Milvus 2.0: A Cloud-Native Vector Database for AI Applications

Explore a comprehensive conference talk on Milvus 2.0, a cloud-native vector database designed for AI applications. Delve into the growing importance of vector data in AI-driven unstructured data analytics and learn how Milvus, an open-source project, has rapidly gained over 1000 enterprise users worldwide. Discover the key design considerations, system architecture, and major components of Milvus 2.0, which features a flexible, scalable, and cloud-native design with read-write and compute-storage decoupling. Gain insights into the challenges faced during development and understand how Milvus 2.0 addresses the evolving needs of AI applications across various industries. The presentation covers topics such as unstructured data trends, AI-driven data search and analytics, system overview, features, and the innovative "Bazaar Architecture" approach.

Intro
Unstructured Data Trend
Unstructured Data Growth
Al-driven Data Search and Analytics
The Big Picture
Milvus is an LF AI & Data Graduated Project
Milvus Community
Milvus is used in almost every industry
Milvus is used for almost every type of unstructured data
System Overview
Features
Log Sequence Pub-sub as System Backbone
Access Layer
Coordinator Layer
Worker Layer
Storage Layer
Combine Streaming and Batching
The Bazaar Architecture
History of Milvus
Milvus Insight
Milvus CLI
K8s Operator