Scribd
Upload
23 views3 pages

Protect Your Data With Oracle Data Redaction in Oracle Database 23ai

Uploaded by

Parthiban
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
23 views3 pages

Protect Your Data With Oracle Data Redaction in Oracle Database 23ai

Uploaded by

Parthiban
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Protect your data with Oracle

Data Redaction in Oracle


Database 23ai
October 15, 2024 | 4 minute read

Richard Evans
In today's data-driven world, businesses often have a wide array of applications—from online
transactional processing (OLTP) systems and data warehouses to analytics and AI models—all of
which may access the same underlying data. Protecting data privacy and security in these diverse
environments can be challenging, especially when sensitive information needs to be selectively
redacted depending on who's accessing it - the application in use, an analytics tool, or an ad-hoc
query.

This is where Data Redaction shines. It allows you to redact (hide without changing the actual data)
column results through centrally managed security policies that govern how shared sensitive data is
used by different applications, analytic tools, or ad-hoc query tools. Let's explore the benefits of Data
Redaction.

Key Benefits:
1. Ease of Use: Data Redaction is easy to implement, usually requiring no application changes.
Once you create Data Redaction policies, they apply seamlessly to any query, providing
consistent security for all data consumers.
2. Application Independence: There is no need to create individual security policies for each
application. Data Redaction policies are centrally managed and enforced directly within the
database, simplifying security management across diverse environments.
3. Data Layer Application: Since redaction happens at the data layer, it covers every
application, tool, report, and model using the data. Redaction policies automatically apply as
new data is inserted into a protected column, making this solution scalable as your data
grows.
4. Dynamic Redaction: Data redaction can be dynamic and based on expressions defined in the
policy. For instance, you can create policies that redact information based on user name, user
role, specific application, or IP address, supporting varied use cases and ensuring that actual
data is displayed in full only under appropriate conditions.
5. Portability: Both data and redaction policies are transportable across different environments.
This allows for smoother migrations, development, and testing cycles and ensures that data
protection rules stay consistent.

Use-cases for Data Redaction


Complying with regulatory, industry, and organizational
security goals
In the past decade, there has been a significant push to limit access to personally identifiable
information. Unfortunately, not all your software makes it easy to limit this data from being shared
with the end user. Data Redaction can help you here. Instead of changing the report, query, or tool to
redact the data, you can implement a Data Redaction policy on the data, so the actual data is not
returned to the report, query, or tool.

For example, your EMPLOYEES Table has several sensitive columns, including HIRE_DATE,
BIRTH_DATE, GENDER, STREET, CITY, STATE, ZIP_CODE, and BASE_SALARY. You may
create a policy as such:
 The application user and an application role can view actual data in
the HIRE_DATE, STREET, CITY, STATE, and ZIP_CODE columns.
 Only the application user can view data the BIRTH_DATE, GENDER,
and BASE_SALARY columns.
 Regular users cannot access any of the above column data.

Redacting data in aggregate, analytic, and AI


environments
When dealing with analytics and AI, data often gets aggregated or processed through complex queries
and functions. Data Redaction allows sensitive columns to be redacted even when data is used in
aggregate or analytic functions, reducing the risk of exposure during analysis. Data Redaction can
help prevent unauthorized data proliferation when data is used by different applications or reporting
tools.

For example, a financial analyst wants to predict credit usage within the next 12 months. They could
have data in columns such as

CREDIT_SCORE, CREDIT_USAGE, LAST_USAGE, SSN, GENDER, POSTAL_CODE, and ADDRESS. They


may want to redact the SSN, GENDER, and ADDRESS columns because they are irrelevant to the
model. They can create a Data Redaction policy to prevent those columns from being used when
training the model.

Supporting Data Sovereignty and Compliance


Data Redaction can play a crucial role in supporting data sovereignty initiatives. You may want to
create a Data Redaction policy on your CUSTOMER table to redact data when accessed by clients
outside of the approved network, host, IP address, or other identifying context.
When designing your Data Redaction policy, your expression can use contextual information the
Oracle Database knows about the session. For example, this named policy expression allows all hosts
to see redacted data unless they are one of the approved production HR application servers listed. You
can then apply this named policy expression to columns in the CUSTOMERS table or other tables
where you want to enforce the same limitation by the host.
Data Redaction allows you to create policies and named column expressions to control where and
how your sensitive data is provided in an unredacted format.

Performance optimization of data access


Different workloads call for different strategies for data indexing, statistics collection, and Data
Redaction policy expressions.
For example, an application you support always searches the LAST_NAME column in uppercase. To
improve the query's performance, you would apply a function-based index to transform the column to
uppercase within the index (e.g., UPPER(last_name)).
You can use extended statistics to create a correlation between columns. For example, if you have
columns like CITY and ZIP_CODE, which are frequently accessed together, it makes sense to
optimize the statistics collected for them as a pair rather than independently.

Suppose you always want specific columns to be redacted for all non-administrative users. In that
case, you will typically use a policy expression function of '1=1', which always evaluates to TRUE, so
long as the querying user is not exempt from the redaction policy. Using this condition saves CPU
time during the fetch operation from a redacted column by avoiding the need to evaluate any policy
expression.

Conclusion
Oracle Data Redaction is one of many capabilities in Oracle Database 23ai that demonstrates our
commitment to providing robust, flexible, and secure database solutions. Data Redaction enables you
to protect sensitive information more effectively while leveraging the full power of the Oracle
Database.

You might also like