Workshop environment setup
The lab has been set up in form of a service catalog portfolio. To set up the AWS Service catalog portfolio, you need to deploy a CloudFormation template.
Task - Deploy CloudFormation template
- Open AWS CloudFormation console from us-east-2 region.
Choose Create Stack.
Specify stack name as
Accept the acknowledgement, then choose Create Stack.
Congratulations, your AWS Service Catalog portfolio has been set up successfully.
Task - Explore AWS Service Catalog portfolio
Now you need to open AWS Service Catalog and then choose Portfolios from the panel on the left and then select AWS Data and ML module portfolio to see how an AWS Service Catalog has been set up.
You will find following products:
- AWS Data Exchange product - This product when deployed would export data into an S3 bucket.
- AWS Marketplace - Machine Learning Model (CV) - This product when provisioned would deploy ML model in form of an Amazon SageMaker endpoint.
- An Amazon SageMaker Notebook instance product - This product when provisioned would create a notebook instance which you would use to perform an inference on the ML model.
- An IAM roles product - This product vends IAM roles used by other products.
- An application product - This product strings together all the other products in form of an application. To deploy the environment for the workshop, you would deploy the application product.
You can see how application product chains all other products together so you have to deploy only one product that:
- Creates IAM roles required.
- Passes IAM role to the AWS Data Exchange product that exports data to Amazon S3.
- Passes IAM roles to AWS Marketplace - Machine Learning Model (CV) product and deploys the ML model in form of an Amazon SageMaker endpoint.
- Passes IAM roles to An Amazon SageMaker Notebook instance product and associates the role with the notebook instance.
Task - Provision lab environment via AWS Service Catalog
To Provision AWS Service Catalog application:
- Navigate back to AWS CloudFormation service. Once CloudFormation stack with name
ml-workshop has finished execution, in outputs tab, click on the link corresponding to SwitchRoleAwsStudent and then click switch role.
- Open AWS Service Catalog.
- ChooseProducts from the panel on the left.
- Select An Application Product.
- Then choose LaunchProduct.
- Specify App as the Provisioned Product Name.
- Leave values for other parameters as they are and then Choose Launch Product.
Congratulations you just did step 3, step 4, and step 5 from the following diagram.
The next step is to use an Amazon SageMaker Notebook instance to do step 6 and step 7 from the architecture diagram.