady State 00:10:07 Using the Oplog & Change Streams for Indexing 00:12:44 Managing Large Datasets with Iterative Initial Sync 00:15:03 Query Execution Flow & Document Retrieval 00:16:51 Deployment: Sidecar vs. Dedicated Search Nodes 00:19:31 Scaling Search with Sharding & Scatter-Gather 00:23:38 Roadmap: Binary Quantization & Auto-Scaling 00:25:14 Q&A: Search Scoring (BM25) & Vector Similarity Visit Mongodb.com → https://mdb.link/MongoDB Read the MongoDB Blog → https://mdb.link/Blog Read the Developer Blog → https://mdb.link/developerblog

 315      10