All Products
Document Center


Last Updated: Jul 23, 2019

Online inference is a key component in applying algorithm models to businesses. To help users apply algorithms from end to end, Machine Learning Platform for AI provides Elastic Algorithm Service (EAS) for online inference. EAS is used to load CPU compute and GPU compute-based models and handle service requests in real time. EAS provides multiple methods for you to deploy models online as RESTful APIs. It also supports auto scaling and blue-green deployment to deploy high-concurrent and stable online model services with the lowest costs.


EAS allows you to deploy a model online as a RESTful API. You can then send HTTP requests to the API to call the service.

  • Supported regions: China (Beijing) and China (Shanghai)
  • EAS supports version management, blue-green deployment (rolling update), and auto scaling. You can use these features to apply models to your businesses.


  • The service is charged after its status changes to Running. For more information, see Pricing.
  • You must bind the service to your domain name in API Gateway. Otherwise, you can only call the service up to 1,000 times per day by using the endpoint of the service.

Service fees

Model deployment methods

Manage deployed services

On the EAS-Online Model Service page, you can perform these actions to manage the deployed model services:
view the model invocation information,
perform online debugging,
view log data, monitoring data, and service deployment information, and
upgrade, downgrade, run, stop, and delete services.

Authorize RAM user accounts

If you want to use a RAM user account to manage models, you must first use your Alibaba Cloud account to authorize the RAM user account in Alibaba Cloud Resource Access Management (RAM).
Alibaba Cloud RAM:

  1. Log on to the RAM console, select Policies, click Create Policy, and select Script.


Add the following permissions:

Permission Description
eas:EditInstance The write permission for creation, updating, and deletion.
eas:ReadInstance The read permission for online debugging and checking the model service monitoring data, log data, and service endpoint.
eas:OperateInstance The operation permission for running or stopping model services, and distributing network traffic.
eas:ListInstance The permission for managing the models created by the current Alibaba Cloud account and its RAM user accounts. We do not recommend that you grant the full permission to a RAM user account. If you want RAM User A to manage a model created by RAM User B, you can grant RAM User A the permission to manage the model. Example: eas:*:*:eas/{modelId}
  1. {
  2. "Statement": [
  3. {
  4. "Effect": "Allow",
  5. "Action": [
  6. "eas:EditInstance",
  7. "eas:ReadInstance",
  8. "eas:OperateInstance"
  9. ]
  10. "Resource": "*"
  11. }
  12. ],
  13. "Version": "1"
  14. }
  1. Select Users, locate the RAM user account, and then authorized the account.


  1. To use the RAM user account to deploy models, you must manually bind the AccessKey pair of the RAM user account to DTplus. To view the AccessKey information of the RAM user account, use the Alibaba Cloud account to log on to the RAM console, select Users, click the RAM user account, and locate the AccessKey pair. If the RAM user account does not have an AccessKey pair, create an AccessKey pair for it.


  1. Use the RAM user account to log on to the DTplus console, click Personal Information, and bind the AccessKey pair. For more information, see

After you bind the AccessKey pair, you can then use the RAM user account to create model services and perform other operations in EAS.