Simple architecture

StarRocks does not rely on any external systems. The simple architecture makes it easy to deploy, maintain and scale out.

Native vectorized SQL engine

StarRocks adopts vectorization technology to make full use of the parallel computing power of CPU, achieving sub-second query returns in multi-dimensional analyses.

Query optimization

StarRocks can optimize complex queries through CBO (Cost Based Optimizer). With a better execution plan, the data analysis efficiency will be greatly improved.

Query federation

StarRocks allows direct access to data from Hive, MySQL and Elasticsearch without importing.

Efficiently update

The updated model of StarRocks can perform upsert/delete operations according to the primary key, and achieve efficient query while concurrent updates.

Intelligent materialized view

The materialized view of StarRocks can be automatically updated during the data import and automatically selected when the query is executed.

Standard SQL

StarRocks supports standard SQL syntax (TPC-H and TPC-DS). It is also compatible with the MySQL protocol. Various clients and BI software can be used to access StarRocks.

Unified batch and streaming

StarRocks supports batch and streaming data import. Data sources such as Kafka, HDFS and local files can be imported smoothly.

High availability

StarRocks supports multi-replica data storage and multi-instance data deployment. The cluster has the ability of self-healing and elastic recovery.

High scalability

StarRocks adopts a distributed architecture which allows its storage capacity and computing power to be scaled horizontally. StarRocks clusters can be expanded to hundreds of nodes to support up to 10PB data storage.