How apache/doris works
Apache Doris is an MPP analytics and search database. A user submits SQL to the Frontend; query execution is distributed through Backend pipeline fragments, while storage and ingestion support real-time analytics and lakehouse access. Hybrid search extends SQL over structured, JSON, text, and vector data. The tree and excerpts show portions of Backend execution, cloud storage, authentication, and query-profile handling; other capability links are documented in the README, not verified in sampled implementation. Dashed links mark those documented but unsampled paths.
Main entry points: fe-core SQL Frontend · cloud_storage_engine.cpp Cloud storage engine · agent_server.cpp Backend nodes · pipeline_fragment_context.cpp Pipeline fragments
Components 20
Frontend services
- SQL Frontend
fe/fe-core - Authentication
fe/fe-authentication/fe-authentication-handler/src/main/java/org/apache/doris/authentication/handler/AuthenticationService.java - Authorization context
fe/fe-core/src/main/java/org/apache/doris/mysql/privilege/EngineAuthorizationContext.java - Query cancellation
fe/fe-core/src/main/java/org/apache/doris/qe/QueryCancelWorker.java - Query profiles
fe/fe-core/src/main/java/org/apache/doris/httpv2/controller/QueryProfileController.java
Backend execution
- Backend nodes
be/src/agent/agent_server.cpp - Pipeline fragments
be/src/exec/pipeline/pipeline_fragment_context.cpp - Task memory control
be/src/runtime/workload_management/query_task_controller.cpp - Columnar data model
be/src/core/block/block.cpp
Storage and ingestion
- Storage engine
be/src/storage/storage_engine.cpp - Cloud storage engine
be/src/cloud/cloud_storage_engine.cpp - Cloud metadata
be/src/cloud/cloud_meta_mgr.cpp - Cloud compaction
be/src/cloud/cloud_base_compaction.cpp - Stream loading
be/src/cloud/cloud_stream_load_executor.cpp - Object storage
common/cpp/obj-client/obj_storage_client.cpp
Analytics capabilities
- Real-time analytics
be/src/cloud/cloud_delta_writer.cpp - Lakehouse analytics
fe/fe-connector/fe-connector-spi/src/main/java/org/apache/doris/connector/spi/ConnectorStorageContext.java - Hybrid search
be/src/core/column/variant_v2/column_variant_v2.cpp
Other components
- SQL users and tools
- Lake and object data
Connections 18
- SQL users and tools → SQL Frontend: submit SQL
README.md - SQL Frontend → Authentication: authenticates
- SQL Frontend → Authorization context: checks access
fe/fe-core/src/main/java/org/apache/doris/mysql/privilege/EngineAuthorizationContext.java - SQL Frontend → Backend nodes: dispatches queries
README.md - Query cancellation → Backend nodes: checks backends
fe/fe-core/src/main/java/org/apache/doris/qe/QueryCancelWorker.java - Query profiles → SQL Frontend: serves profile views
fe/fe-core/src/main/java/org/apache/doris/httpv2/controller/QueryProfileController.java - Pipeline fragments → Task memory control: exposes task state
be/src/runtime/workload_management/query_task_controller.cpp - Pipeline fragments → Columnar data model: processes data
- Backend nodes → Pipeline fragments: executes fragments
README.md - Cloud storage engine → Cloud metadata: uses metadata
be/src/cloud/cloud_storage_engine.cpp - Cloud storage engine → Cloud compaction: starts compaction
be/src/cloud/cloud_storage_engine.cpp - Cloud storage engine → Object storage: uses storage
- Stream loading → Storage engine: writes data
- Real-time analytics → Stream loading: supports ingestion
README.md - Lakehouse analytics → Lake and object data: reads table data
fe/fe-connector/fe-connector-spi/src/main/java/org/apache/doris/connector/spi/ConnectorStorageContext.java - Hybrid search → SQL Frontend: supports SQL search
README.md - Lake and object data → Cloud storage engine: stores data
README.md - Storage engine → Backend nodes: serves queries
README.md