Relational Database Service

Almost every app stores data permanently: names, orders, account balances. For this, one uses a database, meaning a program that stores data in an organized way and retrieves it quickly. A relational database stores this data in tables with rows and columns, similar to a very large Excel file. A Relational Database Service is exactly such a table-based database that you rent from a cloud provider. Cloud here means: the machines sit in the data center of another company, and you access them over the internet. The provider installs, backs up, and updates everything, the customer pays monthly and only works with their data.

Why companies prefer to rent their database

Running a database yourself is surprisingly much work. Someone has to set up the server, choose the right software version, and close security holes. Someone has to create backups daily and regularly check whether they can really be restored. This is unglamorous work that nobody notices as long as it works.

These are exactly the tasks the provider takes on with a Relational Database Service. A start-up with five people then doesn’t need its own database specialist. This saves not only money but, above all, time until the first working product. For investors, this is one reason why small teams start faster today than fifteen years ago.

The price for this is dependency. Whoever has their data stored with a provider can only leave with considerable effort. Experts call this lock-in, a kind of confinement effect. In addition, renting is often more expensive in the long run than owning hardware. Some large companies therefore deliberately bring their databases back in-house.

What the provider handles in the background

Technically, the database still runs as a completely normal program on a server. What’s new is the layer around it. The customer clicks together in a web interface which database software they want and how much computing power and storage space. A few minutes later, they get an address at which their database can be reached. Their app then connects just as it would with a self-managed database.

In the background, automation runs. Backups are created on a schedule and kept for a few days. Often a second copy of the database runs simultaneously in a different data center. If the first one fails, the second takes over within seconds. This approach is called high availability.

A common misconception: the service does not automatically make the database fast. If an app issues poorly formulated queries, it will also be slow in the cloud. The provider takes care of operations, not the quality of your own software. The service should be distinguished from plain server space, where you get an empty operating system and have to set up everything yourself.

Where these services show up in everyday life

The best-known offerings are Amazon RDS, Google Cloud SQL, and Azure SQL Database. Amazon coined the term with its product name RDS, which is why many people equate the two words. Behind them are usually well-known database systems like PostgreSQL or MySQL, which already existed before.

As a user, you notice none of this directly. But the online shop where you order, and your bank’s app, very likely store their data in such a service. In financial news, these database services show up as a significant revenue segment of the big cloud providers. They are considered a particularly reliable business because customers rarely switch their database.

They also play a role in AI projects. Training data, user accounts, and logs of chat histories are often stored in such a relational database. The spectacular models get the attention, but without this unglamorous infrastructure, no product built around them would work.