Enable Virtual Kubernetes Service (VKS)

Once you have an Alaya NeW enterprise account with sufficient balance, follow the steps below to provision a Virtual Kubernetes Service (VKS) cluster.

Prerequisites

• An Alaya NeW enterprise account and password. If you have not registered, see Sign up.
• Sufficient account balance for the planned inference / training workloads. For current promotions and pricing details, contact us.

Steps

1. Sign in to the Alaya NeW platform with your enterprise account, then go to Products / Virtual Kubernetes Service to open the VKS overview page.

2. Click New Cluster to open the VKS provisioning page and configure the cluster name, description, region, and other basic information.

   Parameter	Description	Requirements	Example	Required
   Cluster name	Identifier for the cluster; uniquely identifies it in the system	Must start with a letter; supports letters, digits, hyphen (-), underscore (_); 4–20 characters	jskfb-zqf	Yes
   Cluster description	A short description of the cluster's function, purpose, and configuration	—	—	Yes
   Region	The data center hosting the LLM inference / training service	Pick an available data center, e.g. Beijing Zone 1, Beijing Zone 2, Beijing Zone 3	—	Yes
   Cluster type	Shared: all authorized users share the entire VKS workspace with no isolation between users — suited for team collaboration. Dedicated: each authorized user gets workspace isolation via a Namespace inside the VKS, with per-user resource limits	Pick by business needs	—	Yes
   Billing mode	The billing method for data-center resources	E.g. pay-as-you-go	—	Yes
   Compute spec	Hardware specification of the data center, including GPU model, DCU per card per hour, quota, etc.	Pick resources that meet your inference / training service needs	—	Yes
   Storage configuration	Storage type (e.g. mass storage), status, capacity, storage service allocation, etc. The "10 GB free" option can be selected	Pick a configuration that meets your data-persistence needs	—	Yes
   Optional	Extra features as needed, e.g. enabling external service exposure	—	—	No

   • For compute spec, only pick the GPU model and the number of GPU cards you need; do not exceed the maximum allowed by your current entitlement.
   • You can scale enterprise storage on demand; mass storage must be allocated to file storage, object storage, or the image registry before it can be used.

3. Once the parameters are configured, click Enable Now to complete cluster provisioning. You can view the new cluster on the Resource Center / Virtual Kubernetes Service page. When the status reads "Running", the cluster is ready to use.

   

Next steps