2
Understanding Data Labels and User Labels

This chapter discusses the fundamental concepts of data labels and user labels, and introduces the terminology that will help you understand Oracle Label Security.

The chapter includes:

• Introduction to Label-Based Security
• Label Components
• Label Syntax and Type
• How Data Labels and User Labels Work Together
• Administering Labels

Introduction to Label-Based Security

Label-based security provides a flexible way of controlling access to sensitive data. Oracle Label Security controls data access based on the identity and label of the user, and the sensitivity and label of the data. Label security adds protections beyond the discretionary access controls that determine the operations users can perform upon data in an object, such as a table or view.

An Oracle Label Security policy controls access to data in three dimensions:

Note that the discussion here concerns access to data. The particular type of access, such as reading or writing the data, is covered in Chapter 3, "Understanding Access Controls and Privileges". Policy privileges are covered in Chapter 7, "Administering User Labels and Privileges"

When an Oracle Label Security policy is applied to a database table, a column is added to the table to contain each row's label. The administrator can choose to display or hide this column.

Label Components

This section describes the three elements defined for use in labels.

• Label Component Definitions and Valid Characters
• Levels
• Compartments
• Groups
• Industry Examples of Levels, Compartments, and Groups

Label Component Definitions and Valid Characters

A sensitivity label is a single attribute with multiple components. All data labels must contain a level component, but the compartment and group components are optional. An administrator must define the label components before creating labels.

Table 2-1 Sensitivity Label Components

Valid characters for specifying all label component include alphanumeric characters, underscores, and spaces. (Leading or trailing spaces are ignored.)

The following figure illustrates the three dimensions in which data can be logically classified, using levels, compartments, and groups.

Figure 2-1 Data Categorization with Levels, Compartments, Groups

Text description of the illustration olsag006.gif

Levels

A level is a ranking that denotes the sensitivity of the information it labels. The more sensitive the information, the higher its level. The less sensitive the information, the lower its level.

Every label must include one level. Oracle Label Security permits defining up to 10,000 levels in a policy. For each level, the Oracle Label Security administrator defines a numeric form and character forms.

For example, you can define a set of levels like the following:

Table 2-2 Level Example

Table 2-3 Forms of Specifying Levels  

Although the administrator defines both long and short names for the level (and for each of the other label components), only the short form of the name is displayed upon retrieval. When users manipulate the labels, they use only the short form of the component names.

Other sets of labels that users commonly define include TOP_SECRET, SECRET, CONFIDENTIAL, and UNCLASSIFIED; or TRADE_SECRET, PROPRIETARY, COMPANY_CONFIDENTIAL, PUBLIC_DOMAIN.

If only levels are used, a level 40 user (in this example) can access or alter any data row whose level is 40 or less.

Compartments

Compartments identify areas that describe the sensitivity of the labeled data, providing a finer level of granularity within a level.

Compartments associate the data with one or more security areas. All data related to a particular project can be labeled with the same compartment. For example, you can define a set of compartments like the following:

Table 2-4 Compartment Example

Table 2-5 Forms of Specifying Compartments  

S:OP,CHEM,FINCL

S:FINCL,OP,CHEM

Compartments are optional; a label can contain zero or more compartments. Oracle Label Security permits defining up to 10,000 compartments.

Not all labels need to have compartments. For example, you can specify HIGHLY_SENSITIVE and CONFIDENTIAL levels with no compartments, and a SENSITIVE level that does contain compartments.

When you analyze your data's sensitivity, you may find that some compartments are only useful at specific levels. Figure 2-2 shows how compartments can be used to categorize data.

Figure 2-2 Label Matrix

Text description of the illustration olsag005.gif

Here, compartments FINCL, CHEM, and OP are used with the level HIGHLY_SENSITIVE (40). The label HIGHLY_SENSITIVE:FINCL, CHEM indicates a level of 40 with the two named compartments. Compartment FINCL is not more sensitive than CHEM, nor is CHEM more sensitive than FINCL. Note also that some data in the protected table may not belong to any compartment.

If compartments are specified, then a user whose level would normally permit access to a row's data will nevertheless be prevented from such access unless the user's label also contains all the compartments appearing in that row's label.

Groups

Groups identify organizations owning or accessing the data, such as EASTERN_REGION, WESTERN_REGION, WR_SALES. All data pertaining to a certain department can have that department's group in the label. Groups are useful for the controlled dissemination of data, and for timely reaction to organizational change. When a company reorganizes, data access can change right along with the reorganization.

Groups are hierarchical: you can label data based upon your organizational infrastructure. A group can thus be associated with a parent group. For example, you can define a set of groups corresponding to the following organizational hierarchy:

Figure 2-3 Group Example

Text description of the illustration olsag009.gif

The WESTERN_REGION group includes three subgroups: WR_SALES, WR_HUMAN_RESOURCES, and WR_FINANCE. The WR_FINANCE subgroup is further subdivided into WR_ACCOUNTS_RECEIVABLE and WR_ACCOUNTS_PAYABLE.

Table 2-6 shows how the organizational structure in this example can be expressed in the form of Oracle Label Security groups. Notice that the numeric form assigned to the groups affects display order only. The administrator specifies the hierarchy (that is, the parent-child relationships) separately.

Table 2-6 Group Example  

Table 2-7 Forms of Specifying Groups  

S:CHEM:WR,WR_HR

Groups are optional; a label can contain zero or more groups. Oracle Label Security permits defining up to 10,000 groups.

All labels need not have groups. When you analyze your data's sensitivity, you may find that some groups are only used at specific levels. For example, you can specify HIGHLY_SENSITIVE and CONFIDENTIAL labels with no groups, and a SENSITIVE label that does contain groups.

Industry Examples of Levels, Compartments, and Groups

Table 2-8 illustrates the flexibility of Oracle Label Security levels, compartments, and groups, by listing typical ways in which they can be implemented in various industries.

Table 2-8 Typical Levels, Compartments, and Groups, by Industry

Label Syntax and Type

After defining the label components, the administrator creates data labels by combining particular sets of level, compartments, and groups. Out of all the possible permutations of label components, the administrator specifies those combinations that will actually be used as valid data labels in the database.

This can be done using the Oracle Policy Manager graphical user interface, or using a command line procedure. Character string representations of labels use the following syntax:

LEVEL:COMPARTMENT1,...,COMPARTMENTn:GROUP1,...,GROUPn

The text string specifying the label can have a maximum of 4,000 characters, including alphanumeric characters, spaces, and underscores. The labels are case-insensitive; you can enter them in uppercase, lowercase, or mixed case, but the string is stored in the data dictionary and displayed in uppercase. A colon is used as the delimiter between components. It is not necessary to enter trailing delimiters in this syntax.

For example, the administrator might create valid labels such as these:

SENSITIVE:FINANCIAL,CHEMICAL:EASTERN_REGION,WESTERN_REGION
CONFIDENTIAL:FINANCIAL:VP_GRP
SENSITIVE
HIGHLY_SENSITIVE:FINANCIAL 
SENSITIVE::WESTERN_REGION

When a valid data label is created, two additional things occur:

• The label is automatically designated as a valid data label. This functionality limits the labels that can be assigned to data. Oracle Label Security can also create valid data labels dynamically at runtime, from those that are pre-defined in Oracle Internet Directory. Most users, however, prefer to create the labels manually in order to limit data label proliferation.
• A numeric label tag is associated with the text string representing the label. It is this label tag--rather than the text string--that is stored in the policy label column of the protected table.

  Note: For Oracle Label Security installations that are not using Oracle Internet Directory, dynamic creation of valid data labels uses the TO_DATA_LABEL function. Its usage should be tightly controlled. See Inserting Labels Using TO_DATA_LABEL within the section Inserting Labeled Data, which starts.

  See Also: Chapter 6, "Creating an Oracle Label Security Policy" for instructions on creating label components and labels The Policy Label Column and Label Tags "Label Tags"

How Data Labels and User Labels Work Together

A user can only access data within the range of his or her own label authorizations. A user has:

• Maximum and minimum levels
• A set of authorized compartments
• A set of authorized groups (and, implicitly, authorization for any subgroups)

For example, if a user is assigned a maximum level of SENSITIVE, then the user potentially has access to SENSITIVE, CONFIDENTIAL, and UNCLASSIFIED data. The user has no access to HIGHLY_SENSITIVE data.

Figure 2-4 shows how data labels and user labels work together, to provide access control in Oracle Label Security. Whereas data labels are discrete, user labels are inclusive. Depending upon authorized compartments and groups, a user can potentially access data corresponding to all levels within his or her range.

Figure 2-4 Example: Data Labels and User Labels

Text description of the illustration olsag007.gif

As shown in the figure, User 1 can access rows 2, 3, and 4 because her maximum level is HS; she has access to the FIN compartment; and her access to group WR hierarchically includes group WR_SAL. She cannot access row 1 because she does not have the CHEM compartment. (A user must have authorization for all compartments in a row's data label, to access that row.)

User 2 can access rows 3 and 4. His maximum level is S, which is less than HS in row 2. Although he has access to the FIN compartment, he only has authorization for group WR_SAL. He cannot, therefore, access row 1.

Figure 2-5 shows how data pertaining to an organizational hierarchy fits in to data levels and compartments.

Figure 2-5 How Label Components Interrelate

Text description of the illustration olsag010.gif

For example, the UNITED_STATES group includes three subgroups: EASTERN_REGION, CENTRAL_REGION, and WESTERN_REGION. The WESTERN_REGION subgroup is further subdivided into CALIFORNIA and NEVADA. For each group and subgroup, there may be data belonging to some of the valid compartments and levels within the database. Thus there may be SENSITIVE data that is FINANCIAL, within the CALIFORNIA subgroup.

Note that data is generally labeled with a single group, whereas users' labels form a hierarchy. If users have a particular group, that group may implicitly include child groups. Thus a user associated with the WESTERN_REGION group has access to all data; but a user associated with CALIFORNIA would only have access to data pertaining to that subgroup.

Administering Labels

Oracle Label Security provides administrative interfaces to define and manage the labels used in a database. You define labels in an Oracle database using Oracle Label Security packages, or using the Oracle Policy Manager. Initially, an administrator must define the levels, compartments, and groups that compose the labels, and then she or he can define the set of valid data labels for the contents of the database.

The administrator can apply a policy to individual tables in the database, or to entire application schemas. Finally, the administrator assigns to each database user the label components (and privileges, if needed) appropriate for the person's job function.