Milvus Standalone

Milvus — High-Performance Vector Database for AI Applications

Milvus is an open-source, high-performance vector database built for managing and searching large-scale unstructured data such as text, images, and multimodal embeddings. It is designed to power AI/ML workloads with fast and scalable similarity search.

Engineered in Go and C++, Milvus leverages hardware acceleration for both CPU and GPU to deliver low-latency, high-throughput vector queries.

Milvus Standalone:

• Distributed as a Docker image for fast and convenient deployment.
• Runs on a single virtual machine — no need for Kubernetes.
• Suitable for environments without high-scale requirements.
• Provides high availability via master-slave replication.

Milvus Standalone is an optimal solution for AI use cases such as recommendation systems, image and video search, and semantic text search — especially when simplicity of deployment and performance are key.

1. Get an SSH key pair to connect to the virtual machine (VM).

2. Create a VM from a public image. In the Image/Boot Disk Selection section, go to the Cloud Marketplace tab and select Milvus Standalone. In the Access section:

   • enter the username in the Login field;
   • paste the contents of the file with the public SSH key into the SSH Key field.

   Save the public IP address of the VM.

3. Connect to the VM via SSH. Use the login you specified when creating the VM and the private SSH key created earlier.

4. To create SSH tunnels and connect, use the commands:

   ssh -L 9091:localhost:9091 <username>@<public_IP_address_of_VM>
   

   Where:

   • <username> is the username.
   • <public_IP_address_of_VM> is the public IP address of the VM.

5. To open the web UI, go to http://localhost:9091/webui.

• Deploying a scalable, enterprise-grade solution
• Utilizing a highly reliable and distributed vector database with a comprehensive toolkit
• Scaling horizontally to manage billions of vectors

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.