Using Alibaba Cloud ACK One Registration Cluster for Elastic Hybrid Cloud Deployment

As enterprises deepen digital transformation, the inflexibility and limited scalability of traditional IDC data‑centers become critical challenges; the ACK One registration cluster offers a highly flexible, resource‑rich solution to address these issues.

By integrating on‑premises Kubernetes clusters with an ACK One registration cluster, organizations gain elastic resource scaling, cost optimization through pay‑as‑you‑go pricing, and unified management.

Key advantages include:

Resource elasticity: on‑demand expansion of compute resources via Alibaba Cloud.

Cost optimization: avoid sunk costs of fixed IDC investments.

The architecture of the ACK One registration cluster is illustrated in the accompanying diagram, and the integration process follows these steps:

Create a registration cluster via console or CLI (see documentation [2] ).

Deploy the ack-cluster-agent on the IDC Kubernetes cluster.

Verify the agent status with kubectl -n kube-system get pod | grep ack-cluster-agent (expected output shows running pods).

The cloud‑side elastic resources include:

Elastic Container Instances (ECI) for on‑demand containers.

Elastic Compute Service (ECS) instances for rapid scaling.

GPU instances for AI training and big‑data workloads.

Three practical scenarios are presented:

Scenario 1: Registration Cluster with Cloud ECS Node Pool

When on‑prem resources are insufficient, a cloud node pool (CPU/GPU) can be added to extend capacity, supporting multiple OS images and billing models.

Scenario 2: ack‑virtual‑node (ECI)

Deploy the virtual‑node component to schedule Pods directly onto ECI without managing ECS nodes, using commands such as kubectl run nginx --image=registry.cn-hangzhou.aliyuncs.com/eci_open/nginx:1.14.2 -l alibabacloud.com/eci=true .

Scenario 3: ack‑co‑scheduler for Multi‑Level Elastic Scheduling

Install the co‑scheduler component and define ResourcePolicy objects to prefer IDC resources, then fall back to ECS node pools or ECI when needed, e.g.:

apiVersion: scheduling.alibabacloud.com/v1alpha1
kind: ResourcePolicy
metadata:
  name: schedued-to-node-pool
spec:
  selector:
    app: nginx
  strategy: prefer
  units:
  - resource: idc
  - resource: ecs
    nodeSelector:
      alibabacloud.com/nodepool-id: np6d13af2bb56546128***

Custom scripts can be used to detect GPU nodes and install drivers, as shown in the sample Bash script:

#!/bin/bash
# Check if lspci is installed
if ! which lspci &>/dev/null; then
  yum -y install pciutils
fi
# Detect NVIDIA GPU
if lspci | grep -i nvidia &>/dev/null; then
  echo "该节点为GPU节点，请安装相关驱动。"
fi

Finally, the guide summarizes that ACK One registration clusters provide unparalleled flexibility and scalability for enterprises, supporting ordinary compute, container instances, and high‑performance workloads, thereby accelerating digital transformation.