Multi-SQL: An Automatic Multi-model Data Management System.

Nowadays, data in applications become diverse and large in scale. In order to meet the increasing demand for multi-model data management, multi-model databases have evolved into huge systems with many knobs. However, such a large system brings great challenges for users to efficiently tune. In extreme cases, it is difficult even for experienced DBAs to be the masters in multi-source engines and provide effective solutions for multi-model data. Meeting this challenge, we firstly propose an automatic multi-model data management system (Multi-SQL) based on machine learning. We design an intelligent middleware in which some artificial intelligence(AI) technologies are embedded, including automatic multi-model storage selection and automatic multi-model index recommendation methods. In this paper, we clarify the architecture, core techniques and key scenarios of Multi-SQL.