Vector Retrieval Service for Milvus (Milvus) is a fully managed vector search engine that is fully compatible with open-source Milvus, ensuring seamless migration. Built upon the open-source version, it delivers enhanced scalability for similarity searches on large-scale AI vector data. With its out-of-the-box usability, flexible scaling, and end-to-end monitoring and alerting, our managed Milvus service is the ideal choice for a wide range of AI applications, including multimodal search, Retrieval-Augmented Generation (RAG), recommendation engines, and content moderation. You can also leverage the open-source Attu tool for visual management to accelerate your application development and deployment.
Background information
Milvus is a cloud-native, open-source vector search engine. It is built on popular libraries such as Faiss, Annoy, and HNSW. Milvus is optimized for high availability, high performance, and easy scaling. It is suitable for the real-time retrieval of massive amounts of vector data. The engine includes advanced features such as data partitioning, sharding, persistence, incremental data ingestion, and hybrid search. It also supports time travel operations. Milvus provides an intuitive API and multi-language SDKs, making it suitable for various AI fields, including recommendation systems, image retrieval, video analysis, and natural language processing (NLP).
Benefits
Cloud-native, high-speed vector retrieval service
Leveraging a rich set of vector search libraries, the service delivers efficient and stable vector data retrieval with high performance, high availability, and support for hybrid search.
Enterprise-grade O&M and Ease of Use
As a fully managed cloud service, it eliminates cluster maintenance overhead and is ready to use out-of-the-box. Its cloud-native architecture delivers high performance and scalability with on-demand node scaling. The service includes built-in configuration management, security controls, and a comprehensive visual monitoring and alerting system to ensure operational stability and efficiency.
Compatibility with the open-source Milvus ecosystem
The service is fully compatible with open-source Milvus and provides rich management tools like Attu, backed by an extensive and active open-source community.
Features
A Managed, Scalable, and Enterprise-grade AI vector database for similarity search
High availability
Built on the Milvus O&M platform, the service guarantees 99.9% availability.
High scalability
Designed with a serverless architecture, our managed Milvus service supports rapid horizontal and vertical cluster scaling.
Open-source compatibility
This fully managed service is 100% compatible with open-source Milvus, offering a user experience identical to the native software. It includes the open-source visualization tool Attu by default.
High security
Deployed within an Alibaba Cloud Virtual Private Cloud (VPC), the service provides secure network access, fine-grained access control, and advanced security features.
Instance observability
The service provides comprehensive cluster metrics monitoring and alerting capabilities.
Advanced Capabilities and Ecosystem Integrations
Integration with Alibaba Cloud AI products
The service integrates with Alibaba Cloud's suite of AI products, including the Platform for AI (PAI) and Tongyi models. This provides a faster and easier-to-use experience for implementing enterprise AI applications.
Integration with upstream and downstream Alibaba Cloud products
The service integrates with Alibaba Cloud storage and big data products. This facilitates seamless data exchange between products and simplifies data engineering for AI applications.
Billing
Milvus is billed based on two components: Compute Units (CUs) and storage. For more information, see Billable items.