ACK Lingjun provides fully managed Container Service for Kubernetes (ACK) clusters on Intelligent Computing LINGJUN ACK Lingjun is a cluster type for Intelligent Computing Lingjun that provides fully managed, highly available Kubernetes clusters with Lingjun compute nodes as worker nodes. This topic covers the product overview, features, and core advantages.
Table of contents
Usage notes
To use ACK Lingjun managed clusters, you must first create a Lingjun cluster with ACK activated in the Intelligent Computing LINGJUN console.
Related topics:
Overview
ACK Lingjun managed clusters provide fully managed, highly available control planes with heterogeneous resource management and task scheduling. These clusters serve as the cloud-native base of Platform for AI, with enhanced capabilities for AI and High Performance Computing (HPC) scenarios. The following figure shows the cluster architecture, which decouples software from hardware and integrates with Alibaba Cloud services for stable, reliable, efficient, and secure cloud-native AI infrastructure.
Features
-
Cluster management
ACK Lingjun shares the same cluster management capabilities as ACK Pro. ACK manages the control plane, deployed across three zones by default for high availability. Lifecycle operations include granting permissions, monitoring, updating, and managing components.
-
Node management
Lingjun node pools host Lingjun computing nodes and provide the same management and O&M features as Elastic Compute Service (ECS) node pools, such as batch add/remove, configuration, maintenance, fully managed nodes, application scheduling, monitoring, diagnostics, and automated O&M.
-
Cloud-native AI
Built-in components enhance cloud-native capabilities, including topology-aware multi-GPU scheduling and eGPU-based GPU virtualization for scheduling and isolation. Additional features include gang scheduling, capacity scheduling, binpack policy, dataset orchestration, and access acceleration.
Competitive advantages
-
Security and stability
ACK Lingjun shares ACK Pro enterprise-class features with highly available managed control planes, eliminating manual cluster setup. SLAs with compensation clauses ensure stability, reliability, and security for large-scale production environments.
-
Simplified O&M
Deeply integrated with Intelligent Computing LINGJUN and Alibaba Cloud services, ACK Lingjun provides Kubernetes-native services that simplify operations and automate O&M. Lingjun nodes offer the same management experience as ECS nodes, reducing adaptation and O&M costs.
-
Improved efficiency and acceleration
GPU sharing and topology-aware GPU scheduling improve heterogeneous resource efficiency. Priority-based job queue management and rich scheduling policies accelerate AI training and inference tasks, providing a unified method to manage and deliver AI resources and workloads.