Dynamic vs. Static Data Masking
When masking data, organizations prevent unauthorized users from viewing sensitive data and protect information for following regulatory needs. Data masking technology provides data security by replacing sensitive information with a non-sensitive content, but doing so in such a way that the copy of data that looks and acts like the original.
In this article, we talk about the different types of data masking and discuss how organizations can use data masking to protect sensitive data.
Masking data isn’t the same as a firewall
Most organizations have a fair amount of security around their most sensitive data in the production (live) databases. Access to databases is restricted in a variety of ways from authentication to firewalls.
Masking limits the duplication of sensitive data within development and testing environments by distributing substitute data sets for analysis. In other cases, masking will dynamically provide masked content if a user’s request for sensitive information is considered ‘risky’. Masking data is designed to fit within existing data management frameworks and mitigate risks to information without sacrificing its usefulness. Masking platforms tend to guard data, locate data, identify risks and protect as information moves in and out of the applications.
Data masking hides the actual data. There are a variety of different algorithms for masking, depending on the requirements.
- Simple masking just turns characters to blank, so, for example, an e-mail address would appear as email@example.com
- More complex masking understands values, so, for example, a real name like “David Patrick” would be transformed into a fake name (with the same gender characteristics), like “John Smith”
- In some algorithms, values are scrambled, so, for example, a table of health conditions might appear with values of the health conditions, but not associated with the correct person for the particular salary
Most data masking tools will offer a variety of levels of masking that can be enabled in your network. Both static and dynamic data masking use these same masking methodologies.
Static data masking
Static data masking is used by most organizations when they create testing and development environments, and, in fact, is the only possible masking method when using outsourced contractors or developers in a separate location or separate company. In these cases, it’s necessary to duplicate the database. When doing so, it is crucial to use a static data masking tools. These tools ensure that all sensitive data is masked before sending it out of the organization.
Static data masking provides a basic level of data protection by creating an offline or testing database using a standard ETL procedure. This procedure replicates a production database, but substitute’s data that has been masked, in other words, the data fields are changed to data that’s not original or is not readable.
As shown in the image above, this means there are two databases, most of the time are not synchronized, because all of the real data may not be relevant for the test database.
It’s important to be aware that static masking can provide a backdoor, especially when outsourced personnel is used for administration, development, or testing. To mask data, the data is extracted from the database, at least for inspection, to comprehend the data before masking. Theoretically, this could provide a backdoor for data breaches, though it is not one of the common methods of malicious data capture.
Also, it’s clear that the static database always lags behind the actual data. The static database can be updated periodically, for example on a daily or weekly basis. This is not a security risk, but it often has implications for a variety of tests and development issues.
Static data masking allows database administrators, quality assurance, and developers to work on a non-live system so that private data is not exposed.
In many cases, in fact, you’ll want a test database anyhow. You don’t want to be running live experiments on a production database, so for R&D and testing, it makes sense to have a test database. There’s nothing wrong with this scenario.
Is your database protected with static data masking?
The answer should be obvious from the image above. Your actual production database is, in fact, not protected in any way when it comes to concealing sensitive information. Anyone or any system that has access to the production database might also have access to sensitive information. For most organizations, the only protection under this scenario is provided by limiting authorization access to the production database.
Concerns about static data masking
With static data masking, most of the DBAs, programmers, and testers never actually get to touch the production database. All of their work is done on the dummy test database. This provides one level of protection and is necessary for many environments. However, it is not a complete solution because it does not protect authorized users from viewing and extracting unauthorized information. The following concerns should be noted when using static database solutions.
- Static solutions actually require extraction of all the data before it is masked, that is, it actually guarantees the data gets out of the database in unmasked form. One of the most disturbing facts about static data masking is the standard ETL (extract, transform, and load) approach. In other words, the database information was extracted as-is from the database, and only afterward transformed. You have to hope or trust that the masking solution successfully deleted the real data, and that the static masking solution is working on a secure platform that was not compromised.
- The live database is not protected from those who do have permissions to access the database. There are always some administrators, QA, developers, and others with access to the actual live database. This personnel can access actual data records, which are not masked.
- For organizations where a test database is not necessary for other purposes, it is wasteful to have a full test database that is a copy of the full production database, minus identifying information. The cost is in the hardware and maintenance of the second system.
- Activities have to be performed twice: once on the test system and then implemented on the live system. There’s no guarantee that it will work on the production system, and then the developers or DBAs who need to debug the system will be either debugging on the testing system, or they will be granted permissions that allow them to see the actual live data.
Dynamic data masking: security for live systems
Dynamic data masking is designed to secure data in real time for live production and non-live systems. Dynamic data masking masks all sensitive data as it is accessed, in real time and the sensitive information never leaves the database. When a DBA or other authorized personal views actual data in the production database, data is masked or garbled, so the real data is never exposed. This way, under no circumstances, is anyone exposed to private data through direct database access.
Using a reverse proxy, the dynamic masking tool investigates each query before it reaches the database server. If the query involves any sensitive data, the data is masked on the database server before it reaches the application or the individual who is requesting the data. This way, the data is fully functional for development or testing purposes but is not displayed to unauthorized users.
Dynamic masking allows all authorized personnel to perform any type of action on the database without seeing real data. Of course, activities that are supposed to show data do show that data, but only to the authorized personnel using the correct access. When using advanced data masking rules, it’s possible to identify whether a particular field should be shown to a particular person, and under what circumstances. For example, someone may be able to access one hospital record at a time but only from a particular terminal or IP address, using a specific application and specific credentials. Accessing that same record using a direct database command would not work or would produce masked data.
Concerns with dynamic data masking
Dynamic data masking requires a reverse proxy, which means adding a component between the data query and response. Different solutions exist, some of which require a separate on-premises server, and others that are software-only based and can be installed on the database server.
Furthermore, when a company uses only dynamic data masking and does not have a production system, there are issues associated with performing functions on the live database.
The following concerns should be noted when using dynamic database masking solutions.
- Response time for real-time database requests. In environments where milliseconds are of crucial importance, dynamic masking needs to be carefully tested to ensure that performance meets the organization standards. Even when a particular item of data is not masked, the proxy does inspect the incoming request.
- Security of the proxy itself. Any type of software installed on the database server needs to be secure. And once a proxy is present, you have to enforce that the entire connections to the database are now passing through this SQL proxy. Bypassing this proxy in any way will result in access to the sensitive data without masking.
- Performing of database development and testing on live systems can cause errors in the production system. In many cases, DBAs perform changes on a limited part of the system before deploying. However, best practices would require a separate database for development and testing.
Static vs. Dynamic Data Masking
The main reason to use data masking is to protect sensitive and confidential information from being breached and protected according to regulatory compliance requirements. At the same time, the data must stay in the same structure, otherwise, the testing will not show accurate results. The data needs to look real and perform exactly as data normally would in the production system. Some companies take real data for non-production environments but sometimes the data may have other uses. For example, in some organizations, when a call center personnel views customer data, the credit card data may be masked on screened.
Generally speaking, most organizations will need some combination of dynamic and static database masking. Even when static data masking is in place, almost any organization with sensitive information in the database should add dynamic data masking to protect live production systems. Organizations with minimal development and testing can rely solely on dynamic data masking, though they may find themselves providing some data with static masking to outside developers or other types of contractors.
Advantages of static data masking
- Allows the development and testing without influencing live systems
- Best practice for working with contractors and outsourced developers, DBAs, and testing teams
- Provides a more in-depth policy of masking capabilities
- Allows organizations to share the database with external companies
Advantages of dynamic data masking
- The sensitive information never leaves the database!
- No changes are required at the application or the database layer
- Customized access per IP address, per user, or per application
- No duplicate or off-line database required
- Activities are performed on real data, saving time and providing real feedback to developers and quality assurance