白皮书

Milvus: A Purpose-Built Vector Data Management System

June 20, 2021

Whitepaper cover image

Used to build performant similarity search solutions

Recently, there has been a trend toward managing high-dimensional vector data in data science and AI applications. To use the proliferation of unstructured data, developers use machine learning (ML) to transform unstructured data into feature vectors for data analytics. Unfortunately, existing systems and algorithms for managing vector data have limited functions and usually incur serious performance issues when handling large-scale and dynamic vector data.

Milvus, a cloud-native, open-source vector database, stands out for its exceptional flexibility, reliability, and remarkable speed. It specializes in powering embedding similarity search and AI applications with a mission to make vector databases accessible to all organizations. Milvus efficiently manages large-scale vector data through purpose-built data management capabilities. This includes support for easy-to-use SDKs and RESTful APIs, optimization for heterogeneous computing platforms, advanced query processing, dynamic data handling for swift updates, and intelligent data distribution for scalability and availability.

Creating a large-scale similarity search service with Milvus takes minutes, thanks to its simple, intuitive SDKs in various languages. The vector database is hardware-efficient, featuring advanced indexing algorithms that result in a remarkable 10x performance boost in retrieval speed. Battle-tested by over a thousand enterprise users across diverse use cases, Milvus demonstrates resilience and reliability by extensively isolating individual system components.

Milvus's distributed and high-throughput nature makes it an ideal choice for serving large-scale vector data, adopting a systemic cloud-native approach that separates compute from storage and allows seamless scaling up and out. The vector search capabilities of Milvus encompass various data types, attribute filtering, UDF support, configurable consistency levels, time travel features, and more. Experience the power of Milvus, where accessibility, efficiency, and scalability converge to redefine the landscape of vector databases.

分享

获取白皮书