The evolution of databases and being open source with MariaDB

Website Planet’s Miguel Amado talked with Daniel Black, Chief Innovation Officer of the MariaDB Foundation. With more than one decade of existence and big companies as users, MariaDB is a reference when we talk about databases. Check out to know more about the initiative and also plans for the future.

Please present MariaDB to our audience

MariaDB started out as a fork from MySQL, a relational database, in 2010 when MySQL was acquired by Oracle. Many original developers of MySQL are working at MariaDB.

It presents a stable interface to users to store and retrieve their data. In many ways, its stability is boring. But this is what the internet is built on; strong reliable products that perform a single function very well.

When you want to store and retrieve data; aggregate monthly sales figures, a history of user interactions, get the version of a Drupal page from several years ago, join data from different sources together to create value; a MariaDB has the mature SQL there that does this.

MariaDB has more than one decade of existence. What changed since the initial idea and what are the challenges that you face today?

The goal when MariaDB started was to ensure the continuity of MariaDB for MySQL users. There was a lot of fear that Oracle was going to dump MySQL, which proved not to be the case.

“Drop in replacement” to MySQL, uninstall MySQL, install MariaDB running on the same data storage, was the goal for many years but MariaDB was (and still is) innovating significantly for its user base in its own right. Now the MySQL compatibility is on the protocol level and to a large degree the SQL layer maintains MySQL compatibility.

We’ve picked up storage engines like RockDB along the way that proves good performance under more modern high data ingestion rates. MariaDB ColumnStore was added to MariaDB-10.5 providing a high performing analytical capability.

As we develop there is a significant ecosystem around products in the community related to high availability, sharding, user database tools, connectors for different languages that are all being produced independently from the MariaDB server development. It’s important to bridge those gaps to ensure that those communities know about new technologies coming in the server and their bugs and issues get communicated back to the server community.

Clouds, containers and automation took off significantly well and we’re still thinking and implementing ways to improve the database server to run better in those environments.

Ten years ago, there was x86 architecture of Intel and AMD, now there’s ARM, POWER as supported platforms. The scale of available RAM and CPUs has risen, and the development needs to support the high end of hundreds of Gigabytes of RAM and CPUs, and at the same time ensuring those running in a single micro instance in a container can still work well.

At the end of the day people are still asking, and getting, better performance for all the SQL queries they come up with. Newer features like a rich processing and handling of JSON are being requested and delivered.

Can you share with us some of MariaDB’s users and uses?

MediaWiki Foundation – Wikipedia

DBS Bank – Running banking systems

ServiceNow – Business Automation

http://www.automattic.com – company behind WordPress.com, WooCommerce

MariaDB is available for every company running cPanel and Plesk

Alibaba / Tencent / Amazon / MariaDB Corporation – DBaaS cloud services

How does being open source help the development of MariaDB?

In a lot of ways, the ecosystem of connectors and libraries is enhanced by being open source.

It helps users trust it more. The continuity of the codebase, one of the MariaDB Foundation’s goals, ensures that the community ecosystem that delivers connectors and adds features can exist. There are very few closed source applications with such a rich amount of independent development.

As it’s all being developed in the open, from source code and the features and bug lists, users are able to request and or change the code in a way that suits them. Hopefully they’ll provide it back for everyone to use, but that’s not a key requirement.

The open source allows large scale providers like Amazon and Alibaba, TenCent to use the code with changes to tightly integrate into their infrastructure and offer DBaaS.

MariaDB has a diverse team and board members, which is great and I believe must bring different ideas and visions for the software. How can you make sure everyone is pushing in the same direction?

There’s a great amount of respect for the differing user cases. Already as a product the SQL ideals of Online Transaction Processing vs Analytics is a continuum. The difference of ideas and visions is a strength that can be mutually achieved and people resolve the cases where there are competing requirements as to how to best technically achieve this.

What are MariaDB’s plans for the future?

To continue on the journey with the technology industry. New hardware and storage, along with software environments will continue to provide avenues to innovate improvements. Data growth is already exceeding the growth of CPU and RAM and architecturally the parallelism and distributed nature of MariaDB will need to close this gap to continue to give users the performance they demand.

As users move to a cloud environment we’re going to be ready to ensure that MariaDB is a superior database product in that environment. Our continual engagement with our community will ensure that users define our plans.

Also, to make sure that the level of support for MariaDB is smooth for the users, and that we’re going to let them know about new features coming and also look at the whole community, so to see what are the issues and features that those users actually want.

Because sometimes they are not coming directly to us: they are just working with their development communities of NodeJS, Laravel, Django for example, so it requires us to be there also, and get their requirements implemented in the MariaDB server and through the ecosystem stack to where our users are.