MySQL and MariaDB are ready for use again

Between the two, MySQL had a slightly longer gestation period (about 2 years of work) than MariaDB, which instead took about 1 year to come out.

But, in the end, what was really done?

Let's start with a simple list comparing what, for me, are the most significant and expected features of both versions released:

MariaDB 10.1 The Galera part (cluster management) has been included in the basic package which, in this version, becomes one with the product

Improvements for the exploitation of the machine's multicores

Automatic compression of InnoDb tables

JSON format to better show how the optimizer thinks

Maximum execution time for individual queries

Optimizations for Union not creating temporary tables

Data obfuscation via encryption in tables/logs and replication.

It is possible to view the complete list of news on the MariaDB website.

MySQL 5.7 Native JSon management

Split some projects from the core project for faster evolution (replication and utilities are an example)

Improvements for the exploitation of the machine's multicores

GIS and TextSearch functionality also brought to Innodb

Creation of virtual columns

New SYS schema for performance and statistics

Expanded possibility of being able to perform operations on tables/databases "hot", i.e. not creating blocks and/or slowdowns to the rest of the system

Improvements and rewriting of some parts of the optimization.

You can view the complete list of new features on the MySQL website.

It should be underlined that both projects were also ported to the Power8 platform, on which we had already carried out preliminary tests thanks to the collaboration with IBM and MariaDB. Both MariaDB 10.1 and MySQL 5.7.9 use this platform to report their highest peak in terms of recorded performance and scalability in multicore management.

What's missing? Everything that isn't there and that I would like to have ;)

Personal considerations and jokes aside, the beauty of an open source product is the possibility of influencing it even if you are not part of the internal development team. This especially applies to new features and implementations. In this regard, in this article I report the link that refers to the new features that we want to implement on MariaDB 10.2 and which will be developed as a priority based on the number of votes obtained from the community:

http://bit.ly/mariadb-10-2-vote-new-features

So… register and vote, vote, vote! ;)

The teams of both projects are working again both to refine the versions just released (minor release) and for the next releases (major release). I had the pleasure of personally meeting many of the developers directly involved in the creation of these two projects at the Percona Live Europe in Amsterdam, which ended recently, where my colleague Andrea Tosatto and I held a session that concerned the use of one of these two databases. The general thought is that all of them were very curious to understand how and if the features they had developed could be useful or improved.

And what happens now? To conclude, my opinion is that both products are not only increasingly faster and scalable products, but that they are products that are increasingly modeled on the end user's requests and try to create more and more innovation in the relational database market which, in my opinion, always remains a sector with continuous progressive innovation, also given by the arrival of NoSQL... but if these are the results... welcome!

At the moment, my R&D team and I are already working to figure out how to make these two products best usable. At the end of this study I will try to report my personal impressions and experiences on both.

For those who would like to follow these two products more closely and exchange opinions with other experts in the sector, below are some interesting links to very active Facebook groups:

MARIADB ITALIAN USER GROUP