Skip to content

[Harness] Implement generic benchmark engine supporting multiple triple stores and datasets #8

Description

@remiceres

Summary

Implement a generic, modular benchmark execution engine capable of running standardized loading and query workloads across multiple triple store implementations and datasets.

Objectives and Technical Scope

1. Modular Driver SPI

• Define a generic store adapter interface (e.g. TripleStoreDriver) specifying lifecycle methods: init(), load(datasetPath), query(sparqlQuery), clear(), close().
• Support automated metric collection before, during, and after each phase (time, heap memory, GC stats).

2. Scenario Runner and Result Export

• Build a CLI/runner executing test matrices: [Store] x [Dataset] x [Query Workload] x [Repetitions].
• Export standardized execution results in machine-readable JSON and CSV formats for subsequent analysis and dashboard rendering.

Acceptance Criteria

[ ] Generic driver SPI is defined and fully decoupled from any single triple store implementation.
[ ] Automated benchmark runner executes configurable benchmark matrices.
[ ] Metric collector accurately captures ingestion speed, query execution latencies, and JVM memory delta.
[ ] Results are exported in structured JSON format.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Labels

featureNew feature or capability

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions