MirrorShip is a new generation of ultra-fast full-scene MPP databases.
MirrorShip fully absorbed the outstanding research results of relational OLAP databases and distributed storage systems in the big data era. On the basis of industry practice, MirrorShip further improved the optimization and upgraded architecture, and added many new functions to form a new enterprise-level product.
MirrorShip is committed to building an extremely fast unified analysis experience to meet a variety of data analysis scenarios for enterprise users, supports multiple data models (detailed models, aggregation models, update models), multiple import methods (batch and real-time), and can integrate and connect to various existing systems (Spark, Flink, Hive, ElasticSearch).
Mirrorship is compatible with the MySQL protocol, and you can use MySQL clients and commonly used BI tools to connect to StarRocks for data analysis.
MirrorShip uses a distributed architecture where data tables are divided horizontally and stored in multiple copies. The cluster size can be flexibly scaled and can support 10PB level data analysis; supports MPP frameworks for parallel accelerated computing; supports multiple copies and has elastic fault tolerance.
MirrorShip uses a relational model, uses strict data types and a columnar storage engine, reduces read/write amplification through coding and compression technology; uses vectorized execution to fully exploit the parallel computing power of multi-core CPUs, thereby significantly improving query performance.