Milvus distributed

Milvus is a high-performance cloud vector database for AI applications with hardware-accelerated search on CPU and GPU.

Benefits

• Horizontal scaling by separating computations and storage: supporting billions of vectors and thousands of concurrent queries.
• Supports all major vector index types: HNSW, IVF, FLAT, SCANN, DiskANN, as well as metadata filtering, hybrid search (dense and sparse vectors), and multi-tenancy.
• Supports use cases of semantic search, RAG, image search, and recommendation systems.
• Multiple deployment options: from a local instance to a fully distributed cluster.

1. Install the product from the Marketplace into an existing Kubernetes cluster.

2. Wait for the installation of all Milvus resources to complete.

3. Add a new rule to the security group of the cluster where Milvus is deployed:

   Traffic direction	Port range	Protocol	Destination / Source	CIDR blocks
   Ingress	443	TCP	CIDR	0.0.0.0/0

4. Get a temporary password for authorization:

   kubectl get cm milvus -n milvus -o jsonpath='{.data.default\.yaml}' | sed -nE 's/.*defaultRootPassword: "?([^"]+)"?.*/\1/p'
   

5. Change the password following the instructions.

6. Milvus will be available at the following addresses:

   • Attu: attu.<domain_name_specified_during_installation>.
   • Milvus gRPC: milvus.<domain_name_specified_during_installation>.
   • Milvus WebUI: Run the kubectl port-forward -n <namespace> service/milvus 27018:19530 command and open http://localhost:27018 in your browser.

Generation with additional retrieval (RAG)

# Use case
> Software engineer: "I need to improve the accuracy of my LLM model's responses using our own knowledge base with millions of documents."

Milvus solution: Integrate Milvus into your RAG pipeline for high-performance vector search. Store and retrieve relevant context from your knowledge base in milliseconds, reducing hallucinations and increasing response accuracy while processing billions of documents at scale.

Image search

# Use case
> User: "I want to find all images in my library similar to this photo of a mountain landscape."

Milvus solution: Build powerful image search systems with Milvus' efficient vector similarity search. Whether you are developing a personal photo library or a commercial image search application, Milvus provides the performance and scalability needed to instantly find visually similar images across massive datasets.

Semantic search

# Use case
> User: "Show me documents about the impact of artificial intelligence on healthcare."

Milvus solution: Go beyond keyword matching with Milvus' semantic search capabilities. Understand user intent and provide more relevant results by searching based on meaning rather than exact matches, significantly improving the search experience for your users.

Recommendation systems

# Use case
> Business requirement: "We need to personalize product recommendations in real time for millions of users."

Milvus solution: Deliver personalized recommendations at scale with Milvus. By efficiently storing and searching through vector representations of users and items, Milvus enables the creation of up-to-date, real-time recommendations that increase engagement and conversion on your platform.

Yandex Cloud technical support is available 24/7. The types of requests you can submit and the relevant response times depend on your pricing plan. You can switch to the paid support plan in the management console. You can learn more about the technical support terms and conditions here.