TuGraph-DB v4.0: New Features Including ISO GQL Support, Enterprise High Availability, and Graph Learning Engine

At the 2022 World AI Conference, Ant Group announced the open‑source release of its self‑developed graph database TuGraph‑DB, and on the first anniversary of its open‑source status, TuGraph‑DB v4.0 was launched after ten major iterations.

The new version adds cloud deployment, a Procedure On Graph (POG) query language, Python algorithm interfaces, a graph learning engine, high‑availability features, and support for the ISO GQL standard, improving completeness, usability, and advanced design.

1. ISO GQL

Just as SQL standardized relational databases, GQL (Graph Query Language) is the international standard for graph queries, acting as the “SQL” of the graph database world. TuGraph‑DB v4.0 follows the latest GQL standard, offering users a rich set of query language options and promoting standardization across the graph ecosystem.

2. Enterprise‑grade High Availability

TuGraph‑DB now provides enterprise‑level high availability with multi‑active hot‑standby and built‑in load balancing, allowing higher read loads without extra configuration. In a cluster, one leader and multiple followers serve reads and writes; if any node fails, the RAFT protocol switches the leader within seconds, ensuring RPO = 0 and RTO < 10 seconds. Future releases will add roles such as Witness and Learner.

3. Graph Learning Engine

v4.0 deeply integrates a graph learning engine compatible with popular frameworks like DGL and PyG, offering real‑time and snapshot‑based graph sampling operators that share the same storage as the query engine. Users can train graph learning models directly on the database, leveraging disk‑based storage to handle massive graphs on a single machine, reducing deployment costs compared to traditional big‑data solutions.

Conclusion

Over the past year, TuGraph‑DB has enhanced both interface richness and learning engine support, striving for better ease‑of‑use and feature completeness. Future plans include continued architectural evolution and a more vibrant graph ecosystem.

Community members are invited to contribute code, discuss technical topics, share insights, or report issues. The project’s source code, documentation, quick‑start guide, and community channels are listed below.

Open‑source repository: https://github.com/tugraph-family/tugraph-db

Official documentation: https://tugraph-db.readthedocs.io

One‑click trial: https://aliyun-computenest.github.io/quickstart-tugrap