Users for Oracle Exadata Database Service on Dedicated Infrastructure can be centrally managed in a Microsoft Azure Active Directory (Azure AD) service.

You can perform this integration in the following Oracle Database environments:

• On-premises Oracle Database release 19.16 and later, but not for Oracle Database 21c
• Oracle Exadata Database Service on Dedicated Infrastructure on Database versions 19.17 and later. This feature is not supported on Oracle Database release 21c.
• Oracle Base Database Service

The instructions for configuring Azure AD use the term "Oracle Database" to encompass these environments.

This type of integration enables the Azure AD user to access an Oracle Exadata Database Service on Dedicated Infrastructure instance. Azure AD users and applications can log in with Azure AD Single Sign On (SSO) credentials to get an Azure AD OAuth2 access token to send to the database.

The administrator creates and configures the application registration (app registration) of the Oracle Exadata Database Service on Dedicated Infrastructure instance with Azure AD. The database administrator also creates application (app) roles for the database app registration in Azure AD, and assigns these roles to Azure AD users, groups, and applications. These app roles will be mapped to the database global schemas and global roles. An Azure AD principal that is assigned to an app role will be mapped to either a database global schema or database global role. An Oracle global schema can also be mapped exclusively to an Azure AD user. When the principal is a guest user or service principal, they can only be mapped to the database schema through an Azure app role. An Oracle global role can only be mapped to an Azure app role.

Tools and applications that are updated to support Azure AD tokens can authenticate users directly with Azure AD and pass the database access token to the Oracle Exadata Database Service on Dedicated Infrastructure instance. You can configure existing database tools such as SQL*Plus to use an Azure AD token from a file location. In these cases, Azure AD tokens can be retrieved using tools like Microsoft PowerShell or Azure CLI and put into a file location. An Azure AD OAuth2 database access token is a bearer token with an expiration time. The Oracle Database client driver will ensure that the token is in a valid format and that it has not expired before passing it to the database. The token is scoped for the database. Assigned app roles for the Azure AD principal are included as part of the access token. The directory location for the Azure AD token should only have enough permission for the user to write the token file to the location and the database client to retrieve these files (for example, just read and write by the process user). Because the token allows access to the database, it should be protected within the file system.

Azure AD users can request a token as a client registered with Azure AD app registration by using methods such as the following:

• Passing the Azure AD user name and password through a command line, script, file, or any other supported method
• Entering the Azure AD credentials into an Azure AD authentication screen with or without multi-factor authentication

Oracle Exadata Database Service on Dedicated Infrastructure supports the following Azure AD authentication flows:

• Resource owner password credential (ROPC), which is used in non-graphic user interface environments when a pop-up window cannot be used to authenticate a user.
• Authorization code, which is used when a browser can be used to enter credentials for the user
• Client credentials, which are for applications that connect as themselves (and not the end-user)
• On-Behalf-Of (OBO), where an application requests an access token on behalf of a logged-in user to send to the database

Oracle Exadata Database Service on Dedicated Infrastructure accepts tokens representing the following Azure AD principals:

• Azure AD user, who is registered user in the Azure AD tenancy
• Guest user, who is registered as a guest user in the Azure AD tenancy
• Service, which is the registered application connecting to the database as itself with the client credential flow (connection pool use case)

Microsoft Azure users must be mapped to an Oracle Database schema and have the necessary privileges (through roles) before being able to authenticate to the Oracle Database instance.

In Microsoft Azure, users, groups, and applications can be assigned to the app roles by an Azure AD administrator.

An Azure AD user can be mapped to a database schema (user) either exclusively or through an app role.

• Creating an exclusive mapping between an Azure AD user and an Oracle Database schema. In this type of mapping, the database schema must be created for the Azure AD user with the appropriate database privileges and roles. The schema must also be updated during the life cycle of the Azure AD user and eventually dropped when the user leaves.
• Creating a shared mapping between an Azure AD app role and an Oracle Database schema. This type of mapping, which is more common than exclusive mappings, is for Azure AD users who have been assigned directly to the app role or is a member of an Azure AD group that is assigned to the app role. The app role is mapped to an Oracle Database schema (shared schema mapping). Shared schema mapping allows multiple Azure AD users to share the same Oracle Database schema so a new database schema is not required to be created every time a new user joins the organization. This operational efficiency allows database administrators to focus on database application maintenance, performance, and tuning tasks instead of configuring new users, updating privileges and roles, and removing accounts.

In addition to database roles and privileges being granted directly to the mapped global schema, additional roles and privileges can be granted through mapped global roles. Different Azure AD users mapped to the same shared global schema may need different privileges and roles. Azure app roles can be mapped to Oracle Database global roles. Azure AD users who are assigned to the app role or are a member of an Azure AD group that is assigned to the app role will be granted the Oracle Database global role when they access the database.

The following diagram illustrates the different types of assignments and mappings that are available.

Figure 5-1 Assignments and Mappings Between Azure AD and Oracle Database

These mappings are as follows:

• An Azure AD user can be mapped directly to an Oracle Database global schema (user).
• An Azure AD user, Azure AD group, or application is assigned to an app role, which is then mapped to either an Oracle Database global schema (user) or a global role.

Oracle Database supports four types of use cases for connecting to an Oracle Database instance using Microsoft Azure Active Directory.

• Connection using OAuth 2.0 authorization flow: The client directs the resource owner to an authorization server, which in turn directs the resource owner back to the client with the authorization code. See the Microsoft Azure article Microsoft identity platform and OAuth 2.0 authorization code flow.
• Connection using the resource owner password credentials: The resource owner password credentials (that is, the user name and password) can be used directly to obtain an access token. Azure AD requires an additional client Id and a secret for this flow. (The secret is not required for public client.) See the Microsoft Azure article Microsoft identity platform and OAuth 2.0 Resource Owner Password Credentials.
• Connection using the client credentials: The client acts on its own behalf (the client is also the resource owner) or requests access to protected resources based on an authorization arranged with the authorization server. This flow is used to get the Azure OAuth2 access token for the service principal. An application can also request an Azure AD OAuth2 access token directly from Azure AD and pass it through a database client API. See the Microsoft Azure article Get Azure AD tokens by using a service principal.
• Connection using on-behalf-of (OBO) token: An Azure application requests an OBO token for a logged in user. The OBO token will also be an access token for the database with the Azure AD user identity and assigned app roles for the database. This enables the Azure AD user to log in to the database as the user and not the application. Only an application can request an OBO token for its Azure AD user and pass it to the database client through the API.

Both the Oracle and the Microsoft Azure administrators play roles in configuring the connection between Oracle Exadata Database Service on Dedicated Infrastructure and Microsoft Azure AD.

The general process is as follows:

1. The Oracle administrator ensures that the Oracle Database environment meets the requirements for the Microsoft Azure AD integration. See Oracle Database Requirements for the Microsoft Azure AD Integration.
2. The Oracle administrator registers the database instance with the Microsoft Azure AD tenancy and then enables the connection between the Oracle Exadata Database Service on Dedicated Infrastructure and the Azure AD endpoint.

   As part of the registration process, the Oracle administrator or the Azure administrator creates or designates Azure app roles to be used for the mappings between the Oracle database and the Microsoft Azure endpoint.

3. The Oracle administrator creates and maps global schemas to either an Azure AD user (exclusive schema mapping) or to an Azure app role (shared schema mapping). The Azure AD user or application must be mapped to one schema.
4. Optionally, the Oracle administrator creates and maps global Oracle Database roles to Azure app roles.
5. The Azure AD end user who wants to connect with the Oracle Exadata Database Service on Dedicated Infrastructure instance registers the client application as an Azure AD client (similar to how the Oracle database is registered).

   The Azure AD client will have a client identification and a client secret, unless the application client is public. If the application client is public, then only the application client identification is necessary.

6. The Azure AD end user (who can be a database administrator) connects using an utility such as PowerShell or the Azure command-line interface to retrieve the token and store it in a local file directory. An application can also request an Azure AD OAuth2 access token directly from Azure AD and pass it through a database client API. Refer to the following Oracle Database client documentation for information about passing Azure AD OAuth2 tokens:
   • JDBC-thin clients: Oracle Database JDBC Developer’s Guide
   • Oracle Call Interface (OCI): Oracle Call Interface Programmer's Guide
   • Oracle Data Provider for .NET (ODP): Oracle Data Provider for .NET Developer's Guide for Microsoft WindowsConnecting to Oracle Database
7. Once connected to the Oracle Exadata Database Service on Dedicated Infrastructure instance, the Azure AD end user performs tasks as needed.

Before using Azure AD authentication on databases in the Exadata Cloud Infrastructure, you must use the Networking service to add a service gateway, a route rule, and an egress security rule to the Virtual Cloud Network (VCN) and subnets where your database resources reside.

1. Create a service gateway in the VCN where your database resources reside by following the instructions in Task 1: Create the service gateway in OCI documentation.
2. After creating the service gateway, add a route rule and an egress security rule to each subnet (in the VCN) where the database resources reside so that these resources can use the gateway to use Azure AD authentication:
   1. Go to the Subnet Details page for the subnet.
   2. In the Subnet Information tab, click the name of the subnet's Route Table to display its Route Table Details page.
   3. In the table of existing Route Rules, check whether there is already a rule with the following characteristics:
      • Destination: 0.0.0.0/0
      • Target Type: NAT Gateway
      • Target: The name of the NAT gateway you just created in the VCN

      If such a rule does not exist, click Add Route Rules and add a route rule with these characteristics.

   4. Return to the Subnet Details page for the subnet.
   5. In the subnet's Security Lists table, click the name of the subnet's security list to display its Security List Details page.
   6. In the side menu, under Resources, click Egress Rules.
   7. In the table of existing Egress Rules, check whether there is already a rule with the following characteristics:
      • Destination Type: CIDR
      • Destination: 0.0.0.0/0
      • IP Protocol: TCP
      • Source Port Range: 443
      • Destination Port Range: All
   8. If such a rule does not exist, click Add Egress Rules and add an egress rule with these characteristics.

When sending Azure AD tokens from the database client to the database server, a TLS connection must be established. The TLS wallet with the database certificate for the ExaDB-D service instance must be stored under the WALLET_ROOT location. Create a tls directory so it looks like: WALLET_ROOT/<PDB GUID>/tls.

When configuring TLS between the database client and server there are several options to consider.

• Using a self-signed database server certificate vs a database server certificate signed by a commonly known certificate authority
• One-way TLS (TLS) vs Mutual or two-way TLS (mTLS)
• Client with or without a wallet

Self-Signed Certificate

Using a self-signed certificate is a common practice for internally facing IT resources since you can create these yourself and it's free. The resource (in our case, the database server) will have a self-signed certificate to authenticate itself to the database client. The self-signed certificate and root certificate will be stored in the database server wallet. For the database client to be able to recognize the database server certificate, a copy of the root certificate will also be needed on the client. This self-created root certificate can be stored in a client-side wallet or installed in the client system default certificate store (Windows and Linux only). When the session is established, the database client will check to see that the certificate sent over by the database server has been signed by the same root certificate.

A Well-Known Certificate Authority

Using a commonly known root certificate authority has some advantages in that the root certificate is most likely already stored in the client system default certificate store. There is no extra step for the client to store the root certificate if it is a common root certificate. The disadvantage is that this normally has a cost associated with it.

One-Way TLS

In the standard TLS session, only the server provides a certificate to the client to authenticate itself. The client doesn't need to have a separate client certificate to authenticate itself to the server (similar to how HTTPS sessions are established). While the database requires a wallet to store the server certificate, the only thing the client needs to have is the root certificate used to sign the server certificate.

Two-Way TLS (also called Mutual TLS, mTLS)

In mTLS, both the client and server have identity certificates that are presented to each other. In most cases, the same root certificate will have signed both of these certificates so the same root certificate can be used with the database server and client to authenticate the other certificate. mTLS is sometimes used to authenticate the user since the user identity is authenticated by the database server through the certificate. This is not necessary for passing IAM tokens but can be used when passing IAM tokens.

Client with a Wallet

A client wallet is mandatory when using mTLS to store the client certificate. However, the root certificate can be stored either in the same wallet or in the system default certificate store.

A Client without a Wallet

Clients can be configured without a wallet when using TLS under these conditions: 1) One-way TLS is being configured where the client does not have its own certificate and 2) the root certificate that signed the database server certificate is stored in the system default certificate store. The root certificate would most likely already be there if the server certificate is signed by a common certificate authority. If it's a self-signed certificate, then the root certificate would need to be installed in the system default certificate store to avoid using a client wallet.

For details on how to configure TLS between the database client and database server including the options described above, see Configuring Transport Layer Security Authentication in the Oracle Database Security Guide.

If you choose to use self-signed certificates and for additional wallet related tasks, see Managing Public Key Infrastructure (PKI) Elements in the Oracle Database Security Guide.

Before you can configure an Oracle Database instance with Microsoft Azure AD, you must ensure that your environment meets special requirements.

Note the following:

• The Oracle Database server must be able to request the Azure AD public key. Depending on the enterprise environment, you may need to configure a proxy setting.
• Users and applications that need to request an Azure AD token must also be able to connect to Azure AD. You may need to configure a proxy setting for the connection.
• You must configure Transport Layer Security (TLS) between the Oracle Database client and the Oracle Database server. This TLS connection can be either one-way or mutual.
• You can create the TLS server certificate to be self-signed or be signed by a well known root certificate. The advantage of using a certificate that is signed by a well known root certificate is that the database client can use the system default certificate store to validate the Oracle Database server certificate instead of having to create and maintain a local wallet with the root certificate. Note that this applies to Linux and Windows clients only.

A user with administrator privileges uses Microsoft Azure AD to register the Oracle Database instance with the Microsoft Azure AD tenancy.

1. Log in to the Azure portal as an administrator who has Microsoft Azure AD privileges to register applications.
2. In the Azure Active directory admin center page, from the left navigation bar, select Azure Active Directory.
3. In the MS - App registrations page, select App registrations from the left navigation bar.
4. Select New registration.
   The Register an application window appears.
   
5. In the Register an application page, enter the following Oracle Database instance registration information:
   • In the Name field, enter a name for the Oracle Database instance connection (for example, Example Database).
   • Under Supported account types, select the account type that matches your use case.
     • Accounts in this organizational directory only (tenant_name only - Single tenant)
     • Accounts in any organizational directory (Any Azure AD directory - Multitenant)
     • Accounts in any organizational directory (Any Azure AD directory - Multitenant) and personal Microsoft accounts (e.g. Skype, Xbox)
     • Personal Microsoft accounts only
6. Bypass the Redirect URI (Optional) settings. You do not need to create a redirect URI.
7. Click Register.
   After you click Register, Azure AD displays the app registration's Overview pane, which will show the Application (client) ID under Essentials. This value is a unique identifier for the application in the Microsoft identity platform.
8. Register a scope, which will set the permission for the registered app.
   1. In the left navigation bar, select Expose an API.
   2. Under Set the App ID URI, in the Application ID URI field, enter the app ID URI for the database connection using the following format, and then click Save:

      your_tenancy_url/application_(client)_id

      In this specification:

      • your_tenancy_url must include https as the prefix and the fully qualified domain name of your Azure AD tenancy.
      • application_(client)_id is the ID that was generated when you registered the Oracle Database instance with Azure AD. It is displayed in the Overview pane of the app registration.

      For example:

      https://sales_west.example.com/1aa11111-1a1z-1a11-1a1a-11aa11a1aa1a

   3. Select Add a scope and then enter the following settings:
      
      

      • Scope name specifies a name for the scope. Enter the following name:

        session:scope:connect

        This name can be any text. However, a scope name must be provided. You will need to use this scope name later when you give consent to the database client application to access the database.

      • Who can consent specifies the necessary permissions. Select Admins and users, or for higher restrictions, Admins only.
      • Admin consent display name describes the scope's purpose (for example, Connect to Oracle), which only administrators can see.
      • Admin consent display name describes the scope's purpose (for example, Connect to Example Database), which only administrators can see.
      • User consent display name is a short description of the purpose of the scope (for example, Connect to Example Database), which users can see if you specify Admins and users in Who can consent.
      • User consent description is a more detailed description of the purpose of the scope (for example, Connect to Example Database), which users can see if you specify Admins and users in Who can consent.
      • State enables or disables the connection. Select Enabled.

To enable the Microsoft Azure AD v2 access token, you must configure it to use the upn attribute from the Azure portal.

1. Check the version of the Azure AD access token that you are using.
2. Log in to the Microsoft Azure portal.
3. Search for and select Azure Active Directory.
4. Under Manage, select App registrations.
5. Choose the application for which you want to configure optional claims based on your scenario and desired outcome.
6. Under Manage, select Token configuration.
7. Click Add optional claim and select upn.

• Checking the Azure AD Access Token Version
  You can check the version of the Microsoft Azure AD access token that your site uses by using the JSON Web Tokens web site.

In Azure AD, you can create and manage app roles that will be assigned to Azure AD users and groups and also be mapped to Oracle Database global schemas and roles.

• Creating a Microsoft Azure AD App Role
  Azure AD users, groups, and applications will be assigned to the app roles.
• Assigning Users and Groups to the Microsoft Azure AD App Role
  Before Microsoft Azure AD users can have access to the Oracle Database instance, they must first be assigned to the app roles that will be mapped to Oracle Database schema users or roles.
• Assigning an Application to an App Role
  You can assign an Azure AD client application to a app role.

You can enable a Microsoft Azure AD external authentication with Oracle Database.

1. Log in to the Oracle Database instance as a user who has been granted the ALTER SYSTEM system privilege.
2. Set the IDENTITY_PROVIDER_TYPE parameter as follows:

   ALTER SYSTEM SET IDENTITY_PROVIDER_TYPE=AZURE_AD SCOPE=BOTH;

3. Ensure that you set the IDENTITY_PROVIDER_TYPE parameter correctly.

   SELECT NAME, VALUE FROM V$PARAMETER WHERE NAME='identity_provider_type';

   The following output should appear:

   NAME                    VALUE 
   ----------------------  ------- 
   identity_provider_type  AZURE_AD

4. Set the IDENTITY_PROVIDER_CONFIG parameter by using the following syntax:

   ALTER SYSTEM SET IDENTITY_PROVIDER_CONFIG =
   {
      application_id_uri : string , // from registered app, to be mapped in jwt "aud" claim; 
                                    // Domain qualified to support cross tenancy resource access
      tenant_id : string,           // from tenant config
      app_id: string                // from registered resource app
   }SCOPE=BOTH;

   For example:

   ALTER SYSTEM SET IDENTITY_PROVIDER_CONFIG =
   {
     "application_id_uri" : "https://www.example.com/11aa1a11-aaaa-1111-1111-1111aa11111",
     "tenant_id" : 111a1111-a11a-111a-1a1a-1111111111a,
     "app_id" : 11aa1a11-aaaa-1111-1111-1111aa11111
   }SCOPE=BOTH;

To disable Azure AD External authentication for an Oracle Database instance, you must set parameters with the ALTER SYSTEM statement.

1. Log in to the Oracle Database instance as a user who has been granted the ALTER SYSTEM system privilege.
2. Set the identity provider parameters as follows:

   ALTER SYSTEM RESET IDENTITY_PROVIDER_CONFIG SCOPE=BOTH;
   ALTER SYSTEM RESET IDENTITY_PROVIDER_TYPE SCOPE=BOTH;

You can exclusively map an Oracle Database schema to a Microsoft Azure AD user.

1. Log in to the Oracle Database instance as a user who has been granted the CREATE USER or ALTER USER system privilege.
2. Run the CREATE USER or ALTER USER statement with the IDENTIFIED GLOBALLY AS clause specifying the Azure AD user name.
   For example, to create a new database schema user named peter_fitch and map this user to an existing Azure AD user named peter.fitch@example.com:

   CREATE USER peter_fitch IDENTIFIED GLOBALLY AS 
   'AZURE_USER=peter.fitch@example.com';

3. Grant the CREATE SESSION privilege to the user.

   GRANT CREATE SESSION TO peter_fitch;

In this mapping, an Oracle schema is mapped to an app role.

1. Log in to the Oracle Database instance as a user who has the CREATE USER or ALTER USER system privilege.
2. Run the CREATE USER or ALTER USER statement with the IDENTIFIED GLOBALLY AS clause specifying the Azure application role name.
   For example, to create a new database global user account (schema) named dba_azure and map it to an existing Azure AD application role named AZURE_DBA:

   CREATE USER dba_azure IDENTIFIED GLOBALLY AS 'AZURE_ROLE=AZURE_DBA';

Oracle Database global roles that are mapped to Azure app roles give Azure users and applications additional privileges and roles above those that they have been granted through their login schemas.

1. Log in to the Oracle Database instance as a user who has been granted the CREATE ROLE or ALTER ROLE system privilege
2. Run the CREATE ROLE or ALTER ROLE statement with the IDENTIFIED GLOBALLY AS clause specifying the name of the Azure AD application role.
   For example, to create a new database global role named widget_sales_role and map it to an existing Azure AD application role named WidgetManagerGroup:

   CREATE ROLE widget_sales_role IDENTIFIED GLOBALLY AS 
   'AZURE_ROLE=WidgetManagerGroup';

There are numerous ways that you can configure a client to connect with an Oracle Database instance using Azure AD tokens.

You should choose the client connection method that works best with your environment. This guide provides examples of connecting SQL*Plus with different methods of getting an Azure AD OAuth2 access token. All Oracle Database clients version 19.16 and above can accept a token that is passed as a file. The JDBC-thin, Instant Client, and ODP.net drivers also accept the token through the database client API from an application. Oracle Database tools such as SQL*Plus cannot retrieve the tokens directly, so tools such as PowerShell or Azure CLI must be used to retrieve the Azure AD OAuth2 access token. To retrieve an Azure AD token, the client must be registered through the Azure AD app registration process. Registering the client is similar to registering the Oracle Database server with Azure AD using app registration. Both the database and client must be registered with Azure AD.

The database must be registered so the client can get permission to get an access token for the database. The client must be registered so that Azure AD can recognize a trusted client is asking for an access token.

See the following Microsoft Azure articles for more information about connecting clients to Azure AD:

• Quickstart: Configure a client application to access a web API
• Choose the right Azure command-line tool
• Get Azure AD tokens by using the Microsoft Authentication Library
• Install the Azure CLI on Linux

Oracle Database supports several types of client drivers for Azure AD connections.

• JDBC-thin: Oracle Database 19.16 (July 2022), Oracle Database 21.8 (October 2022)
• OCI (C driver): Oracle Database 19.16 (July 2022)
• Oracle Instant Client based on OCI
• Oracle Data Provider (core): Oracle Database 19.16, Oracle Database 21.7
• Oracle Data Provider (unmanaged): based on OCI
• Oracle Data Provider (managed): Oracle Database 19.16, Oracle Database 21.7
• All other drivers built on OCI adopts the OCI compatibility

The connection between the Azure user, Azure AD, and the Oracle Database instance relies on the passing of the OAuth2 token throughout these components.

This example shows the use of the Resource Owner Password Credential (ROPC) flow with a public client. See the Microsoft Azure article Microsoft identity platform and OAuth 2.0 Resource Owner Password Credentials for detailed information about ROPC.

Figure 5-2 ROPC Operational Flow with a Public Client

1. The Azure user requests an Azure AD access token for the database in PowerShell and the returned token is written into a file called token at a file location.
2. The Azure user connects to the database using / slash login. Either the sqlnet.ora or tnsnames.ora connection string tells the instant client that an Azure AD OAuth2 token is needed and to retrieve it from a specified file location. The access token is sent to the database.
3. The database verifies that the access token came from Azure AD (using the Azure AD public key) and then checks the token for additional claims.
4. The database finds the schema mapping (exclusive or shared) and creates the session. The database will also grant any global roles that the Azure user is also assigned to through an app role.

This type of registration is similar to registering Oracle Database with Azure AD app registration.

• Confidential and Public Client Registration
  You can register the database client with Azure as either confidential or public depending on your use case.
• Registering a Database Client App with Azure AD
  Creating the client app registration is similar to creating the Oracle Database instance with the Microsoft Azure AD tenancy.

These examples show different ways that you can retrieve Azure AD OAuth2 tokens.

• Example: Using PowerShell to Get a Token Using Resource Owner Password Credentials
  This example shows how to use PowerShell to get an Azure AD access token by using Resource Owner Password Credentials (ROPC).
• Example: Using Python with Microsoft Authentication Library Using an Authorization Flow
  Because this example with the Microsoft Authentication Library (MSAL) is in Python, it can be run on a variety of platforms such as PowerShell and Linux.
• Example: Using Curl with a Resource Owner Password Credential Flow
  This example shows how to use the curl command against the Azure AD API uses a Resource Owner Password Credential (ROPC) flow with a public Azure AD client.
• Example: Azure CLI Using Authorization Flow
  This example shows how to use the Azure CLI to retrieve an access token and then write the token to a file.

You must configure SQL*Plus to retrieve the Azure AD database access token from a location and use it when the / slash login is used.

1. Ensure that you have an Azure AD user account.
2. Check with an Azure AD administrator or Oracle Database administrator for one of the following:
   • An application client ID that you can use to get Azure AD tokens. If you have Azure AD privileges to do so, then create your own client app registration, similar to registering the Oracle Database instance with an Azure AD tenancy.
   • You are mapped to a global schema in the database.
3. Ensure that you are using the latest release updates for the Oracle Database client releases 19c.
   This configuration only works with the Oracle Database client release 19.16 and above. Oracle Database release 21c clients do not support the full range of features.
4. Configure the database server and client to use a TLS connection with the ExaDB-D database server.
5. On the client, set the following parameters in the sqlnet.ora file:
   • Check for the parameter SSL_SERVER_DN_MATCH = ON to ensure that DN matching is enabled.
   • Set the TOKEN_AUTH parameter to enable the client to use the Azure AD token. Include the TOKEN_LOCATION parameter to point to the token location. For example:

     TOKEN_AUTH=OAUTH 
     TOKEN_LOCATION="token_location" 

     Note that there is no default location. If the token is named token, then you only need to specify the file directory (for example, /test/oracle/aad-token). If the token name is different from token (for example, azure.token), then you must include this name in the path (for example, /test/oracle/aad-token/azure.token).

(description= 
  (retry_count=20)(retry_delay=3)
  (address=(protocol=tcps)(port=1522)
  (host=example.us-phoenix-1.oraclecloud.com))
  (connect_data=(service_name=aaabbbccc_exampledb_high.example.oraclecloud.com))
  (security=(ssl_server_cert_dn="CN=example.uscom-east-1.oraclecloud.com, 
     OU=Oracle BMCS US, O=Example Corporation, 
     L=Redwood City, ST=California, C=US")
  (TOKEN_AUTH=OAUTH)(TOKEN_LOCATION="/test/oracle/aad-token"))

connect /@exampledb_high

connect /@(description= 
  (retry_count=20)(retry_delay=3)
  (address=(protocol=tcps)(port=1522)
  (host=example.us-phoenix-1.oraclecloud.com))
  (connect_data=(service_name=aaabbbccc_exampledb_high.example.oraclecloud.com))
  (security=(ssl_server_cert_dn="CN=example.uscom-east-1.oraclecloud.com, 
     OU=Oracle BMCS US, O=Example Corporation, 
     L=Redwood City, ST=California, C=US") (TOKEN_AUTH=OAUTH)(TOKEN_LOCATION="/test/oracle/aad-token")

You can generate two levels of trace files to troubleshoot Microsoft Azure AD connections on client side.

The two levels of trace files that you can generate are as follows:

• Low level tracing prints traces in case of failures:
  • If TCPS is not set up for the Azure AD connection, then it prints a message that the protocol has to be TCPS.
  • If SSL_SERVER_DN_MATCH is not set to TRUE, then it prints a message that the value is FALSE.
  • If TOKEN_LOCATION has not been specified, then it prints a message that the token location does not exist.
  • If the token is not present at the specified TOKEN_LOCATION, then it prints a message.
  • If the application has passed in the token without setting OCI_ATTR_TOKEN_ISBEARER to true, it prints a message for the missing attribute.
  • If the application has set OCI_ATTR_TOKEN_ISBEARER to TRUE and not passed in the token, it prints a message for the missing attribute.
  • If the token has expired, then it prints a message.
• High level tracing prints traces in case of failure as mentioned above. In addition, it prints traces in case of success, as follows:
  • It prints where SSL_SERVER_DN_MATCH is present, tnsnames.ora or sqlnet.ora. It also prints the value as TRUE if set to TRUE.
  • If both the token and OCI_ATTR_TOKEN_ISBEARER=true are set by the application, then it prints a message.
  • If TOKEN_AUTH has the correct value OAUTH, then it prints the value.
  • If the token is not expired, then it prints a message.

You can add EVENT settings to the client-side sqlnet.ora file to control client tracing.

• Use either of the following methods:
  • Add the following settings to the client side sqlnet.ora file:
    • EVENT_25701=14 for low level tracing
    • EVENT_25701=15 for high level tracing
  • Set the environment variable EVENT_25701:
    • EVENT_25701=14 for low level tracing
    • EVENT_25701=15 for high level tracing
  Client trace files are created in the following locations:

  • Linux: $ORACLE_HOME/log/diag/clients
  • Windows: %ORACLE_HOME%\log\diag\clients

  You can use the ADR_BASE parameter in the client side sqlnet.ora to specify the directory in which tracing messages are stored. Ensure that the directory path is valid and has write permissions. Ensure that the DIAG_ADR_ENABLED parameter is not set to FALSE.

  An example of setting ADR_BASE is as follows:

  ADR_BASE=/oracle/oauth2/trace