Milvus: Scaling Vector Data Solutions with Gen AI
summary
Milvus is an open source, cloud‑native vector database built on MinIO and Kafka and designed to be stateless on Kubernetes with etcd for metadata and S3-compatible object storage. Deployments range from Milvus Lite to Milvus Distributed that claims up to ~100 billion vectors; supported indexes include HNSW and DISKANN and GPU acceleration. The demo uses Cohere embeddings with ~35 million vectors and reports a search P99 of about 5 ms. Features: real‑time ingestion via Kafka/Pulsar, metadata filtering, hybrid search, dynamic autoscaling.
AI-written summary · about
chapters (11)
- 0:00Milvus and vector databases
- 2:28Vector databases in GenAI
- 4:54Traditional databases vs vector search
- 7:24Generating vectors from data
- 9:11Vector search in practice
- 10:13Scaling Milvus on Kubernetes
- 14:53Index strategies for search
- 18:20Filtering and search reduction
- 19:44Retrieval Augmented Generation with Milvus
- 22:54Live demo: 35M vectors
- 33:27Q&A
related talks
37mMilvus: Scaling Vector Data Solutions with Gen AI
36mMilvus: Scaling Vector Data Solutions with Gen AI
30mMilvus Vector Database: Unlocking the Future of Open Source Vector Databases
35mMilvus 2.0: Building the World's Most Advanced Open-source Vector Database
29mThe Evolution of Milvus: A Cloud-Native Vector Database
31mMilvus 2.0: A Vector Database With Cloud-Native Architecture
51mUnleashing the Potential of Cloud-Native Vector Databases
40mEvolution of Milvus Cloud-Scalable Vector Database