Lingjun AI Computing Service does not include built-in network connectivity. After activation, compute nodes are isolated from the Alibaba Cloud public cloud. Before running workloads, configure the companion services described in this topic to establish network connectivity, enable monitoring, and attach storage and container resources.
Required services at a glance
| Service | Purpose | Configuration required |
|---|---|---|
| Cloud Enterprise Network (CEN) | Connects Lingjun compute nodes to the Alibaba Cloud public cloud | Yes — create a CEN instance and transit router |
| Virtual Private Cloud (VPC) | Connects Lingjun AI Computing Service to cloud services in other VPCs | Yes — create a VPC and vSwitches |
| Application Real-Time Monitoring Service (ARMS) | Monitors cluster instances in real time; displays status on the Intelligent Computing LINGJUN console dashboard | No — automatically activated when you activate Lingjun AI Computing Service |
| Container Service for Kubernetes (ACK) | Manages large-scale data computing and high-performance data processing workloads | Yes — required for ACK Lingjun managed clusters |
| File Storage NAS (NAS) | Provides shared file storage for ACK Lingjun managed clusters | Yes — create a file system and mount target |
| Container Registry | Stores and serves container images for cluster workloads | Yes — create an instance, push images, and create a Kubernetes Secret |
| OAuth (Open Authorization) | Authenticates applications accessing the cluster | Yes — create an OAuth application, add scopes, and create a secret |
| ARMS application monitoring agent | Provides application-level monitoring within ACK Lingjun managed clusters | Yes — install and add to the cluster |
ARMS cluster monitoring is activated automatically. You do not need to configure it manually.
Network architecture overview
Choose your network structure
The applicable network structure depends on the instance type and region.
| Condition | Network structure | Next steps |
|---|---|---|
| Instance types gu7ef, gu7xf, or some gu8xf in the Ulanqab region | With Lingjun Connection | Configure CEN, VPC, Lingjun connection instances, and Cloud Parallel File Storage (CPFS) |
| All other regions and instance types | Without Lingjun Connection | Configure CEN and VPC only |
Network structure without Lingjun Connection
Network structure with Lingjun Connection
Compute nodes start in an isolated network. Connect them to CEN using Lingjun connection instances to reach the Alibaba Cloud public cloud. In addition to compute nodes, Lingjun connection instances, and CPFS, configure the following services:
CEN: Implements network connectivity between Lingjun AI Computing Service and the Alibaba Cloud public cloud. See the CEN configurations section.
VPC: Connects Lingjun AI Computing Service to cloud services in other VPCs after CEN is established. See the VPC configurations section.
ARMS: Monitors cluster instances in real time and displays their status on the dashboard. Activated automatically — no manual configuration needed.
CEN configurations
CEN connects Lingjun AI Computing Service to the Alibaba Cloud public cloud. For full service documentation, see the CEN documentation.
Create a CEN instance. See CEN instances.
Create a transit router for the CEN instance. See Transit routers.
Connect the VPCs where your other cloud services are deployed to the transit router:
Intra-region connection: Connect VPCs in the same region to the transit router so they can communicate with each other. See Create a VPC connection.
Inter-region connection: To access Alibaba Cloud services in a different region, create inter-region connections and allocate bandwidth. See Manage inter-region connections.
VPC configurations
Lingjun AI Computing Service and the cloud services it connects to run in separate VPCs. Create a VPC and vSwitches to enable network connectivity between them. See Create a VPC with an IPv4 CIDR block.
If a VPC already exists, reuse it for Lingjun services. Make sure the vSwitches in that VPC have idle IP addresses available.
The VPC IP addresses serve two purposes:
Monitoring networks: One IP address is assigned to the network used to monitor Lingjun AI Computing Service connectivity.
ARMS and other cloud services: VPCs where other cloud services are deployed connect to the CEN transit router, enabling Lingjun AI Computing Service to reach them.
After creating the VPC and vSwitches, use them when configuring your cluster.
Configurations of an ACK Lingjun managed cluster
If you need to process large-scale data computing and high-performance data processing workloads, you can activate Container Service for Kubernetes (ACK) to run them on Lingjun AI Computing Service.
If this is your first time using ACK, assign default roles to ACK and activate the required cloud services. See What is ACK Lingjun?
For the full list of cloud services to activate before using an ACK Lingjun managed cluster, see the "Cloud service fee" section of Billing for ACK LINGJUN clusters.
NAS configurations
Create a File Storage NAS (NAS) file system. See Create a file system.
ImportantThe NAS file system must be in the same VPC as the ACK Lingjun managed cluster.
Create a mount target. See CreateMountTarget.
Container Registry configurations
Create a Container Registry instance in the same VPC as the ACK Lingjun managed cluster.
Run
docker pushto push a base image to the image repositories in the Container Registry instance. See Use images of Container Registry to deploy applications in other cloud services.Create a secret object for the Docker registry from which images are pulled to the ACK Lingjun managed cluster.
OAuth configurations
Create an OAuth application and note the application name. See Create an application.
Add OAuth scopes to the application. See Manage OAuth scopes.
Create a secret for the application. See Create an application secret.
ARMS application monitoring agent configurations
Install the ARMS application monitoring agent. Follow "Step 1: Install the ARMS application monitoring agent" in Connect a registered cluster to Application Real-Time Monitoring Service.
Add the ARMS application monitoring agent to the ACK Lingjun managed cluster. When creating the cluster, deploy the following add-ons: prometheus, kube-state-metrics, gpu-exporter, and node-exporter.