Conceptual documentation explaining BenchBox architecture, terminology, and design principles.
- Architecture - System architecture and component overview
- Workflow - Benchmark execution workflow and lifecycle
- Data Model - Data structures and relationships
- Glossary - Terms and definitions
- Benchmarking Tools Compared - BenchBox vs HammerDB, BenchBase, and LakeBench
BenchBox is designed around a few key concepts:
Standard workloads that test database performance across different dimensions (analytics, transactions, data loading).
Database adapters that execute benchmarks on specific database systems (DuckDB, Databricks, Snowflake, etc.).
Different ways to run benchmarks:
- Power Test: Sequential query execution measuring single-user performance
- Throughput Test: Concurrent query streams measuring multi-user scalability
- Data Generation: Creating benchmark datasets at various scale factors
Structured output capturing execution metrics, query timings, and optional validation against expected results.
- Getting Started - Start using BenchBox
- Design Documents - Detailed design specifications
- Development Guide - Contributing to BenchBox
:maxdepth: 1
:caption: Concept Guides
:hidden:
architecture
workflow
data-model
glossary
benchmarking-tools-compared