Main content

MariaDB CEO takes aim at Oracle and launches new Labs

Derek du Preez Profile picture for user ddpreez February 26, 2018
MariaDB, the open source database that offers a “drop-in replacement for MySQL”, is trying to make it easier for customers to switch from Oracle.

MariaDB CEO Michael Howard

MariaDB CEO Michael Howard took aim at Oracle today at the company’s second annual user conference in New York, stating that he is on a mission to provide enhanced capabilities and familiarity for those customers still using MySQL.

For those unaware, MariaDB aims to provide an open-source, “drop-in replacement for MySQL” and is a fast-growing vendor in the database market. My colleague Jon Reed last year spoke to Howard about why relational databases still matter, despite the growing popularity of NoSQL alternatives. And about why it doesn’t mean the relational database giants, such as Oracle, will necessarily win in the future.

MariaDB was created by the original developers of MySQL and Howard is himself an ex-Oracle executive. The company is fresh off the back of a new round of funding, led by Alibaba, where it secured $54 million.

Since it started, the company’s database has now been adopted in all the Linux distributions, becoming the default in both Red Hat and Suse. It has also been selected by both Google and Wikipedia.

Howard kicked off the keynote address in New York today by urging customers to resist the “fear and the mystery of the unknown”. He said:

If you do not change, you will be beholden to a company that does not have your best interests in mind. I know this. I was there. You will pay ten times more for that privilege. You know this. Your presence here acknowledges it. You don’t want someone in particular - you know who I mean - or some ideology, to dictate the criteria of success and your future. MariaDB is not going to let that happen. That’s not how we role.

Howard also referenced the release of MariaDB 10.3 today, which he hopes will make it even easier for those using Oracle to switch over to MariaDB. He said:

With this presence, with these capabilities and with these new resources, comes bigger expectations. We have to make it easier for global enterprises to be able to easily change and migrate. Just this morning we released 10.3, which offers Oracle capability and a set of other advanced features. To not only to provide familiarity and portability in terms of code, but to provide portability and familiarity in terms of skill sets. Mastering of skill sets of one era, to move across into a new era.

Isn’t it ironic that MariaDB is offering an Oracle compatibility layer when MySQL, a part of Oracle, doesn’t? Isn’t it ironic that Oracle Enterprise, MySQL’s bigger brother, provides data warehousing, yet it’s MariaDB that’s delivering it to you? They don’t want you to succeed with MySQL. They certainly don’t want MySQL to cannibalise things like Exadata. We have to dare to be better and different. We have to have freedom to make a great database. Freedom is so important here.

It’s incumbent upon us to work with global enterprises and proven innovators to drive features and capabilities into the product for everyone in the community.

Cloud ambitions

During the keynote it became clear that Howard wants MariaDB to be at the forefront of certain technology trends, where perhaps it has been lagging slightly. For example, he said that it will be using the new investment to further expand its cloud offering to customers, which is clearly an imperative in this digital age. Howard said:

When we think of things like clouds, containers and micro-services, these are new ways in which MariaDB can be managed and deployed. So one of the things that we are going to do with our investment dollars is invest in our own cloud and DBaaS service. So let anyone manage and deploy MariaDB, anywhere. Be it on a private or a public cloud, or both. The first project to support this endeavour is something called MariaDB Manager, which is a visual interface allowing the management and deployment of MariaDB.

Now we are not stopping there for this cloud endeavour. We are also working with some partners on lightweight file systems to leverage shared storage for low latency, high availability. In essence, to create a response to what Aurora offers on AWS, but is restricted within their cloud environment. To let a a similar technology be ubiquitously available. So mark my works, you will see a lot coming out of MariaDB with regards to this cloud endeavour.

More to come

Howard admitted during his keynote that “there are some things that have not yet been conquered”, things that “take time”. In response to this, the company is launching MariaDB Labs, which aims to solve ubiquitous technology problems that can’t be tied to any particular release. Specifically, MariaDB Labs will focus on (in its words):

  • Machine learning to transform the ways in which MariaDB is deployed and managed in public and private clouds, and employ learning methods for database self-configuration and self-optimization.
  • Distributed computing to break through long-held beliefs in relational technology and compromises in NoSQL, raising the bar for web-scale, geo-distributed deployments and executing on new approaches for one of the industry’s most difficult to solve problems.
  • Next-generation chips, memory and storage to radically change the infrastructure on which databases run, advancing performance, high availability, costs and energy conservation.

Howard said that these are “hard problems that need different types of financial resources”. Providing some more detail, Howard said:

In terms of machine learning, [we are going to] make the database smarter. To go beyond what cost optimisers do. We are also going to apply this set of methods to our cloud. Allowing MariaDB to instantiated via an SLA approach, rather than a template approach. Imagine, if you will, that you no longer have to declare how much memory, storage or compute you need for a particular application. The system does it on your behalf. Machine learning is wonderful for this.

[Also], machine learning lends itself well to distributed computing. In that, we at MariaDB want to handle web scale, geo distributed workloads better, via seamless, scalable architecture. Now we can’t walk down the road and look at real application clusters, albeit the gold standard, because it’s based on old technical precepts. It needs a refresh. And we can’t search out and collaborate with our NoSQL relatives, because of their naivety regarding asset properties, consistency, isolation, reliability. We need a more modern and widened approach to distributed computing and that’s what MariaDB Labs is all about.

My take

An interesting start to the day with MariaDB. Howard clearly is on a mission to make MariaDB as useable and as competitive as the databases offered by the incumbents. However, the detail about its future requirements will be found in our discussions with customers - which we will be having plenty of over the coming days.

A grey colored placeholder image