Explore the evolution of Milvus, a cloud-native vector database, in this informative conference talk. Delve into the challenges posed by the rapid increase in unstructured data and learn how embeddings can effectively represent and analyze this information. Discover the design principles behind Milvus 2.0, a fully distributed and horizontally scalable solution trusted by over 1000 organizations. Gain insights into the paradigm shift in vector database architecture, from version 1.0 to 2.0, and examine real-world applications. Understand key concepts such as unstructured data, embeddings, and vector databases, while exploring Milvus' features, cloud-native architecture, and general structure including access, coordinator, worker, and storage layers. Conclude with practical takeaways and diverse use cases, from art gallery experiences to threat detection, property search, and product recommendations.