How to Ensure that 'Agile' Data Is Adequately Secured
- By Caio Milani
- Apr 13, 2017
It’s not unusual for companies to have siloed data, but it’s not OK for that data to remain siloed—not if a company wants to stay competitive. The ability to integrate data from disparate sources in order to analyze and make sense of it is not only key to doing business today; it’s the difference between a company’s success and failure. But so, too, is data security.
And therein lies the rub. Companies are trying to strike a delicate balance between an agile data model and strong data security. Companies must be able to zero in quickly on customer needs and just as quickly provide products and services that meet those needs. But this pace cannot come at the price of data security, privacy and compliance. And nowhere is this more true than at the database level, where all of this information is stored.
The level of integration and agility that companies must achieve today requires a high-level of flexibility and security. It’s a huge challenge: When your most sensitive and valuable data is being integrated across multiple silos of data, it takes a combination of products and processes to ensure that data at rest and in motion is saved in a secure and well-governed manner. Implemented in a strategic way, these capabilities can protect against some of the most sophisticated security threats companies are facing today—and in doing so provide competitive advantage.
True data security in this dynamic new model requires
Advanced encryption: Encryption is not a new feature in databases, but encryption must be implemented in an increasingly more strategic and systematic way to protect data from cyber criminals and insider threats. Advanced encryption involves the selective and transparent encryption of data, configuration and logs. This includes granular, role-based access, standards-based cryptography, advanced key management, granular separation of duties, and state-of-art algorithms that drastically decrease exposure. Advanced encryption is important due to the rise in frequency and complexity of internal and external security threats, expanding security requirements, and the growing use of the cloud among companies large and small.
Redaction: Companies need to balance protection of data with the ability to share it. Redaction enables companies to share information with minimal effort by concealing or masking sensitive information—such as names and Social Security numbers--when data is exported for sharing purposes. Companies must also be able to implement policy-based redaction using both custom and out-of-the-box rules, including partial masking, full masking and concealment. Some purposes require sample data for testing or anonymized information for data analysis that need policy-based redaction using consistent data outputs based on dictionary and deterministic masking.
Element-level security: While redaction in and of itself is important, companies need to be able to do it in real-time, as close to the data as possible. Security at the element, or property, level--based on an employee’s role--enables companies to protect sensitive information during queries and updates. Element-level security should be extended into document elements and built into indexes for performance. Rich XML and JSON document models can even describe in the data itself, using a concept called markings, how element-level security should protect all parts of the document. This allows security definitions to actually travel with the data in contrast to definitions in a schema table.
Certification: There are a lot of options out there, and it can be difficult to determine which products are needed in the first place--not to mention which actually do what they say they will do. Certification such as Common Criteria ensures that a product can be evaluated to determine that is meets specific security properties at a certain level of assurance.
These data security features are important across company sizes and industries. Take, for instance, the activities involved in serving customers in healthcare, or the financial industry. A doctor or financial analyst should have access to a great deal of patient or customer information; on the other hand, a call center user should have a restricted view.
Implementing this type of security at the application level leaves too many open entry points, giving attackers plenty of opportunity to bypass security and collect information. Implementing security rules in a database schema reduce the business agility, as schema changes are costly and brittle.
True data security has to be enforced at the database level in real-time and at rest, and better yet, based on information in the data itself and not schemas. A certified multi-model database with capabilities such as encryption at rest, redaction and element-level security ensures that companies can meet their security goals, while quickly adapting to evolving customer needs .
With data driving business decisions, companies must be able to effectively manage the entities and relationships that define the business. And with this enhanced data agility comes the need for enhanced security. Companies must ensure that they are implementing not only database technology that supports the integration needed to connect the dots between disparate data, but also the security required to protect that data.