Skip to content

Latest commit

 

History

History
645 lines (498 loc) · 26.5 KB

Authentication.md

File metadata and controls

645 lines (498 loc) · 26.5 KB
Raw

Authentication

The python-sdk-core project supports the following types of authentication:

  • Basic Authentication
  • Bearer Token Authentication
  • Identity and Access Management (IAM) Authentication (grant type: apikey)
  • Identity and Access Management (IAM) Authentication (grant type: assume)
  • Container Authentication
  • VPC Instance Authentication
  • Cloud Pak for Data Authentication
  • Multi-Cloud Saas Platform (MCSP) Authentication
  • No Authentication (for testing)

The SDK user configures the appropriate type of authentication for use with service instances. The authentication types that are appropriate for a particular service may vary from service to service, so it is important for the SDK user to consult with the appropriate service documentation to understand which authenticators are supported for that service.

The python-sdk-core allows an authenticator to be specified in one of two ways:

  1. programmatically - the SDK user invokes the appropriate function(s) to create an instance of the desired authenticator and then passes the authenticator instance when constructing an instance of the service client.
  2. configuration - the SDK user provides external configuration information (in the form of environment variables or a credentials file) to indicate the type of authenticator, along with the configuration of the necessary properties for that authenticator. The SDK user then invokes the configuration-based service client constructor method to construct an instance of the authenticator and service client that reflect the external configuration information.

The sections below will provide detailed information for each authenticator which will include the following:

  • A description of the authenticator
  • The properties associated with the authenticator
  • An example of how to construct the authenticator programmatically
  • An example of how to configure the authenticator through the use of external configuration information. The configuration examples below will use environment variables, although the same properties could be specified in a credentials file instead.

Basic Authentication

The BasicAuthenticator is used to add Basic Authentication information to each outbound request in the Authorization header in the form:

   Authorization: Basic <encoded username and password>

Properties

  • username: (required) the basic auth username

  • password: (required) the basic auth password

Programming example

from ibm_cloud_sdk_core.authenticators import BasicAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator.
authenticator = BasicAuthenticator(<your_username>, <your_password>)

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.

Configuration example

External configuration:

export EXAMPLE_SERVICE_AUTH_TYPE=basic
export EXAMPLE_SERVICE_USERNAME=myuser
export EXAMPLE_SERVICE_PASSWORD=mypassword

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.

Bearer Token Authentication

The BearerTokenAuthenticator will add a user-supplied bearer token to each outbound request in the Authorization header in the form:

    Authorization: Bearer <bearer-token>

Properties

  • bearerToken: (required) the bearer token value

Programming example

from ibm_cloud_sdk_core.authenticators import BearerTokenAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator.
authenticator = BearerTokenAuthenticator(<your_bearer_token>)

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.
...
# Later, if your bearer token value expires, you can set a new one like this:
new_token = '<new token>'
authenticator.set_bearer_token(new_token);

Configuration example

External configuration:

export EXAMPLE_SERVICE_AUTH_TYPE=bearertoken
export EXAMPLE_SERVICE_BEARER_TOKEN=<the bearer token value>

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.

Note that the use of external configuration is not as useful with the BearerTokenAuthenticator as it is for other authenticator types because bearer tokens typically need to be obtained and refreshed programmatically since they normally have a relatively short lifespan before they expire. This authenticator type is intended for situations in which the application will be managing the bearer token itself in terms of initial acquisition and refreshing as needed.

Identity and Access Management (IAM) Authentication (grant type: apikey)

The IamAuthenticator will accept a user-supplied apikey and will perform the necessary interactions with the IAM token service to obtain a suitable bearer token for the specified apikey. The authenticator will also obtain a new bearer token when the current token expires. The bearer token is then added to each outbound request in the Authorization header in the form:

   Authorization: Bearer <bearer-token>

Properties

  • apikey: (required) the IAM api key to be used to obtain an IAM access token.

  • url: (optional) The base endpoint URL of the IAM token service. The default value of this property is the "prod" IAM token service endpoint (https://iam.cloud.ibm.com). Make sure that you use an IAM token service endpoint that is appropriate for the location of the service being used by your application. For example, if you are using an instance of a service in the "production" environment (e.g. https://resource-controller.cloud.ibm.com), then the default "prod" IAM token service endpoint should suffice. However, if your application is using an instance of a service in the "staging" environment (e.g. https://resource-controller.test.cloud.ibm.com), then you would also need to configure the authenticator to use the IAM token service "staging" endpoint as well (https://iam.test.cloud.ibm.com).

  • client_id/client_secret: (optional) The client_id and client_secret fields are used to form a "basic auth" Authorization header for interactions with the IAM token server. If neither field is specified, then no Authorization header will be sent with token server requests. These fields are optional, but must be specified together.

  • scope: (optional) the scope to be associated with the IAM access token. If not specified, then no scope will be associated with the access token.

  • disable_ssl_verification: (optional) A flag that indicates whether verification of the server's SSL certificate should be disabled or not. The default value is false.

  • headers: (optional) A set of key/value pairs that will be sent as HTTP headers in requests made to the IAM token service.

Programming example

from ibm_cloud_sdk_core.authenticators import IAMAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator.
authenticator = IAMAuthenticator('myapikey')

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.

Configuration example

External configuration:

export EXAMPLE_SERVICE_AUTH_TYPE=iam
export EXAMPLE_SERVICE_APIKEY=myapikey

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.

Identity and Access Management (IAM) Authentication (grant type: assume)

The IAMAssumeAuthenticator performs a two-step token fetch sequence to obtain a bearer token that allows the application to assume the identity of a trusted profile:

  1. First, the authenticator obtains an initial bearer token using grant type urn:ibm:params:oauth:grant-type:apikey. This initial token will reflect the identity associated with the input apikey.
  2. Second, the authenticator uses the grant type urn:ibm:params:oauth:grant-type:assume to obtain a bearer token that reflects the identity of the trusted profile, passing in the initial bearer token from the first step, along with the trusted profile-related inputs.

The authenticator will also obtain a new bearer token when the current token expires. The bearer token is then added to each outbound request in the Authorization header in the form:

   Authorization: Bearer <bearer-token>

Properties

  • apikey: (required) the IAM apikey to be used to obtain the initial IAM access token.

  • iam_profile_crn: (optional) the Cloud Resource Name (CRN) associated with the trusted profile for which an access token should be fetched. Exactly one of iam_profile_crn, iam_profile_id or iam_profile_name must be specified.

  • iam_profile_id: (optional) the ID associated with the trusted profile for which an access token should be fetched. Exactly one of iam_profile_crn, iam_profile_id or iam_profile_name must be specified.

  • iam_profile_name: (optional) the name associated with the trusted profile for which an access token should be fetched. When specifying this property, you must also specify the iam_account_id property as well. Exactly one of iam_profile_crn, iam_profile_id or iam_profile_name must be specified.

  • iam_account_id: (optional) the ID associated with the IAM account that contains the trusted profile referenced by the iam_profile_name property. The iam_account_id property must be specified if and only if the iam_profile_name property is specified.

  • url: (optional) The base endpoint URL of the IAM token service. The default value of this property is the "prod" IAM token service endpoint (https://iam.cloud.ibm.com). Make sure that you use an IAM token service endpoint that is appropriate for the location of the service being used by your application. For example, if you are using an instance of a service in the "production" environment (e.g. https://resource-controller.cloud.ibm.com), then the default "prod" IAM token service endpoint should suffice. However, if your application is using an instance of a service in the "staging" environment (e.g. https://resource-controller.test.cloud.ibm.com), then you would also need to configure the authenticator to use the IAM token service "staging" endpoint as well (https://iam.test.cloud.ibm.com).

  • client_id/client_secret: (optional) The client_id and client_secret fields are used to form a "basic auth" Authorization header for interactions with the IAM token server when fetching the initial IAM access token. These fields are optional, but must be specified together.

  • scope: (optional) the scope to be used when obtaining the initial IAM access token. If not specified, then no scope will be associated with the access token.

  • disable_ssl_verification: (optional) A flag that indicates whether verification of the server's SSL certificate should be disabled or not. The default value is false.

  • headers: (optional) A set of key/value pairs that will be sent as HTTP headers in requests made to the IAM token service.

Usage Notes

  • The IAMAssumeAuthenticator is used to obtain an access token (a bearer token) from the IAM token service that allows an application to "assume" the identity of a trusted profile.

  • The authenticator first uses the apikey, url, client_id/client_secret, scope, disable_ssl_verification, and headers properties to obtain an initial access token by invoking the IAM get_token (grant_type=urn:ibm:params:oauth:grant-type:apikey) operation.

  • The authenticator then uses the initial access token along with the url, iam_profile_crn, iam_profile_id, iam_profile_name, iam_account_id, disable_ssl_verification, and headers properties to obtain an access token by invoking the IAM get_token (grant_type=urn:ibm:params:oauth:grant-type:assume) operation. The access token resulting from this second step will reflect the identity of the specified trusted profile.

  • When providing the trusted profile information, you must specify exactly one of: iam_profile_crn, iam_profile_id or iam_profile_name. If you specify iam_profile_crn or iam_profile_id, then the trusted profile must exist in the same account that is associated with the input apikey. If you specify iam_profile_name, then you must also specify the iam_account_id property to indicate the IAM account in which the named trusted profile can be found.

Programming example

from ibm_cloud_sdk_core.authenticators import IAMAssumeAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator.
authenticator = IAMAssumeAuthenticator('myapikey', iam_profile_id='my_profile_id')

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.

Configuration example

External configuration:

export EXAMPLE_SERVICE_AUTH_TYPE=iamAssume
export EXAMPLE_SERVICE_APIKEY=myapikey
export EXAMPLE_SERVICE_IAM_PROFILE_ID=myprofile-1

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.

Container Authentication

The ContainerAuthenticator is intended to be used by application code running inside a compute resource managed by the IBM Kubernetes Service (IKS) in which a secure compute resource token (CR token) has been stored in a file within the compute resource's local file system. The CR token is similar to an IAM apikey except that it is managed automatically by the compute resource provider (IKS). This allows the application developer to:

  • avoid storing credentials in application code, configuration files or a password vault
  • avoid managing or rotating credentials

The ContainerAuthenticator will retrieve the CR token from the compute resource in which the application is running, and will then perform the necessary interactions with the IAM token service to obtain an IAM access token using the IAM "get token" operation with grant-type cr-token. The authenticator will repeat these steps to obtain a new IAM access token when the current access token expires. The IAM access token is added to each outbound request in the Authorization header in the form:

   Authorization: Bearer <IAM-access-token>

Properties

  • cr_token_filename: (optional) The name of the file containing the injected CR token value. If not specified, then the authenticator will first try /var/run/secrets/tokens/vault-token and then /var/run/secrets/tokens/sa-token as the default value (first file found is used). The application must have read permissions on the file containing the CR token value.

  • iam_profile_name: (optional) The name of the linked trusted IAM profile to be used when obtaining the IAM access token (a CR token might map to multiple IAM profiles). One of iam_profile_name or iam_profile_id must be specified.

  • iam_profile_id: (optional) The ID of the linked trusted IAM profile to be used when obtaining the IAM access token (a CR token might map to multiple IAM profiles). One of iam_profile_name or iam_profile_id must be specified.

  • url: (optional) The base endpoint URL of the IAM token service. The default value of this property is the "prod" IAM token service endpoint (https://iam.cloud.ibm.com). Make sure that you use an IAM token service endpoint that is appropriate for the location of the service being used by your application. For example, if you are using an instance of a service in the "production" environment (e.g. https://resource-controller.cloud.ibm.com), then the default "prod" IAM token service endpoint should suffice. However, if your application is using an instance of a service in the "staging" environment (e.g. https://resource-controller.test.cloud.ibm.com), then you would also need to configure the authenticator to use the IAM token service "staging" endpoint as well (https://iam.test.cloud.ibm.com).

  • client_id/client_secret: (optional) The client_id and client_secret fields are used to form a "basic auth" Authorization header for interactions with the IAM token server. If neither field is specified, then no Authorization header will be sent with token server requests. These fields are optional, but must be specified together.

  • scope (optional): the scope to be associated with the IAM access token. If not specified, then no scope will be associated with the access token.

  • disable_ssl_verification: (optional) A flag that indicates whether verification of the server's SSL certificate should be disabled or not. The default value is False.

  • proxies (optional): The proxy endpoint to use for HTTP(S) requests.

  • headers: (optional) A set of key/value pairs that will be sent as HTTP headers in requests made to the IAM token service.

Programming example

from ibm_cloud_sdk_core.authenticators import ContainerAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator.
authenticator = ContainerAuthenticator(iam_profile_name='iam-user-123')

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.

Configuration example

External configuration:

export EXAMPLE_SERVICE_AUTH_TYPE=container
export EXAMPLE_SERVICE_IAM_PROFILE_NAME=iam-user-123

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.

VPC Instance Authentication

The VPCInstanceAuthenticator is intended to be used by application code running inside a VPC-managed compute resource (virtual server instance) that has been configured to use the "compute resource identity" feature. The compute resource identity feature allows you to assign a trusted IAM profile to the compute resource as its "identity". This, in turn, allows applications running within the compute resource to take on this identity when interacting with IAM-secured IBM Cloud services. This results in a simplified security model that allows the application developer to:

  • avoid storing credentials in application code, configuration files or a password vault
  • avoid managing or rotating credentials

The VPCInstanceAuthenticator will invoke the appropriate operations on the compute resource's locally-available VPC Instance Metadata Service to (1) retrieve an instance identity token and then (2) exchange that instance identity token for an IAM access token. The authenticator will repeat these steps to obtain a new IAM access token whenever the current access token expires. The IAM access token is added to each outbound request in the Authorization header in the form:

   Authorization: Bearer <IAM-access-token>

Properties

  • iam_profile_crn: (optional) the crn of the linked trusted IAM profile to be used when obtaining the IAM access token.

  • iam_profile_id: (optional) the id of the linked trusted IAM profile to be used when obtaining the IAM access token.

  • url: (optional) The VPC Instance Metadata Service's base URL.
    The default value of this property is http://169.254.169.254. However, if the VPC Instance Metadata Service is configured with the HTTP Secure Protocol setting (https), then you should configure this property to be https://api.metadata.cloud.ibm.com.

Usage Notes:

  1. At most one of iam_profile_crn or iam_profile_id may be specified. The specified value must map to a trusted IAM profile that has been linked to the compute resource (virtual server instance).

  2. If both iam_profile_crn and iam_profile_id are specified, then an error occurs.

  3. If neither iam_profile_crn nor iam_profile_id are specified, then the default trusted profile linked to the compute resource will be used to perform the IAM token exchange. If no default trusted profile is defined for the compute resource, then an error occurs.

Programming example

from ibm_cloud_sdk_core.authenticators import VPCInstanceAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator.
authenticator = VPCInstanceAuthenticator(iam_profile_crn='crn:iam-profile-123')

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.

Configuration example

External configuration:

export EXAMPLE_SERVICE_AUTH_TYPE=vpc
export EXAMPLE_SERVICE_IAM_PROFILE_CRN=crn:iam-profile-123

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.

Cloud Pak for Data

The CloudPakForDataAuthenticator will accept a user-supplied username value, along with either a password or apikey, and will perform the necessary interactions with the Cloud Pak for Data token service to obtain a suitable bearer token. The authenticator will also obtain a new bearer token when the current token expires. The bearer token is then added to each outbound request in the Authorization header in the form:

   Authorization: Bearer <bearer-token>

Properties

  • username: (required) the username used to obtain a bearer token.

  • password: (required if apikey is not specified) the password used to obtain a bearer token.

  • apikey: (required if password is not specified) the apikey used to obtain a bearer token.

  • url: (required) The URL representing the Cloud Pak for Data token service endpoint's base URL string. This value should not include the /v1/authorize path portion.

  • disable_ssl_verification: (optional) A flag that indicates whether verification of the server's SSL certificate should be disabled or not. The default value is false.

  • headers: (optional) A set of key/value pairs that will be sent as HTTP headers in requests made to the Cloud Pak for Data token service.

Programming example

from ibm_cloud_sdk_core.authenticators import CloudPakForDataAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator using username/apikey.
authenticator = CloudPakForDataAuthenticator(username='myuser', apikey='myapikey', url='https://mycp4dhost.com')

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.

Configuration example

External configuration:

# Configure "example_service" with username/apikey.
export EXAMPLE_SERVICE_AUTH_TYPE=cp4d
export EXAMPLE_SERVICE_USERNAME=myuser
export EXAMPLE_SERVICE_PASSWORD=myapikey
export EXAMPLE_SERVICE_URL=https://mycp4dhost.com

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.

Multi-Cloud Saas Platform (MCSP) Authentication

The MCSPAuthenticator can be used in scenarios where an application needs to interact with an IBM Cloud service that has been deployed to a non-IBM Cloud environment (e.g. AWS). It accepts a user-supplied apikey and performs the necessary interactions with the Multi-Cloud Saas Platform token service to obtain a suitable MCSP access token (a bearer token) for the specified apikey. The authenticator will also obtain a new bearer token when the current token expires. The bearer token is then added to each outbound request in the Authorization header in the form:

   Authorization: Bearer <bearer-token>

Properties

  • apikey: (required) the apikey to be used to obtain an MCSP access token.

  • url: (required) The URL representing the MCSP token service endpoint's base URL string. Do not include the operation path (e.g. /siusermgr/api/1.0/apikeys/token) as part of this property's value.

  • disable_ssl_verification: (optional) A flag that indicates whether verification of the server's SSL certificate should be disabled or not. The default value is false.

  • headers: (optional) A set of key/value pairs that will be sent as HTTP headers in requests made to the MCSP token service.

Usage Notes

  • When constructing an MCSPAuthenticator instance, you must specify the apikey and url properties.

  • The authenticator will use the token server's POST /siusermgr/api/1.0/apikeys/token operation to exchange the user-supplied apikey for an MCSP access token (the bearer token).

Programming example

from ibm_cloud_sdk_core.authenticators import MCSPAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator.
authenticator = MCSPAuthenticator(apikey='myapikey', url='https://example.mcsp.token-exchange.com')

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.

Configuration example

External configuration:

export EXAMPLE_SERVICE_AUTH_TYPE=mcsp
export EXAMPLE_SERVICE_APIKEY=myapikey
export EXAMPLE_SERVICE_AUTH_URL=https://example.mcsp.token-exchange.com

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.

No Auth Authentication

The NoAuthAuthenticator is a placeholder authenticator which performs no actual authentication function. It can be used in situations where authentication needs to be bypassed, perhaps while developing or debugging an application or service.

Properties

None

Programming example

from ibm_cloud_sdk_core.authenticators import NoAuthAuthenticator
from <sdk-package-name>.example_service_v1 import *

# Create the authenticator using username/apikey.
authenticator = NoAuthAuthenticator()

# Construct the service instance.
service = ExampleServiceV1(authenticator=authenticator)

# 'service' can now be used to invoke operations.

Configuration example

External configuration:

export EXAMPLE_SERVICE_AUTH_TYPE=noauth

Application code:

from <sdk-package-name>.example_service_v1 import *

# Construct the service instance.
service = ExampleServiceV1.new_instance(service_name='example_service')

# 'service' can now be used to invoke operations.