The open source database company MariaDB is launching a research division aimed at tackling the most pressing issues in the database field.
Speaking at the M18 user conference in New York yesterday, MariaDB CEO Michael Howard identified that the labs will focus on three key areas: "Machine learning, distributed computing and the use and exploitation of new chips, persistent storage and in-memory processing."
This is in addition to a launch collaboration with Intel to look into a new reference architecture for distributed databases.
In terms of machine learning, the labs will be tasked with investigating how to use supervised and unsupervised techniques to drive better automation, including self-configuration and self-optimisation when running databases in the cloud.
Secondly the research will look at distributed computing, specifically improving upon web-scale, geo-distributed deployments.
Lastly, the labs will look to develop next generation chips, memory and storage in a bid to rethink the underlying infrastructure on which databases run.
Howard further identified some key areas of the research division as part of the announcement press release: "Geo-distributed applications, non-volatile memory, predictive optimisers and self-driving databases are the new inflection points in this era of modern data computing," he said.
The first challenge the lab is taking on is in collaboration with Intel, as they try to develop a reference architecture for databases being run on distributed storage.
The idea is to develop a new open source database architecture which will work on disaggregated cloud-centric storage, with fast recovery and cloning through an automated recovery process a key aim.
Alper Ilkbahar, Intel vice president and data centre group general manager said: "Our research collaboration with MariaDB will focus on an innovative distributed log architecture using Intel Optane SSDs and Intel Persistent Memory woven together through high-performance fabrics. Our shared goal is to enable faster recovery and cloning, increase overall performance and resilience, and reduce solution TCO from today's level."