Goals

1. Authorize a subset of operations on CDAP entities using Apache Sentry

2. Make the authorization system pluggable. Support the following two systems to begin with:

   1. Sentry based

   2. CDAP Dataset based

Checklist

• User stories documented (Rohit/Bhooshan)
• User stories reviewed (Nitin)
• Design documented (Rohit/Bhooshan)
• Design reviewed (Andreas)
• Feature merged (Rohit/Bhooshan)
• Examples and guides (Rohit)
• Integration tests (Bhooshan) 
• Documentation for feature (Rohit/Bhooshan)
• Blog post 

User Stories

• As a CDAP system, I should be able to integrate with Apache Sentry for fine-grained role-based access controls of select CDAP operations 
• As a CDAP admin, I should be able to create/update/delete roles in Apache Sentry
• As a CDAP admin, I should be able to add users/groups to roles in Apache Sentry
• As a CDAP admin, I should be able to turn authorization on/off easily
• As a CDAP system, I should be able to authorize the following requests
• Namespace create/update/delete
• Application deployment
• Program start/stop
• Stream read/write
  These operations are a subset that represents the various 'kinds' of operations allowed in CDAP

Scenarios

Scenario #1

• D-Rock is an IT-Admin extra-ordinaire who has just been tasked with adding authorizing access to entities in CDAP on the cluster he manages. 
• D-Rock is already familiar with Apache Sentry, since he has used it for authorization in other projects like Apache HDFS, Apache Hive, Apache Sqoop, etc. 
• He would rather not learn a new authorization system. He would instead prefer that Apache Sentry be used to provide Role Based Access Control to CDAP entities as well.

Scenario #2

• D-Rock manages a variety of CDAP clusters in dev/smoke/qa/staging environments along with the prod environment.
• For these environments, he would like to be able to turn authorization on/off easily with a switch, depending on the need at a given time.

Scenario #3

• Ideally, D-Rock would like to be able to authorize all operations on all entities in CDAP. 
• However, this can be rolled out in phases. In the initial phase, he would like to control who can:
  • Create/update/delete a namespace
    • Only users with WRITE permission on CDAP instance should be able to perform this operation.
    • A property in cdap-site.xml should decide a set of users who have admin permission on cdap instance. These admins can then later grant permissions to other users.
  • Deploy an application in a namespace
    • Only users with WRITE permission on the namespace should be able to perform this operation
    • One the application is deployed the the user who deployed becomes the owner
  • Start/stop a program
    • Only users with READ permission on the namespace and application, and EXECUTE permission on the program should be able to perform this operation
    • Only users with ADMIN permission on the program can set preference for the program
    • Only users with WRITE permission can provide runtime args
  • Read/write to a stream
    • Only users with READ privilege on the namespace and READ permission on the stream should be able to read from the stream
    • Only users with READ privilege on the namespace and WRITE permission on the stream should be able to write to the stream

Design

This feature can be broken down into the following main parts, in no specific order:

Authorization Hooks in CDAP

This would include the authorization system in CDAP. External systems like Apache Sentry/Ranger could be plugged into this system. It provides authorization hooks during various operations within CDAP, that throw AuthorizationException if the operation is not authorized.

This system exposes a set of interfaces defined below. 

AuthChecker

The AuthChecker interface provides a way to check if an operation is authorized. At various points in the CDAP code (NamespaceHttpHandler, AppLifecycleHttpHandler, ProgramLifecycleHttpHandler, StreamHandler in 3.4), this interface will be used to check if an operation is authorized.

interface AuthChecker {
	/**
     * Checks if a user is allowed to perform a set of actions on an entity.
     *
     * @param Principal the Principal that performs the actions. This could be a user, group or a role
     * @param entity the entity on which an action is being performed
     * @param action the action being performed
     * @throws AuthorizationException if the Principal is not authorized to perform action on the entity
     */
	void checkAuthorized(Principal Principal, Entity entity, Action action) throws AuthorizationException;
}

Authorizer

This interface allows CDAP admins to grant/revoke permissions for specific operations on specific CDAP entities to specified Principals. It will be used by the ACL Management module, which may or may not reside in CDAP for the purposes of integration with Apache Sentry  TBD.

interface Authorizer extends AuthChecker {
	/**
     * Grants a principal authorization to perform a set of actions on an entity.
     *
     * @param entity the entity on which an action is being performed
     * @param principal the Principal that performs the actions. This could be a user, group or a role
     * @param actions the set of actions to grant
     */
    void grant(EntityId entity, Principal principal, Set<Action> actions);

	/**
     * Grants a Principal authorization to perform all actions on an entity.
     *
     * @param entity the entity on which an action is being performed
     * @param principal the Principal that performs the actions. This could be a user, group or a role
     */
    void grant(EntityId entity, Principal principal, Set<Action> actions);
	/**
     * Revokes a principal's authorization to perform a set of actions on an entity.
     *
     * @param entity the entity on which an action is being performed
     * @param principal the principal that performs the actions. This could be a user, group or a role
     * @param actions the set of actions to revoke permissions on
     */
    void revoke(EntityId entity, Principal principal, Set<Action> actions);

	/**
     * Revokes a principal's authorization to perform any action on an entity.
     *
     * @param entity the entity on which an action is being performed
     * @param principal the principal that performs the actions. This could be a user, group or a role
     */
    void revoke(EntityId entity, Principal Principal);

    /**
     * Revokes all principals' authorization to perform any action on an entity.
     *
     * @param entity the entity on which an action is being performed
     */
    void revoke(EntityId entity);
}

Where Principal is the entity performing actions defined as below:

public class Principal {
	enum PrincipalType {
		USER,
		GROUP,
		ROLE
	}
 
	private final String name;
	private final PrincipalType type;
 
	public Principal(String name, PrincipalType type) {
		this.name = name;
		this.type = type;
	}
 
	public String getName() {
		return name;
	}
 
	public PrincipalType getType() {
		return type;
	}
}

Integration with Apache Sentry will be achieved by implementations of these interfaces that delegate to Apache Sentry.

Integration with Apache Sentry

Integration with Apache Sentry involves the development of three main modules:

CDAP Sentry Binding

Here we will bind CDAP to SentryGenericServiceClient and to the operations on the client.

public class SentryAuthorizer implements Authorizer {

    void grant(EntityId entity, Principal Principal, Set<Action> actions){
		// do grant operation on sentry client with needed mapping/conversion
	}

    void grant(EntityId entity, Principal Principal, Set<Action> actions){
		// do grant operation on sentry client with needed mapping/conversion
	}

    void revoke(EntityId entity, Principal Principal, Set<Action> actions){
		// do grant operation on sentry client with needed mapping/conversion
	}

    void revoke(EntityId entity, Principal Principal){
		// do revoke operation on sentry client with needed mapping/conversion
	}

    void revoke(EntityId entity){
		// do revoke operation on sentry client with needed mapping/conversion
	}
 	void checkAuthorized(Principal Principal, Entity entity, Action action) throws AuthorizationException{
		// do authorization check operation on sentry client with needed mapping/conversion
	}
 
	private SentryGenericServiceClient getClient() throws Exception {
	  return SentryGenericServiceClientFactory.create(conf); // create sentry client from Configuration 
	}
}

CDAP Sentry Model

The CDAP Sentry Model defines the CDAP entities for whom access needs to be authorized via Apache Sentry. It will based off of the Sentry Generic Authorization Model. The CDAP Sentry Model will have the following components:

CDAPAuthorizable

This interface defines the CDAP entities that need to be authorized. It must implement Authorizable.

/**
 * This interface represents an authorizable resource in the CDAP component.
 */
public interface CDAPAuthorizable extends Authorizable {

  public enum AuthorizableType {
	Instance,
    Namespace,
    Artifact,
    Application,
    Program,
    Dataset,
    Stream,
	Stream_View
  };
  AuthorizableType getAuthzType();
}

The CDAPAuthorizable interface will have to be implemented for each authorizable entity defined by the AuthorizableType enum above.

CDAPAction and CDAPActionFactory

These classes will implement BitFieldAction and BitFieldActionFactory to define the types of actions on CDAP entities. These classes also allow you to define implies relationships between actions.

TODO: Think about ALL, ADMIN_ALL

public class CDAPActionConstants {
  public static final String READ = "read";
  public static final String EXECUTE = "execute";
  public static final String WRITE = "write";
  public static final String ADMIN = "admin"; // this is read + write + execute + admin (create/update/delete)
}

Sentry Policy Engine

Resource URIs

Using the above authorizable model, resource URIs for CDAP entities in the Sentry Policy Engine will be as follows:

Note: Will have to whitelist the cdap user for the Sentry Service.

ACL management

• TBD: either using CDAP CLI or via external systems like Sentry CLI or Hue 

Questions

for a given user/group and type of access
	if allowed:
		perform operation
	else:
		throw AuthException

References