The Graph Data Council (GDC) works in the area of graph data management. GDC defines graph benchmarks and carries out research in graph schemas and query languages. GDC’s scope includes linked data standards (e.g., RDF and SPARQL), property graphs (e.g., the ISO GQL language), and relational database management systems (e.g., the SQL/PGQ extension of SQL:2023).
The organization was originally named the Linked Data Benchmark Council (LDBC). It now operates under the name, Graph Data Council (GDC), to reflect its focus on graph data management technology.
The GDC defines standard graph database benchmarks to accelerate the progress in the field, similar to how the TPC has created competition in the relational data management space. GDC is also active in promoting standard graph query languages such as GQL and SQL/PGQ. For a brief introduction, see the presentation on GDC and the accompanying paper.
GDC is run by the Board of Directors. The board frequently consults with the Policy Council, which consists of representatives from member companies and individual members.
GDC’s history starts with the Linked Data Benchmark Council, a research project under the European Union’s Seventh Framework Programme (FP7), which originally started in September 2012 with the participation of five academic institutions and three database vendors. The non-profit Linked Data Benchmark Council was established in 2013. In 2025, the organization changed its name to Graph Data Council.
The GDC is registered as the “Linked Data Benchmark Council” in the United Kingdom.
The GDC releases all of its software under the Apache Software License v2.
The GDC collects membership fees from its members and an auditing fee for supervising benchmark audits.
Members of the GDC publish papers at top academic conferences (e.g., SIGMOD and VLDB) and the GDC hosts Technical User Community meetings co-located with these conferences. Additionally, the GDC has a blog and is active on social media: Twitter/X and LinkedIn.
Membership is based on a calendar year, which runs from 1 January through to 31 December.
No. Your membership fee will be prorated to the number of full months remaining in the year.
Once your application form has been submitted it will be reviewed by the Board of Directors and the Policy Council, and you will be notified of the result. See Becoming a Member for details.
We aim to process your application within four weeks of receipt of the completed application form(s).
GQL (Graph Query Language) and SQL/PGQ (Property Graph Queries) are ISO-standard query languages to query graph data. The first version of SQL/PGQ was published in 2023, while GQL was published in 2024. The two languages share their pattern matching syntax.
No. GQL is a fully-fledged graph query language, while GraphQL is a query language to retrieve data from APIs (typically returned as JSON trees). GQL is an ISO standard, while GraphQL is maintained by the GraphQL Foundation, which is hosted under the Linux Foundation.
The GDC has a liaison with ISO, which allows members to access drafts of the GQL and SQL/PGQ standards documents. Several members have actively contributed to these standards.
Gábor Szárnyas maintains a list of pointers for GQL and SQL/PGQ with links to learning materials.
LDBC has four auditable benchmarks: the LDBC Financial Benchmark; the LDBC Semantic Publishing Benchmark; and the LDBC’s Social Network Benchmark, Interactive, and Business Intelligence workloads. Other benchmarks include LDBC Graphalytics, a graph analytics benchmark where submissions are run in a competition.
You can find them on the LDBC benchmark datasets site. They can be downloaded free of charge.
Getting your system audited allows you to disseminate your performance results as “LDBC Benchmark Results”, a term protected by trademark. The auditing process often benefits system developers from a technical perspective because preparing for an audit often uncovers performance bottlenecks in your system. Addressing these bottlenecks likely benefits customer workloads.
Please see the Auditing process for the LDBC Social Network Benchmark document, which details the process and its expected cost for the LDBC Social Network Benchmark. The process and the cost are similar for other LDBC benchmarks (e.g., FinBench, SPB).
The processes of GDC and TPC audits are similar. However, there are a few notable differences:
- TPC audits can be commissioned by any organization. GDC audits require the Test Sponsor to be a GDC member.
- A TPC audit can only target a single scale factor, while a GDC audit may involve multiple scale factors.
- For a TPC audit, the Test Sponsor has to produce the Full Disclosure Report of the benchmark run, which is accompanied by an attestation letter from a TPC-certified auditor. For a GDC audit, the auditor produces the Full Disclosure Report and signs the document electronically.
- TPC’s analytical benchmarks (e.g., TPC-H and TPC-DS) require audited systems to use SQL for implementing the queries, while GDC does not require the use of a standard query language.
Yes, but please study and apply the Trademark and Fair Use Policy for LDBC Benchmarks. In short, you have to state that your results are not official LDBC results.
Please consult the Process for defining new LDBC benchmarks document.