CedarDB Platform

CLI name: cedardb - use benchbox run --platform cedardb

CedarDB (formerly Umbra, from TU Munich) is a high-performance standalone RDBMS supporting both OLAP and OLTP workloads over the PostgreSQL wire protocol. BenchBox drives CedarDB through its PostgreSQL adapter, with OLAP-tuned session defaults and standard COPY for bulk loading.

Features

• HTAP engine - OLAP and OLTP on the same storage, compiled query execution
• PostgreSQL-compatible - standard Postgres drivers (psycopg), tools (psql, DBeaver, DataGrip)
• Standalone - not a Postgres extension; its own engine and storage
• Bulk loading - standard PostgreSQL COPY is fully supported
• OLAP-tuned defaults - session-level settings applied at connection time

Deployment Modes

Mode	Description
self-hosted	Self-hosted CedarDB server over PG wire protocol (default)

benchbox run --platform cedardb --benchmark tpch --scale 1.0

This uses the self-hosted mode, which is the default.

Installation

CedarDB uses the same Python driver as PostgreSQL:

uv add benchbox --extra postgresql

Run a local CedarDB server (see CedarDB docs for install options). Default PG-wire port is 5432.

Configuration

Connection parameters mirror PostgreSQL and are passed with --platform-option:

benchbox run --platform cedardb --benchmark tpch --scale 1.0 \
  --platform-option host=localhost \
  --platform-option port=5432 \
  --platform-option username=postgres \
  --platform-option database=benchbox

Unset options fall back to host=localhost, port=5432, username=postgres, and schema=public. There is no dedicated CEDARDB_* environment-variable surface.

Supported Benchmarks

CedarDB inherits the PostgreSQL dialect, so it can in principle run any BenchBox benchmark supported on PostgreSQL - TPC-H, TPC-DS, SSB, Join Order, Read Primitives, Write Primitives, Transaction Primitives, and similar. Individual benchmark/tuning combinations require validation; partitioning and clustering tuning are not yet verified.

Notes

• Data types, functions, and COPY syntax follow PostgreSQL. OLAP session defaults (planner GUCs, page-cost tuning) are applied automatically during configure_for_benchmark() - no CLI flag is required.
• CedarDB is positioned as an HTAP engine - transactional workloads such as Transaction Primitives will also run.
• For large scale factors, prefer --compression zstd on staged data files to speed up COPY.

See Also

• PostgreSQL Platform - parent dialect and driver
• pg_duckdb, pg_mooncake - Postgres-compatible analytics alternatives