Milvus 2.0 vs. 1.x: Cloud-native, distributed architecture, highly scalable, and more

	Milvus 1.x	Milvus 2.0
Architecture	Shared storage	Cloud native
Scalability	1 to 32 read nodes with only one write node	500+ nodes
Durability	Local disk Network file system (NFS)	Object storage service (OSS) Distributed file system (DFS)
Availability	99%	99.9%
Data consistency	Eventual consistency	Three levels of consistency: Strong Bounded Staleness Session Consistent prefix
Data types supported	Vectors	Vectors Fixed-length scalars String and text (in planning)
Basic operations supported	Data insertion Data deletion Approximate nearest neighbor (ANN) Search	Data insertion Data deletion (in planning) Data query Approximate nearest neighbor (ANN) Search Recurrent neural network (RNN) search (in planning)
Advanced features	Mishards Milvus DM	Scalar filtering Time Travel Multi-site deployment and multi-cloud integration Data management tools
Index types and libraries	Faiss Annoy Hnswlib RNSG	Faiss Annoy Hnswlib RNSG ScaNN (in planning) On-disk index (in planning)
SDKs	Python Java Go RESTful C++	Python Node Go (under development) Java (under development) RESTful (in planning) C++ (in planning)
Release status	Long-term support (LTS)	Release candidate. A stable version will be released in November.

Provide feedback