Introducing HAMi

What is HAMi?

HAMi (Heterogeneous AI Computing Virtualization Middleware), formerly known as k8s-vGPU-scheduler, manages heterogeneous AI computing devices within Kubernetes clusters. It enables sharing of various AI devices while enforcing resource isolation between tasks, and provides a unified interface for different device types.

Why Choose HAMi?

Kubernetes Native API Compatibility

HAMi is compatible with Kubernetes' native API. Users can adopt HAMi without changing existing configurations, and Kubernetes default behavior is preserved.

Open and Neutral

HAMi is developed by stakeholders from internet services, finance, manufacturing, and cloud providers. Governance is open under the Cloud Native Computing Foundation (CNCF).

Avoid Vendor Lock-in

HAMi integrates with mainstream cloud providers without requiring proprietary vendor orchestration. Organizations can use their preferred cloud solutions alongside HAMi.

Resource Isolation

HAMi enforces resource isolation within containers. Each task is restricted to its allocated resources, preventing quota overuse. Hard limits improve security and stability.

Support for a Variety of Heterogeneous Computing Devices

HAMi supports GPUs, MLUs, NPUs, and other accelerators from multiple vendors. Devices can be shared across tasks, improving resource efficiency.

Unified Management

HAMi includes a unified monitoring system and configurable scheduling policies such as bin packing and spreading.

Get Started

HAMi works with existing Kubernetes configurations and supports multiple device vendors. See the installation guide to deploy it in your cluster.