Dark Reading is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Application Security

04:45 PM
Connect Directly

SQL Server 2019 Tool Tells Attackers Which Data Is Sensitive

The design of SQL Data Discovery & Classification could let attackers pinpoint sensitive information while flying under organizations' radars.

SQL Data Discovery & Classification, a tool within Microsoft's SQL Server 2019, could inform attackers which data within a database is labeled sensitive, and which isn't, researchers report.

Imperva security research engineer Avidan Reich emphasizes this tool does not grant attackers access to sensitive data, nor is this finding a vulnerability in SQL Server 2019. Rather, the research reveals a security issue that exists in the way SQL Data Discovery & Classification is designed to work.

This tool is built into SQL Server Management Studio (SSMS) to let users detect, classify, and report sensitive data stored within their databases. The classification engine first scans a database and identifies which columns hold potentially sensitive information. From there, the tool gives employees a simpler way to apply classification recommendations and manually classify columns, either by SSMS GUI or by SQL statements "Add Sensitivity Classification."

When the data's classification state is determined, it's added to the audit log so employees can better monitor access to sensitive information for compliance and auditing, Reich explains.

There are tools designed to complete this process outside the database, he points out in a blog post on his findings. The "segregation of duties" principle advises a best practice of only giving database administrators the tools they need to do their jobs: designing databases, managing the database, and monitoring its usage and performance. The idea behind this principle is to separate responsibilities so no single employee has access to too much sensitive information.

If organizations adhere to the segregation of duties, this means only security personnel would perform a classification, he explains. An application owner would review and label sensitive data; a security expert would apply the appropriate controls. If a scan is conducted outside the database, the admin is not involved and doesn't have the opportunity to classify the data.

The issue Reich found with SQL Data Discovery & Classification is it shows where sensitive information is stored by labeling it within the database itself. This makes it easier for a malicious insider, or an employee whose credentials have been compromised, to figure out which database columns contain sensitive data and then gain access to it, Reich explains. They would only need the "view any sensitivity classification" permission and a simple query.

"For insider threats, such as malicious employees or employees whose security has been compromised, it would be very easy to use the tool output to access sensitive data under the radar of security tools, such as behavior analytics solutions," he says. It's worth noting that database admins and accounts used by applications connecting to the database usually are able to view both nonencrypted and encrypted information.

"Without the tool output, insider threats would have to scan all application tables for sensitive data, which is a noisy approach that can trigger an incident within any behavior analytics security solution," Reich continues. An attacker who is able to ascertain exactly where valuable data is located can bypass the added step of scanning application tables.

An attacker who holes the server permission Control Server, or the database permission Alter, could also update sensitive data with the label "Drop Sensitivity Classification." This could render a sensitive column nonsensitive; in doing this, an attacker could make the data accessible under data activity monitoring and behavioral analytics tools.

Imperva shared the details of their findings with Microsoft, which says the tool's purpose is to discover and help classify data. "The reported findings do not pose a security risk and we do not plan to address it with a security update," officials said in an email to Dark Reading.

If your business uses the SQL Data Discovery & Classification tool, Reich advises the following mitigation steps:

  • Monitor access to the catalog view "sys[.]sensitivity_classifications, which has the location of the sensitive data.
  • Monitor executions of SQL statement "Drop Sensitivity Classification," which deletes the classification label.
  • Verify that only authorized accounts can execute the SQL statement "Drop Sensitivity Classification."

Related Content:

Check out The Edge, Dark Reading's new section for features, threat data, and in-depth perspectives. Today's top story: "Disarming Disinformation"

Kelly Sheridan is the Staff Editor at Dark Reading, where she focuses on cybersecurity news and analysis. She is a business technology journalist who previously reported for InformationWeek, where she covered Microsoft, and Insurance & Technology, where she covered financial ... View Full Bio

Recommended Reading:

Comment  | 
Print  | 
More Insights
Newest First  |  Oldest First  |  Threaded View
Look Beyond the 'Big 5' in Cyberattacks
Robert Lemos, Contributing Writer,  11/25/2020
Why Vulnerable Code Is Shipped Knowingly
Chris Eng, Chief Research Officer, Veracode,  11/30/2020
Register for Dark Reading Newsletters
White Papers
Cartoon Contest
Write a Caption, Win an Amazon Gift Card! Click Here
Latest Comment: I think the boss is bing watching '70s TV shows again!
Current Issue
2021 Top Enterprise IT Trends
We've identified the key trends that are poised to impact the IT landscape in 2021. Find out why they're important and how they will affect you today!
Flash Poll
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
PUBLISHED: 2020-12-02
CAPI (Cloud Controller) versions prior to 1.101.0 are vulnerable to a denial-of-service attack in which an unauthenticated malicious attacker can send specially-crafted YAML files to certain endpoints, causing the YAML parser to consume excessive CPU and RAM.
PUBLISHED: 2020-12-02
Editors/LogViewerController.cs in Umbraco through 8.9.1 allows a user to visit a logviewer endpoint even if they lack Applications.Settings access.
PUBLISHED: 2020-12-02
A security vulnerability has been identified in the HPE Edgeline Infrastructure Manager, also known as HPE Edgeline Infrastructure Management Software. The vulnerability could be remotely exploited to bypass remote authentication leading to execution of arbitrary commands, gaining privileged access,...
PUBLISHED: 2020-12-02
HCL Domino is susceptible to a Buffer Overflow vulnerability in DXL due to improper validation of user input. A successful exploit could enable an attacker to crash Domino or execute attacker-controlled code on the server system.
PUBLISHED: 2020-12-02
An out-of-bounds memory write flaw was found in how the Linux kernel’s Voice Over IP H.323 connection tracking functionality handled connections on ipv6 port 1720. This flaw allows an unauthenticated remote user to crash the system, causing a denial of service. The highest threat ...