1 Introduction to Oracle Label Security

Oracle Label Security enables you to control the display of individual table rows using labels that are assigned to individual table rows and application users.

Oracle Label Security is based on multi-level security (MLS) requirements that are found in government and defense organizations. You can easily restrict sensitive information to only authorized users. Oracle Label Security is based on Oracle Virtual Private Database (VPD). However, unlike VPD, Oracle Label Security provides the access mediation functions, data dictionary tables, and policy based architecture out of the box, eliminating customized coding and providing a consistent label based access control model that can be used by multiple applications. Oracle Label Security policies can be applied to one or more application tables. Oracle Label Security works by comparing the row label with a user's label authorizations. Oracle Label Security software is installed by default, but not automatically enabled. You can easily enable Oracle Label Security in either SQL*Plus or by using the Oracle Database Configuration Assistant (DBCA). To manage Oracle Label Security, you can use either Oracle Enterprise Manager Cloud Control or a set of PL/SQL packages and standalone functions at the command line level. The default administrator for Oracle Label Security is the user LBACSYS. To find information about Oracle Label Security policies, you can query ALL_SA_*, DBA_SA_*, or USER_SA_* data dictionary views.

Oracle Label Security provides several benefits.

• It enables row level data classification and provides out of the box access mediation based on the data classification and the user label authorization or security clearance.

• It enables you to assign label authorizations or security clearances to both database users and application users.

• It provides both a graphical user interface and APIs for defining and storing data classification labels and user label authorizations.

• It integrates with Oracle Database Vault and Oracle Advanced Security Data Redaction, enabling security clearances to be use in both Database Vault command rules and Data Redaction policy definitions.

When you register Oracle Label Security with a database, the registration process creates an administrative user named LBACSYS, who has the LBAC_DBA database role.

You can grant this role to any database user who will be responsible for managing Oracle Label Security policies. In addition, you can grant Oracle Label Security administrators the EXECUTE privilege for the Oracle Label Security packages, and privileges to manage individual Oracle Label Security policies.

You can manage the administration of an Oracle Label Security policy by using several privileges.

These privileges are:

• Package-specific privileges: Most of the Oracle Label Security PL/SQL packages, except for the public SA_SESSION and SA_UTL packages, require the EXECUTE privilege.

• Role-based privileges: The Oracle Label Security-specific roles are:

  • The policy_DBA role, which is created and granted to the user when he or she creates a policy. For example, for a policy named ols_hr_pol, the role created is named ols_hr_pol_DBA. This role adds a layer of granularity for access control for your site's Oracle Label Security policies.

  • The LBAC_DBA role, which provides the EXECUTE privilege for the LBAC_SYSDBA package. This role is owned by the LBACSYS user account. The LBAC_SYSDBA package enables the user to create, alter, enable, disable, and drop Oracle Label Security policies. This package is a wrapper for the SA_SYSDBA package.

You can use the Oracle Label Security package EXECUTE privilege grants along with grants of the policy_DBA role to achieve additional separation of duty. The packages are categorized based on different tasks. For example, you could grant the EXECUTE privilege on the SA_COMPONENTS and SA_LABEL_ADMIN packages to one user or role to manage label definitions, and then grant EXECUTE on SA_USER_ADMIN to a different user or role to manage user labels and privileges. Both of these users or roles must also be granted the policy_DBA role for the policies for which they are responsible. In this way, different users can be responsible for the management of different aspects of the policies for which they are responsible. For example, user psmith could be responsible for the label definitions of the ols_hr_pol policy, and user tjones could be responsible for the label definitions of the ols_oe_pol policy. However, user psmith cannot modify label definitions for the ols_oe_pol policy, nor can tjones modify the ols_hr_pol policy label definitions.

An Oracle Label Security policy has a standard set of components.

These components are as follows:

• Labels. Labels for data and users, along with authorizations for users and program units, govern access to specified protected objects. Labels are composed of the following:

  • Levels. Levels indicate the type of sensitivity that you want to assign to the row, for example, SENSITIVE or HIGHLY SENSITIVE.

  • Compartments. (Optional) Data can have the same level (Public, Confidential and Secret), but can belong to different projects inside a company, for example ACME Merger and IT Security. Compartments represent the projects in this example, that help define more precise access controls. They are most often used in government environments.

  • Groups. (Optional) Groups identify organizations owning or accessing the data, for example, UK, US, Asia, Europe. Groups are used both in commercial and government environments, and frequently used in place of compartments due to their flexibility.

• Policy. A policy is a name associated with these labels, rules, authorizations, and protected tables.

For example, assume that a user has the SELECT privilege on an application table. As illustrated in Figure 1-1, when the user runs a SELECT statement, Oracle Label Security evaluates each row selected to determine whether the user can access using the privileges and labels assigned to the user and the label on the row. You can configure Oracle Label Security to perform security checks on UPDATE, DELETE, and INSERT statements as well.

Figure 1-1 Oracle Label Security Label-Based Security

Description of "Figure 1-1 Oracle Label Security Label-Based Security"

Figure 1-2 shows how data is accessed under Oracle Label Security and the sequence of label security checks.

Figure 1-2 Oracle Label Security Architecture

Description of "Figure 1-2 Oracle Label Security Architecture"

In this scenario, the following actions take place:

1. An application user in an Oracle Database session sends a SQL request to query a table.

2. Oracle Database checks the user's data access control (DAC) privileges for performing a SELECT statement on the table.

3. If the user does have the appropriate privileges, then Oracle Database checks if there are any Oracle Virtual Private Database (VPD) policies attached to the table.

4. Oracle Database then checks if there are any Oracle Label Security policies that are assigned to the table.

5. Oracle Label Security then compares the labels that are assigned to individual rows with the users' label authorizations, allowing or denying access. The session label is based on label authorizations that are assigned to the user.

You can perform Oracle Label Security development and administrative tasks using either of two interfaces.

Topics:

• Oracle Label Security Packages

• Oracle Enterprise Manager Cloud Control

You can integrate Oracle Label Security with Oracle Internet Directory (OID) and in a multitenant environment.

Topics:

• Oracle Label Security Integration with Oracle Internet Directory

• Oracle Label Security Integration in a Multitenant Environment