Milvus (vector database)

<h2 id="history">History</h2>
Milvus has been developed by Zilliz since 2017.<a class="footnote-ref" id="fnref:2" href="#fn:2">2</a>
Milvus joined <a href="/facts/Linux_Foundation/5wbe2Vp0">Linux Foundation</a> as an incubation project in January 2020 and became a graduate in June 2021.<a class="footnote-ref" id="fnref:3" href="#fn:3">3</a> The details about its architecture and possible applications were presented on <a href="/facts/ACM_SIGMOD_Conference/WjPo3dWP">ACM SIGMOD Conference</a> in 2021<a class="footnote-ref" id="fnref:4" href="#fn:4">4</a>
Milvus 2.0, a major redesign of the whole product with a new architecture,<a class="footnote-ref" id="fnref:5" href="#fn:5">5</a> was released in January 2022.

<h2 id="features">Features</h2>
<h3>Similarity search</h3>
Major <a href="/facts/Similarity_search/L55wjqpJ">similarity search</a> related features that are available in the active 2.4.x Milvus branch:<a class="footnote-ref" id="fnref:6" href="#fn:6">6</a>

<ul><li>In-memory, on-disk and GPU indices,</li>
<li>Single query, batch query and range query search,</li>
<li>Support of <a href="/facts/Sparse_vector/FFtN0XjR">sparse vectors</a>, <a href="/facts/Bit_array/PAfzTEHB">binary vectors</a>, <a href="/facts/JSON/bs1rUAhd">JSON</a> and arrays,</li>
<li><a href="/facts/FP32/e6Fez7BM">FP32</a>, <a href="/facts/FP16/7xPj7B3w">FP16</a> and <a href="/facts/BF16/wo1Q6IJY">BF16</a> data types,</li>
<li><a href="/facts/Euclidean_distance/9qDoQKQe">Euclidean distance</a>, <a href="/facts/Dot_product/tNz8MLjT">inner product distance</a> and <a href="/facts/Cosine_distance/vrbNb6wH">cosine distance</a> support for <a href="/facts/Floating-point_numbers/eIckahxe">floating-point data</a>,</li>
<li><a href="/facts/Hamming_distance/MgdtbYPo">Hamming distance</a> and <a href="/facts/Jaccard_distance/dUurmTet">jaccard distance</a> for <a href="/facts/Binary_data/PSFa5S2Q">binary data</a>,</li>
<li>Support of <a href="/facts/Nearest_neighbor_search/R6QTyAph">graph indices</a> (including <a href="/facts/Hierarchical_navigable_small_world/FT9gAjj2">HNSW</a>), <a href="/facts/Inverted_index/Kyva3prq">Inverted-lists</a> based indices and a <a href="/facts/Brute-force_search/YU2fzBIt">brute-force search</a>.</li>
<li>Support of <a href="/facts/Vector_quantization/rKNehCok">vector quantization</a> for <a href="/facts/Lossy_compression/jHW7DDv3">lossy input data compression</a>, including product quantization (PQ) and scalar quantization (SQ), that trades stored data size for accuracy,</li>
<li>Re-ranking.</li></ul>
Milvus <a href="/facts/Similarity_search/L55wjqpJ">similarity search</a> engine relies on heavily-modified <a href="/facts/Fork_(software_development)/D6khxhpr">forks</a> of third-party open-source <a href="/facts/Similarity_search/L55wjqpJ">similarity search</a> libraries, such as Faiss,<a class="footnote-ref" id="fnref:7" href="#fn:7">7</a><a class="footnote-ref" id="fnref:8" href="#fn:8">8</a> DiskANN<a class="footnote-ref" id="fnref:9" href="#fn:9">9</a><a class="footnote-ref" id="fnref:10" href="#fn:10">10</a> and hnswlib.<a class="footnote-ref" id="fnref:11" href="#fn:11">11</a>
Milvus includes optimizations for I/O data layout, specific to graph search indices.<a class="footnote-ref" id="fnref:12" href="#fn:12">12</a>

<h3>Database</h3>
As a <a href="/facts/Database/VYcpkevb">database</a>, Milvus provides the following features:<a class="footnote-ref" id="fnref:13" href="#fn:13">13</a>

<ul><li><a href="/facts/Column-oriented_DBMS/RHleA9BY">Column-oriented database</a></li>
<li>Four supported <a href="/facts/Data_consistency/WUxG1d41">data consistency</a> levels, including <a href="/facts/Strong_consistency/4cbcSQGc">strong consistency</a> and <a href="/facts/Eventual_consistency/Un19vCr0">eventual consistency</a>.<a class="footnote-ref" id="fnref:14" href="#fn:14">14</a></li>
<li><a href="/facts/Shard_(database_architecture)/YSE3YS0M">Data sharding</a></li>
<li>Streaming data ingestion, which allows to process and ingest data in real-time as it arrives</li>
<li>A dynamic schema, which allows inserting the data without a predefined schema</li>
<li>Independent storage and compute layers</li>
<li><a href="/facts/Multi-tenancy/vAxVpeUu">Multi-tenancy</a> scenarios (database-oriented, collection-oriented, partition-oriented)<a class="footnote-ref" id="fnref:15" href="#fn:15">15</a></li>
<li><a href="/facts/Memory-mapped_file/veYrNQnl">Memory-mapped</a> data storage</li>
<li><a href="/facts/RBAC/dPN1ohBE">Role-based access control</a></li>
<li>Multi-vector and hybrid search<a class="footnote-ref" id="fnref:16" href="#fn:16">16</a></li></ul>
<h3>Deployment options</h3>
Milvus can be deployed as an embedded database, standalone server, or distributed cluster. Zillis Cloud offers a fully managed version.<a class="footnote-ref" id="fnref:17" href="#fn:17">17</a>

<h3>GPU support</h3>
Milvus provides GPU accelerated index building and search using <a href="/facts/Nvidia/8QfCgqbu">Nvidia</a> <a href="/facts/CUDA/fvPpolot">CUDA</a> technology<a class="footnote-ref" id="fnref:18" href="#fn:18">18</a><a class="footnote-ref" id="fnref:19" href="#fn:19">19</a> via Nvidia RAFT library,<a class="footnote-ref" id="fnref:20" href="#fn:20">20</a> including a recent GPU-based graph indexing algorithm Nvidia CAGRA<a class="footnote-ref" id="fnref:21" href="#fn:21">21</a>

<h3>Integration</h3>
Milvus provides official SDK clients for <a href="/facts/Java_(programming_language)/9ScgFyAL">Java</a>, <a href="/facts/NodeJS/FKzhV4Qa">NodeJS</a>, <a href="/facts/Python_(programming_language)/YbuGqofa">Python</a> and <a href="/facts/Go_(programming_language)/oWlsmAFJ">Go</a>.<a class="footnote-ref" id="fnref:22" href="#fn:22">22</a> An additional <a href="/facts/C_sharp_(programming_language)/UC2gxeyb">C#</a> SDK client was contributed by <a href="/facts/Microsoft/nGIDDXdx">Microsoft</a>.<a class="footnote-ref" id="fnref:23" href="#fn:23">23</a><a class="footnote-ref" id="fnref:24" href="#fn:24">24</a> The database can integrate with <a href="/facts/Prometheus_(software)/Vlcfymdn">Prometheus</a> and <a href="/facts/Grafana/TRPcySe0">Grafana</a> for monitoring and alerts, frameworks Haystack<a class="footnote-ref" id="fnref:25" href="#fn:25">25</a> and LangChain,<a class="footnote-ref" id="fnref:26" href="#fn:26">26</a> <a href="/facts/Watsonx/SNyFMBng">IBM Watsonx</a>,<a class="footnote-ref" id="fnref:27" href="#fn:27">27</a> and OpenAI models.<a class="footnote-ref" id="fnref:28" href="#fn:28">28</a><a class="footnote-ref" id="fnref:29" href="#fn:29">29</a>

<h2 id="see-also">See also</h2>
<ul>
<li>Free and open-source software portal</li></ul>
<ul><li><a href="/facts/Nearest_neighbor_search/R6QTyAph">Nearest neighbor search</a></li>
<li><a href="/facts/Similarity_search/L55wjqpJ">Similarity search</a></li>
<li><a href="/facts/Vector_database/aKotamhB">Vector database</a></li>
<li><a href="/facts/Vector_embedding/7uRcBPqo">Vector embedding</a></li>
<li><a href="/facts/Vector_quantization/rKNehCok">Vector quantization</a></li></ul>

<h2 id="references">References</h2>

<ol>
<li id="fn:1">"LF AI & Data Foundation Announces Graduation of Milvus Project". June 23, 2021. <a href="https://lfaidata.foundation/blog/2021/06/23/lf-ai-data-foundation-announces-graduation-of-milvus-project/" target="_blank">https://lfaidata.foundation/blog/2021/06/23/lf-ai-data-foundation-announces-graduation-of-milvus-project/</a> <a href="#fnref:1" class="footnote-back-ref">↩</a></li>
<li id="fn:2">Liao, Ingrid Lunden and Rita (2022-08-24). "Zilliz raises $60M, relocates to SF". TechCrunch. Retrieved 2024-10-21. <a href="https://techcrunch.com/2022/08/24/zilliz-the-startup-behind-the-milvus-open-source-vector-database-for-ai-applications-raises-60m-and-relocates-to-sf/" target="_blank">https://techcrunch.com/2022/08/24/zilliz-the-startup-behind-the-milvus-open-source-vector-database-for-ai-applications-raises-60m-and-relocates-to-sf/</a> <a href="#fnref:2" class="footnote-back-ref">↩</a></li>
<li id="fn:3">"LF AI & Data Foundation Announces Graduation of Milvus Project". June 23, 2021. <a href="https://lfaidata.foundation/blog/2021/06/23/lf-ai-data-foundation-announces-graduation-of-milvus-project/" target="_blank">https://lfaidata.foundation/blog/2021/06/23/lf-ai-data-foundation-announces-graduation-of-milvus-project/</a> <a href="#fnref:3" class="footnote-back-ref">↩</a></li>
<li id="fn:4">"Milvus: A Purpose-Built Vector Data Management System". SIGMOD '21: Proceedings of the 2021 International Conference on Management of Data. June 18, 2021. pp. 2614–2627. doi:10.1145/3448016.3457550. ISBN 978-1-4503-8343-1. <a href="978-1-4503-8343-1" target="_blank">978-1-4503-8343-1</a> <a href="#fnref:4" class="footnote-back-ref">↩</a></li>
<li id="fn:5">Guo, Rentong; Luan, Xiaofan; Xiang, Long; Yan, Xiao; Yi, Xiaomeng; Luo, Jigao; Cheng, Qianya; Xu, Weizhi; Luo, Jiarui; Liu, Frank; Cao, Zhenshan; Qiao, Yanliang; Wang, Ting; Tang, Bo; Xie, Charles (2022). "Manu: A Cloud Native Vector Database Management System". arXiv:2206.13843 [cs.DB]. <a href="/wiki/ArXiv_(identifier)" target="_blank">/wiki/ArXiv_(identifier)</a> <a href="#fnref:5" class="footnote-back-ref">↩</a></li>
<li id="fn:6">"Milvus overview". Retrieved September 23, 2024. <a href="https://milvus.io/docs/overview.md" target="_blank">https://milvus.io/docs/overview.md</a> <a href="#fnref:6" class="footnote-back-ref">↩</a></li>
<li id="fn:7">"Faiss". GitHub. Retrieved September 23, 2024. <a href="https://github.com/facebookresearch/faiss" target="_blank">https://github.com/facebookresearch/faiss</a> <a href="#fnref:7" class="footnote-back-ref">↩</a></li>
<li id="fn:8">Douze, Matthijs; Guzhva, Alexandr; Deng, Chengqi; Johnson, Jeff; Szilvasy, Gergely; Mazaré, Pierre-Emmanuel; Lomeli, Maria; Hosseini, Lucas; Jégou, Hervé (2024). "The Faiss library". arXiv:2401.08281 [cs.LG]. <a href="/wiki/ArXiv_(identifier)" target="_blank">/wiki/ArXiv_(identifier)</a> <a href="#fnref:8" class="footnote-back-ref">↩</a></li>
<li id="fn:9">"DiskANN library". GitHub. Retrieved September 23, 2024. <a href="https://github.com/microsoft/DiskANN" target="_blank">https://github.com/microsoft/DiskANN</a> <a href="#fnref:9" class="footnote-back-ref">↩</a></li>
<li id="fn:10">Subramanya, Suhas Jayaram; Kadekodi, Rohan; Krishaswamy, Ravishankar; Simhadri, Harsha Vardhan (8 December 2019). "DiskANN: fast accurate billion-point nearest neighbor search on a single node". Proceedings of the 33rd International Conference on Neural Information Processing Systems. Curran Associates Inc.: 13766–13776. <a href="https://dl.acm.org/doi/abs/10.5555/3454287.3455520" target="_blank">https://dl.acm.org/doi/abs/10.5555/3454287.3455520</a> <a href="#fnref:10" class="footnote-back-ref">↩</a></li>
<li id="fn:11">"Hnswlib - fast approximate nearest neighbor search". GitHub. Retrieved September 23, 2024. <a href="https://github.com/nmslib/hnswlib" target="_blank">https://github.com/nmslib/hnswlib</a> <a href="#fnref:11" class="footnote-back-ref">↩</a></li>
<li id="fn:12">Wang, Mengzhao; Xu, Weizhi; Yi, Xiaomeng; Wu, Songlin; Peng, Zhangyang; Ke, Xiangyu; Gao, Yunjun; Xu, Xiaoliang; Guo, Rentong; Xie, Charles (2024). "Starling: An I/O-Efficient Disk-Resident Graph Index Framework for High-Dimensional Vector Similarity Search on Data Segment". Proceedings of the ACM on Management of Data. 2: 1–27. arXiv:2401.02116. doi:10.1145/3639269. <a href="/wiki/ArXiv_(identifier)" target="_blank">/wiki/ArXiv_(identifier)</a> <a href="#fnref:12" class="footnote-back-ref">↩</a></li>
<li id="fn:13">"Milvus overview". Retrieved September 23, 2024. <a href="https://milvus.io/docs/overview.md" target="_blank">https://milvus.io/docs/overview.md</a> <a href="#fnref:13" class="footnote-back-ref">↩</a></li>
<li id="fn:14">"Consistency levels in Milvus". Retrieved September 29, 2024. <a href="https://milvus.io/docs/consistency.md" target="_blank">https://milvus.io/docs/consistency.md</a> <a href="#fnref:14" class="footnote-back-ref">↩</a></li>
<li id="fn:15">"Multi-tenancy strategies". Retrieved September 29, 2024. <a href="https://milvus.io/docs/multi_tenancy.md" target="_blank">https://milvus.io/docs/multi_tenancy.md</a> <a href="#fnref:15" class="footnote-back-ref">↩</a></li>
<li id="fn:16">"Hybrid Search". Retrieved September 23, 2024. <a href="https://milvus.io/docs/multi-vector-search.md" target="_blank">https://milvus.io/docs/multi-vector-search.md</a> <a href="#fnref:16" class="footnote-back-ref">↩</a></li>
<li id="fn:17">"Zilliz cloud". Retrieved October 10, 2024. <a href="https://zilliz.com/cloud" target="_blank">https://zilliz.com/cloud</a> <a href="#fnref:17" class="footnote-back-ref">↩</a></li>
<li id="fn:18">"What's New In Milvus 2.3 Beta - 10X faster with GPUs". Retrieved September 29, 2024. <a href="https://zilliz.com/blog/milvus-2-3-beta-new-features-and-updates" target="_blank">https://zilliz.com/blog/milvus-2-3-beta-new-features-and-updates</a> <a href="#fnref:18" class="footnote-back-ref">↩</a></li>
<li id="fn:19">"Milvus 2.3 Launches with Support for Nvidia GPUs". 23 March 2023. Retrieved September 29, 2024. <a href="https://www.datanami.com/2023/03/23/milvus-2-3-launches-with-support-for-nvidia-gpus/" target="_blank">https://www.datanami.com/2023/03/23/milvus-2-3-launches-with-support-for-nvidia-gpus/</a> <a href="#fnref:19" class="footnote-back-ref">↩</a></li>
<li id="fn:20">"NVIDIA RAFT library". GitHub. <a href="https://github.com/rapidsai/raft" target="_blank">https://github.com/rapidsai/raft</a> <a href="#fnref:20" class="footnote-back-ref">↩</a></li>
<li id="fn:21">Ootomo, Hiroyuki; Naruse, Akira; Nolet, Corey; Wang, Ray; Feher, Tamas; Wang, Yong (August 2023). "CAGRA: Highly Parallel Graph Construction and Approximate Nearest Neighbor Search for GPUs". arXiv:2308.15136 [cs.DS]. <a href="/wiki/ArXiv_(identifier)" target="_blank">/wiki/ArXiv_(identifier)</a> <a href="#fnref:21" class="footnote-back-ref">↩</a></li>
<li id="fn:22">"Install Milvus Go SDK". Retrieved September 29, 2024. <a href="https://milvus.io/docs/install-go.md" target="_blank">https://milvus.io/docs/install-go.md</a> <a href="#fnref:22" class="footnote-back-ref">↩</a></li>
<li id="fn:23">"Milvus overview". Retrieved September 23, 2024. <a href="https://milvus.io/docs/overview.md" target="_blank">https://milvus.io/docs/overview.md</a> <a href="#fnref:23" class="footnote-back-ref">↩</a></li>
<li id="fn:24">"Get Started with Milvus Vector DB in .NET". March 6, 2024. Retrieved September 29, 2024. <a href="https://devblogs.microsoft.com/dotnet/get-started-milvus-vector-db-dotnet/" target="_blank">https://devblogs.microsoft.com/dotnet/get-started-milvus-vector-db-dotnet/</a> <a href="#fnref:24" class="footnote-back-ref">↩</a></li>
<li id="fn:25">"Integration HayStack + Milvus". Retrieved September 23, 2024. <a href="https://haystack.deepset.ai/integrations/milvus-document-store" target="_blank">https://haystack.deepset.ai/integrations/milvus-document-store</a> <a href="#fnref:25" class="footnote-back-ref">↩</a></li>
<li id="fn:26">"Milvus connector for LangChain". Retrieved September 23, 2024. <a href="https://python.langchain.com/docs/integrations/vectorstores/milvus/" target="_blank">https://python.langchain.com/docs/integrations/vectorstores/milvus/</a> <a href="#fnref:26" class="footnote-back-ref">↩</a></li>
<li id="fn:27">"IBM watsonx.data's integrated vector database: unify, prepare, and deliver your data for AI". IBM. April 9, 2024. Retrieved September 29, 2024. <a href="https://www.ibm.com/blog/announcement/ibm-watsonx-data-vector-database-ai-ready-data-management/" target="_blank">https://www.ibm.com/blog/announcement/ibm-watsonx-data-vector-database-ai-ready-data-management/</a> <a href="#fnref:27" class="footnote-back-ref">↩</a></li>
<li id="fn:28">"Getting started with Milvus and OpenAI". Mar 28, 2023. Retrieved September 23, 2024. <a href="https://cookbook.openai.com/examples/vector_databases/milvus/getting_started_with_milvus_and_openai" target="_blank">https://cookbook.openai.com/examples/vector_databases/milvus/getting_started_with_milvus_and_openai</a> <a href="#fnref:28" class="footnote-back-ref">↩</a></li>
<li id="fn:29">"OpenAI and Milvus simple app". GitHub. Retrieved September 23, 2024. <a href="https://github.com/Snaiel/OpenAI-Milvus-QA-Over-Docs" target="_blank">https://github.com/Snaiel/OpenAI-Milvus-QA-Over-Docs</a> <a href="#fnref:29" class="footnote-back-ref">↩</a></li>
</ol>

Milvus (vector database) open-in-new

Milvus (vector database)