Milvus
Milvus is an open-source, cloud-native vector database built for storing, indexing, and searching massive collections of embedding vectors at low latency. Developed by Zilliz, it is commonly used as the retrieval layer behind similarity search, recommendation, and RAG applications.
Baseline topologies
| Topology | Minimum nodes | Roles |
|---|---|---|
| standalone-vector-store | 1 | milvus |
| distributed-vector-cluster | 3 | milvus-query, milvus-data, milvus-coordination |
What drives the cost
- Vector count (millions)
- Approximate number of stored vectors before replica or distributed-system overhead.
- Embedding dimension
- Dimension of the embeddings stored in the primary collections.
- Peak query concurrency
- Expected peak concurrent similarity or filtered retrieval requests.
Where it runs
bare_metal, dedicated, cloud_vm
Size Milvus and get a cost estimate