ModelArts Resource Pools

When using ModelArts to implement AI Development Lifecycle, you can use two different resource pools to train and deploy models.

• Public Resource Pool: provides public large-scale computing clusters, which are allocated based on job parameter settings. Resources are isolated by job. Billing of public resource pools is based on the resource specifications, duration, and instance quantity, regardless of the tasks (including training, deployment, and development) where the public resource pools are used. Public resource pools are provided by ModelArts by default and do not need to be created or configured. You can directly select a public resource pool during AI development.
• Dedicated Resource Pool: provides exclusive compute resources, which can be used for notebook instances, training jobs, and model deployment. It delivers higher efficiency and cannot be shared with other users.

  Buy a dedicated resource pool and select the dedicated resource pool during AI development. For details about the dedicated resource pool, see the following:

  Dedicated Resource Pool

  Creating a Dedicated Resource Pool

  Scaling a Dedicated Resource Pool

  Deleting a Dedicated Resource Pool

Dedicated Resource Pool

• Dedicated resource pools can be used in the following jobs and tasks: notebook instances, training, TensorBoard, and deployment.
• Dedicated resource pools are classified into two types: Dedicated for Development/Training and Dedicated for Service Deployment. The Dedicated for Development/Training type can be used only for notebook instances, training, and TensorBoard. The Dedicated for Service Deployment type can be used only for model deployment.
• Dedicated resource pools are available only when they are in the Running status. If a dedicated resource pool is unavailable or abnormal, rectify the fault before using it.
• After a dedicated resource pool is created, the billing starts based on the selected specifications.
• Dedicated resource pools can be billed in pay-per-use or yearly/monthly mode.

Creating a Dedicated Resource Pool

1. Log in to the ModelArts management console and choose Dedicated Resource Pools on the left.
2. On the Dedicated Resource Pools page, select Dedicated for Development/Training or Dedicated for Service Deployment.
3. Click Create in the upper left corner. The page for creating a dedicated resource pool is displayed.
4. Set the parameters on the page. For details about how to set parameters, see Table 1 and Table 2.

   Table 1 Parameters of the Dedicated for Development/Training type

   Parameter	Description
   Resource Type	The default value is Dedicated for Development/Training and cannot be changed.
   Billing Mode	Select a billing mode, Yearly/Monthly or Pay-per-use.
   Name	Name of a dedicated resource pool. The value can contain uppercase and lowercase letters, digits, hyphens (-), and underscores (_).
   Description	Brief description of a dedicated resource pool.
   Nodes	Select the number of nodes in a dedicated resource pool. More nodes mean higher computing performance and a higher cost.
   Specifications	Required specifications. The GPU delivers better performance, and the CPU is more cost-effective. If a flavor is sold out, you can purchase it again only after other users delete the resource pool.
   Required Duration	Select the time length when you want to use the resource pool. This parameter is mandatory only when the Yearly/Monthly billing mode is selected. The duration ranges from one month to one year. ModelArts provides a 1-year preference package, which allows you to enjoy the product for 1 year by paying for only 10 months. Auto renewal. With auto-renewal, the system automatically renews your product before the product expires.

   Table 2 Parameters of the Dedicated for Service Deployment type

   Parameter	Description
   Resource Type	The default value is Dedicated for Service Deployment and cannot be changed.
   Billing Mode	Select a billing mode, Yearly/Monthly or Pay-per-use. The Yearly/Monthly billing mode is supported only in CN North-Beijing4.
   Name	Name of a dedicated resource pool. The value can contain uppercase and lowercase letters, digits, hyphens (-), and underscores (_).
   Description	Brief description of a dedicated resource pool.
   Custom Network Configuration	If you enable Custom Network Configuration, the service instance runs on the specified network and can communicate with other cloud service resource instances on the network. If you do not enable Custom Network Configuration, ModelArts allocates a dedicated network to each user and isolates users from each other. If you enable Custom Network Configuration, set VPC, Subnet, and Security Group. If no network is available, go to the VPC service and create a network. .
   AZ	You can select Random, AZ 1, AZ 2, or AZ 3 based on site requirements. An AZ is a physical region where resources use independent power supplies and networks. AZs are physically isolated but interconnected through an internal network. To enhance workload availability, create nodes in different AZs.
   Nodes	Select the number of nodes in a dedicated resource pool. More nodes mean higher computing performance and a higher cost.
   Specifications	Required specifications. The GPU delivers better performance, and the CPU is more cost-effective. If a flavor is sold out, you can purchase it again only after other users delete the resource pool.
   Required Duration	Select the time length when you want to use the resource pool. This parameter is mandatory only when the Yearly/Monthly billing mode is selected. The duration ranges from one month to 11 months. Auto renewal. With auto-renewal, the system automatically renews your product before the product expires.

5. After confirming that the specifications are correct, create an dedicated resource pool as prompted. After a dedicated resource pool is created, its status changes to Running.

Scaling a Dedicated Resource Pool

After a dedicated resource pool is used for a period of time, you can scale out or in the capacity of the resource pool by increasing or decreasing the number of nodes.

A dedicated resource pool in Yearly/Monthly billing mode does not support scaling. If you buy a dedicated resource pool in Pay-per-use billing mode, you need to pay for the nodes after scaling.

The procedure for scaling is as follows:

1. Go to the dedicated resource pool management page, locate the row that contains the desired dedicated resource pool, and click Scale in the Operation column.
2. On the scaling page, increase or decrease the number of nodes. Increasing the node quantity scales out the resource pool whereas decreasing the node quantity scales in the resource pool. Scale the capacity based on service requirements.
   • During scale-out, select the quota of the current account and increase the number of nodes. Otherwise, the scale-out will fail.
   • During scale-in, switch off the desired node in the Operation column to delete the node. To reduce one node, you need to switch off the node in Node List to delete the node. See Figure 1.
   Figure 1 Switching off a node to delete the node during scale-in
   
3. Click Submit. After the request is submitted, the dedicated resource pool management page is displayed.

Deleting a Dedicated Resource Pool

If a dedicated resource pool is no longer needed during AI service development, you can delete the resource pool to release resources and reduce costs.

• After a dedicated resource pool is deleted, the training jobs, notebook instances, and deployment that depend on the resource pool are unavailable. A dedicated resource pool cannot be restored after being deleted. Exercise caution when deleting a dedicated resource pool.
• A dedicated resource pool in Yearly/Monthly billing mode cannot be deleted.

1. Go to the dedicated resource pool management page, locate the row that contains the desired dedicated resource pool, and click Delete in the Operation column.
2. In the dialog box that is displayed, click OK.