Introduction

Our world is full of data that is constantly changing. We need to keep track of that data and extract anything valuable from it. Relationships between different kinds of data can tell us more than the data itself, and those relationships help detect any anomalies in network systems. If we want to make our network safe, we must be aware of all changes and how they affect our system as a whole.

Applying graphs

Data privacy

Nowadays, protecting an individual's privacy preferences is maybe more important than ever. It is possible to track private information with graph databases and when and how someone used that information. Also, systems that are using specific data can be tracked. Using graph databases, the personal data possession of some companies can be General Data Protection Requirements (GDPR) compliant.

Besides that, data classification is the process of classifying data based on user-configured characteristics, which represents a big part of data security strategies. Data classification can help you find the location of sensitive and regulated data.

Track data lineage

Data lineage tells us more about the life cycle of our data. It shows the whole data flow from its creation until its final form. That can be easily visualized using a graph database. By tracking data lineage with a graph database, you can follow any error in data processes or implement changes to some processes more efficiently. With data lineage, you can track whether your data has been correctly transformed and catch if some untrusted source has created the data.

Previously mentioned data classification can be combined with data lineage. After data classification has discovered the location of sensitive data, data lineage finds out the whole life cycle of those files and figures all the possible security threats.

Data access control

Organizations can authorize certain employees to access company data securely with data access control. In that way, the organization's access to resources complies with the company's policies and any official regulations. There are different types of access control. Usually, employees are divided into roles, and each role has access to some part of the data. Graph databases come as a perfect solution to that problem since each employee can be represented with node, and based on their role property, they could have access (be connected with) certain data. Another way to approach access control is data-centric access control. Each data has a type, and different users can access data based on the type of data they have permission to access. The third way is context-centric access control, where everything depends on the nature of access. For example, permissions are given based on the time of access and the amount of data that an employee is trying to fetch. Companies usually combine these access controls and, with graph databases, are able to create unique data access control.

Anomaly detection in data security

Anomaly detection identifies rare occurrences or events of concern due to their differing characteristics from the majority of the data. Detecting anomalies in data can help organizations track security errors, structural defects, or bank frauds. There are numerous techniques for spotting anomalies in a large set of data, and graph algorithms on your data stored in a graph database are among them. When you find some anomaly, you can use found features to predict future frauds.

Where to next?

This text is a summary of one area that fits perfectly with the application of graphs. Therefore, we would like to have you with us when implementing some of these solutions. Share opinions, experiences and problems you encounter when working with Memgraph on our Discord server. We are here for you and we will help you along the way.