starrocks

StarRocks is an open-source query engine designed for sub-second, ad-hoc analytics both on and off the data lakehouse. It is a Linux Foundation project licensed under Apache 2.0 and is implemented mainly in Java and C++. Key capabilities described in the README: - Native vectorized SQL engine: uses vectorization to exploit CPU parallelism, aiming at sub-second query returns for multi-dimensional analysis. - Standard SQL: supports ANSI SQL syntax (including TPC-H and TPC-DS) and is compatible with the MySQL protocol, so many clients and BI tools can connect. - Smart query optimization: uses a cost-based optimizer (CBO) to produce better execution plans. - Real-time update: supports upsert/delete operations by primary key while serving concurrent queries. - Intelligent materialized views: can refresh automatically during data import and be selected automatically at query time. - Data lake querying: allows direct access to Apache Hive, Apache Iceberg, Delta Lake and Apache Hudi data without importing it. - Resource management: limits query resource consumption and provides tenant isolation within a cluster. - Easy maintenance: a simple architecture intended to simplify deployment, maintenance and scaling, with automatic rebalancing and replica recovery after node failure. Architecture: the system consists of Frontend (FE) and Backend (BE) modules, with horizontal scaling and metadata/data replication to avoid single points of failure. Since version 3.0 it also supports a shared-data architecture for greater scalability and lower cost. The README links to downloads, documentation, benchmarks, a demo repository, Slack, YouTube, GitHub issues and a contributing guide, and lists numerous companies reported as users. It also mentions community events and a Gurubase assistant.