S3 Access Grants and corporate directory identities - Amazon Simple Storage Service
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

S3 Access Grants and corporate directory identities

You can use Amazon S3 Access Grants to grant access to Amazon Identity and Access Management (IAM) principals (users or roles), both in the same Amazon Web Services account and in others. However, in many cases, the entity accessing the data is an end user from your corporate directory. Instead of granting access to IAM principals, you can use S3 Access Grants to grant access directly to your corporate users and groups. With S3 Access Grants, you no longer need to map your corporate identities to intermediate IAM principals in order to access your S3 data through your corporate applications.

This new functionality—support for using end-user identities access to data—is provided by associating your S3 Access Grants instance with an Amazon IAM Identity Center instance. IAM Identity Center supports standards-based identity providers and is the hub in Amazon for any services or features, including S3 Access Grants, that support end-user identities. IAM Identity Center provides authentication support for corporate identities through its trusted identity propagation feature. For more information, see Trusted identity propagation across applications.

To get started with workforce identity support in S3 Access Grants, as a prerequisite, you start in IAM Identity Center by configuring identity provisioning between your corporate identity provider and IAM Identity Center. IAM Identity Center supports corporate identity providers such as Okta, Microsoft Entra ID (formerly Azure Active Directory), or any other external identity provider (IdP) that supports the System for Cross-domain Identity Management (SCIM) protocol. When you connect IAM Identity Center to your IdP and enable automatic provisioning, the users and groups from your IdP are synchronized into the identity store in IAM Identity Center. After this step, IAM Identity Center has its own view of your users and groups, so that you can refer to them by using other Amazon Web Services and features, such as S3 Access Grants. For more information about configuring IAM Identity Center automatic provisioning, see Automatic provisioning in the Amazon IAM Identity Center User Guide.

IAM Identity Center is integrated with Amazon Organizations so that you can centrally manage permissions across multiple Amazon Web Services accounts without configuring each of your accounts manually. In a typical organization, your identity administrator configures one IAM Identity Center instance for the entire organization, as a single point of identity synchronization. This IAM Identity Center instance typically runs in a dedicated Amazon Web Services account in your organization. In this common configuration, you can refer to user and group identities in S3 Access Grants from any Amazon Web Services account in the organization.

However, if your Amazon Organizations administrator hasn't yet configured a central IAM Identity Center instance, you can create a local one in the same account as your S3 Access Grants instance. Such a configuration is more common for proof-of-concept or local development use cases. In all cases, the IAM Identity Center instance must be in the same Amazon Web Services Region as the S3 Access Grants instance to which it will be associated.

In the following diagram of an IAM Identity Center configuration with an external IdP, the IdP is configured with SCIM to synchronize the identity store from the IdP to the identity store in IAM Identity Center.

IAM Identity Center integration with an external identity store through automatic provisioning.

To use your corporate directory identities with S3 Access Grants, do the following:

How directory identities can access S3 data

Suppose that you have corporate directory users who need to access your S3 data through a corporate application, for example, a document-viewer application, that is integrated with your external IdP (for example, Okta) to authenticate users. Authentication of the user in these applications is typically done through redirects in the user's web browser. Because users in the directory are not IAM principals, your application needs IAM credentials with which it can call the S3 Access Grants GetDataAccess API operation to get access credentials to S3 data on the users' behalf. Unlike IAM users and roles who get credentials themselves, your application needs a way to represent a directory user, who isn't mapped to an IAM role, so that the user can get data access through S3 Access Grants.

This transition, from authenticated directory user to an IAM caller that can make requests to S3 Access Grants on behalf of the directory user, is done by the application through the trusted token issuer feature of IAM Identity Center. The application, after authenticating the directory user, has an identity token from the IdP (for example, Okta) that represents the directory user according to Okta. The trusted token issuer configuration in IAM Identity Center enables the application to exchange this Okta token (the Okta tenant is configured as the "trusted issuer") for a different identity token from IAM Identity Center that will securely represent the directory user within Amazon Web Services. The data application will then assume an IAM role, providing the directory user's token from IAM Identity Center as additional context. The application can use the resulting IAM session to call S3 Access Grants. The token represents both the identity of the application (the IAM principal itself) as well as the directory user's identity.

The main step of this transition is the token exchange. The application performs this token exchange by calling the CreateTokenWithIAM API operation in IAM Identity Center. Of course, that too is an Amazon API call and requires an IAM principal to sign it. The IAM principal that makes this request is typically an IAM role that's associated with the application. For example, if the application runs on Amazon EC2, the CreateTokenWithIAM request is typically performed by the IAM role that's associated with the EC2 instance on which the application runs. The result of a successful CreateTokenWithIAM call is a new identity token, which will be recognized within Amazon Web Services.

The next step, before the application can call GetDataAccess on the directory user's behalf, is for the application to obtain an IAM session that includes the directory user's identity. The application does this with an Amazon Security Token Service (Amazon STS) AssumeRole request that also includes the IAM Identity Center token for the directory user as additional identity context. This additional context is what enables IAM Identity Center to propagate the directory user's identity to the next step. The IAM role that the application assumes is the role that will need IAM permissions to call the GetDataAccess operation.

Having assumed the identity bearer IAM role with the IAM Identity Center token for the directory user as additional context, the application now has everything it needs to make a signed request to GetDataAccess on behalf of the authenticated directory user.

Token propagation is based on the following steps:

Create an IAM Identity Center application

First, create a new application in IAM Identity Center. This application will use a template that allows IAM Identity Center to identify which type of application settings that you can use. The command to create the application requires you to provide the IAM Identity Center instance Amazon Resource Name (ARN), an application name, and the application provider ARN. The application provider is the SAML or OAuth application provider that the application will use to make calls to IAM Identity Center.

To use the following example command, replace the user input placeholders with your own information:

aws sso-admin create-application \ --instance-arn "arn:aws-cn:sso:::instance/ssoins-ssoins-1234567890abcdef" \ --application-provider-arn "arn:aws-cn:sso::aws:applicationProvider/custom" \ --name MyDataApplication


{ "ApplicationArn": "arn:aws-cn:sso::123456789012:application/ssoins-ssoins-1234567890abcdef/apl-abcd1234a1b2c3d" }

Create a trusted token issuer

Now that you have your IAM Identity Center application, the next step is to configure a trusted token issuer that will be used to exchange your IdToken values from your IdP with IAM Identity Center tokens. In this step you need to provide the following items:

  • The identity provider issuer URL

  • The trusted token issuer name

  • The claim attribute path

  • The identity store attribute path

  • The JSON Web Key Set (JWKS) retrieval option

The claim attribute path is the identity provider attribute that will be used to map to the identity store attribute. Normally, the claim attribute path is the email address of the user, but you can use other attributes to perform the mapping.

Create a file called oidc-configuration.json with the following information. To use this file, replace the user input placeholders with your own information.

{ "OidcJwtConfiguration": { "IssuerUrl": "https://login.microsoftonline.com/a1b2c3d4-abcd-1234-b7d5-b154440ac123/v2.0", "ClaimAttributePath": "preferred_username", "IdentityStoreAttributePath": "userName", "JwksRetrievalOption": "OPEN_ID_DISCOVERY" } }

To create the trusted token issuer, run the following command. To use this example command, replace the user input placeholders with your own information.

aws sso-admin create-trusted-token-issuer \ --instance-arn "arn:aws-cn:sso:::instance/ssoins-1234567890abcdef" \ --name MyEntraIDTrustedIssuer \ --trusted-token-issuer-type OIDC_JWT \ --trusted-token-issuer-configuration file://./oidc-configuration.json


{ "TrustedTokenIssuerArn": "arn:aws-cn:sso::123456789012:trustedTokenIssuer/ssoins-1234567890abcdef/tti-43b4a822-1234-1234-1234-a1b2c3d41234" }

Connect the IAM Identity Center application with the trusted token issuer

The trusted token issuer requires a few more configuration settings to work. Set the audience that the trusted token issuer will trust. The audience is the value inside the IdToken that's identified by the key and can be found in the identity provider settings. For example:


Create a file named grant.json that contains the following content. To use this file, change the audience to match your identity provider settings and provide the trusted token issuer ARN that was returned by the previous command.

{ "JwtBearer": { "AuthorizedTokenIssuers": [ { "TrustedTokenIssuerArn": "arn:aws-cn:sso::123456789012:trustedTokenIssuer/ssoins-1234567890abcdef/tti-43b4a822-1234-1234-1234-a1b2c3d41234", "AuthorizedAudiences": [ "1234973b-abcd-1234-abcd-345c5a9c1234" ] } ] } }

Run the following example command. To use this command, replace the user input placeholders with your own information.

aws sso-admin put-application-grant \ --application-arn "arn:aws-cn:sso::123456789012:application/ssoins-ssoins-1234567890abcdef/apl-abcd1234a1b2c3d" \ --grant-type "urn:ietf:params:oauth:grant-type:jwt-bearer" \ --grant file://./grant.json \

This command sets the trusted token issuer with configuration settings to trust the audience in the grant.json file and link this audience with the application created in the first step for exchanging tokens of the type jwt-bearer. The string urn:ietf:params:oauth:grant-type:jwt-bearer is not an arbitrary string. It is a registered namespace in OAuth JSON Web Token (JWT) assertion profiles. You can find more information about this namespace in RFC 7523.

Next, use the following command to set up which scopes the trusted token issuer will include when exchanging IdToken values from your identity provider. For S3 Access Grants, the value for the --scope parameter is s3:access_grants:read_write.

aws sso-admin put-application-access-scope \ --application-arn "arn:aws-cn:sso::111122223333:application/ssoins-ssoins-111122223333abcdef/apl-abcd1234a1b2c3d" \ --scope "s3:access_grants:read_write"

The last step is to attach a resource policy to the IAM Identity Center application. This policy will allow your application IAM role to make requests to the API operation sso-oauth:CreateTokenWithIAM and receive the IdToken values from IAM Identity Center.

Create a file named authentication-method.json that contains the following content. Replace 123456789012 with your account ID.

{ "Iam": { "ActorPolicy": { "Version": "2012-10-17", "Statement": [ { "Effect": "Allow", "Principal": { "Amazon": "arn:aws-cn:iam::123456789012:role/webapp" }, "Action": "sso-oauth:CreateTokenWithIAM", "Resource": "*" } ] } } }

To attach the policy to the IAM Identity Center application, run the following command:

aws sso-admin put-application-authentication-method \ --application-arn "arn:aws-cn:sso::123456789012:application/ssoins-ssoins-1234567890abcdef/apl-abcd1234a1b2c3d" \ --authentication-method-type IAM \ --authentication-method file://./authentication-method.json

This completes the configuration settings for using S3 Access Grants with directory users through a web application. You can test this setup directly in the application or you can call the CreateTokenWithIAM API operation by using the following command from an allowed IAM role in the IAM Identity Center application policy:

aws sso-oidc create-token-with-iam \ --client-id "arn:aws-cn:sso::123456789012:application/ssoins-ssoins-1234567890abcdef/apl-abcd1234a1b2c3d" \ --grant-type urn:ietf:params:oauth:grant-type:jwt-bearer \ --assertion IdToken

The response will be similar to this:

{ "accessToken": "<suppressed long string to reduce space>", "tokenType": "Bearer", "expiresIn": 3600, "refreshToken": "<suppressed long string to reduce space>", "idToken": "<suppressed long string to reduce space>", "issuedTokenType": "urn:ietf:params:oauth:token-type:refresh_token", "scope": [ "sts:identity_context", "s3:access_grants:read_write", "openid", "aws" ] }

If you decode the IdToken value that is encoded with base64, you can see the key-value pairs in JSON format. The key sts:identity_context contains the value that your application needs to send in the sts:AssumeRole request to include the identity information of the directory user. Here is an example of the IdToken decoded:

{ "aws:identity_store_id": "d-996773e796", "sts:identity_context": "AQoJb3JpZ2luX2VjEOTtl;<SUPRESSED>", "sub": "83d43802-00b1-7054-db02-f1d683aacba5", "aws:instance_account": "123456789012", "iss": "https://identitycenter.amazonaws.com/ssoins-1234567890abcdef", "sts:audit_context": "AQoJb3JpZ2luX2VjEOT<SUPRESSED>==", "aws:identity_store_arn": "arn:aws:identitystore::232642235904:identitystore/d-996773e796", "aud": "abcd12344U0gi7n4Yyp0-WV1LWNlbnRyYWwtMQ", "aws:instance_arn": "arn:aws-cn:sso:::instance/ssoins-6987d7fb04cf7a51", "aws:credential_id": "EXAMPLEHI5glPh40y9TpApJn8...", "act": { "sub": "arn:aws-cn:sso::232642235904:trustedTokenIssuer/ssoins-6987d7fb04cf7a51/43b4a822-1020-7053-3631-cb2d3e28d10e" }, "auth_time": "2023-11-01T20:24:28Z", "exp": 1698873868, "iat": 1698870268 }

You can get the value from sts:identity_context and pass this information in an sts:AssumeRole call. The following is a CLI example of the syntax. The role to be assumed is a temporary role with permissions to invoke s3:GetDataAccess.

aws sts assume-role \ --role-arn "arn:aws-cn:iam::123456789012:role/temp-role" \ --role-session-name "TempDirectoryUserRole" \ --provided-contexts ProviderArn="arn:aws-cn:iam::aws:contextProvider/IdentityCenter",ContextAssertion="value from sts:identity_context"

You can now use the credentials received from this call to invoke the s3:GetDataAccess API operation and receive the final credentials with access to your S3 resources.