Database and application vendors may want to take a look at expanding their data masking capabilities as its relevance grows. While the market is currently small, Forrester Research expects it to grow in the years ahead as businesses look to ensure security in nonproduction environments.
Data masking is emerging as an important-if underutilized-element of
ensuring data protection.
With its use expected to grow, it may be time for database and
application vendors to consider a more aggressive push into the area.
Forrester analyst Noel Yuhanna believes data masking will eventually become
a standard feature found in database management systems and packaged
applications. Currently, the market is small-Yuhanna estimates about $22
million-but is expected to increase in the years ahead on the back
of compliance regulations.
"Data privacy is one of the foundational pillars needed to ensure compliance
with many emerging compliance standards like PCI, SarBox, Basel II, etc.," said
Al Smith, director of IBM's Optim
Engineering. "By de-identifying personal and sensitive data, significant risks
can be mitigated and costs can be saved while providing support for compliance
with privacy regulations and corporate governance standards."
IBM made a significant play in this space
when it purchased Princeton Softech in 2007. The company now offers up the
technology in its Optim Data Privacy product. IBM
has also made a number of recent investments around ensuring data privacy
during its entire life cycle, such as integration with IBM
Rational Data Architect.
"In the design phase a compliance analyst can specify privacy policies as
part of the data model to the Develop/Test phase, where [the] Optim Test Data
Management Solution can use the policies to de-identify sensitive data as they
generate subset test data for the dev/test/offshore process," said Jim
Sinisgalli, product manager for IBM Optim.
Rival database vendors Oracle and Microsoft have made plays of their own.
Oracle offers a data masking solution as an add-on, while Microsoft enables
users to create a transformation function through the use of SQL Server
Integration Server that can mask data as it's moved to a test database.
Microsoft's Visual Studio Team Server Database Edition provides the ability to
generate test data using its Data Generation Plans feature as well.
Still, there is plenty of room in the market for third-party companies such
as DataGuise. Joe Lawless, the company's vice president of sales and marketing,
said security and compliance officers know insider threats are a significant
risk in nonproduction environments, where development, testing and training
occur. He contended, however, that for the most part developers and database
administrators are using "patchwork solutions" at best.
That sentiment jibes with what Forrester has found. Yuhanna estimated about
14 percent of enterprises do data masking today, with the majority deploying a
solution developed in-house. Many, however, use production data out in the open
in test environments. A
by the Independent Oracle Users Group in
September reported that more than 40 percent of organizations exposed live
data in nonproduction databases.
"The key challenges in implementing data masking are often not technology
related, but knowing what to mask, defining the policies, integrating with
applications, testing with other apps, changing the data movement process, and
getting business and management buy-in," he said. "Unlike auditing, where you
see the benefits of it right away after installing it, data masking projects
often tend to take time and take anywhere between six to nine months to implement
from the ground up for one or two large applications."
There is also the economy to consider. Given the current situation, it is
possible that the market may fall victim to economics if businesses slash their
IT budgets, Yuhanna said.
"Data masking is definitely a good solution for test and development
databases, but most organizations still struggle in securing their production
databases-forget about test and dev ones-which is why data masking often gets a
lower priority," he said.