LDBC SNB Documentation

[latest PDF (2.2.4-SNAPSHOT)] [stable PDF (2.2.3)]

Benchmark specification

For a quick overview of LDBC SNB, start with our presentation (PDF).

For a guide on how to develop SNB Interactive implementations, please check out the READMEs of the reference implementation repositories linked below.

Software components

The SNB workloads are stored in different repositories:

• SNB Interactive v1:
  • Data generator: https://github.com/ldbc/ldbc_snb_datagen_hadoop
  • Driver: https://github.com/ldbc/ldbc_snb_interactive_v1_driver
  • Reference implementations: https://github.com/ldbc/ldbc_snb_interactive_v1_impls
• SNB Interactive v2:
  • Data generator: https://github.com/ldbc/ldbc_snb_datagen_spark
  • Driver: https://github.com/ldbc/ldbc_snb_interactive_v2_driver
  • Reference implementations: https://github.com/ldbc/ldbc_snb_interactive_v2_impls
• SNB Business Intelligence (BI):
  • Data generator: https://github.com/ldbc/ldbc_snb_datagen_spark
  • Driver and reference implementations: https://github.com/ldbc/ldbc_snb_bi

How to cite LDBC benchmarks

• Social Network Benchmark:
  • Detailed specification: The LDBC Social Network Benchmark by the LDBC Social Network Benchmark task force and contributors, arXiv/CoRR abs/2001.02299, 2020. [bib]
  • BI workload: An early look at the LDBC Social Network Benchmark's Business Intelligence workload, GRADES-NDA at SIGMOD 2018 by G. Szárnyas et al. [bib]
  • Interactive workload: The LDBC Social Network Benchmark: Interactive Workload, SIGMOD 2015 by O. Erling et al. [bib]
• Related benchmarks:
  • LDBC Graphalytics: LDBC Graphalytics: A Benchmark for Large-Scale Graph Analysis on Parallel and Distributed Platforms, VLDB 2016 paper by A. Iosup et al. [bib], The LDBC Graphalytics Benchmark, technical report [bib]
  • LDBC Semantic Publishing Benchmark: Benchmarking RDF Query Engines: The LDBC Semantic Publishing Benchmark, BLINK at ISWC 2016 by V. Kotsev et al. [bib]
  • LSQB (Labelled Subgraph Query Benchmark): a microbenchmark focusing on subgraph queries (graph pattern matching) using labelled graphs produced by the LDBC data generator. [bib]

How to build the this document

This repository contains the LaTeX source for the specification of the LDBC Social Network Benchmark. This README discusses how to build the specification PDF from source.

Generating query cards

To get consistent formatting, query cards are generated from query specifications defined in YAML format. This is a necessary step to compile to the document.

Install Pandoc, Python, and the required packages:

scripts/install-dependencies.sh

Building the document

To build the document locally, run make or make texfot. The latter requires Perl but produces a cleaner output.

We also provide a GitHub Action repository and a Docker container and images on Docker Hub. To use this locally, run:

docker run --rm --volume=`pwd`:"/github/workspace" ldbc/document-builder:2021 texfot query_cards workloads && sudo chown -R ${USER}:${USER} .