With the rapid evolution of AI, the application scenarios for graph data have become remarkably diverse and widespread. Iterative graph computing, Graph Neural Networks (GNN), and GraphRAG are driving growing demand for graph-structured data across industries. However, in most enterprises today, graph data remains siloed – hidden in log files, scattered across MySQL databases, or locked inside proprietary graph databases. When different scenarios require graph data to flow between systems, engineers face repeated ETL efforts, format incompatibilities, and significant friction. Meanwhile, a fundamental question arises: should we always load graph data into a transactional graph database just to serve diverse workloads? Doing so often leads to unnecessary data redundancy and operational overhead.

Fortunately, Apache GraphAr (Incubating) effectively addresses this challenge by providing an open graph storage format.In this talk, we will introduce GraphAr’s core philosophy of enabling seamless graph data flow across systems, and share recent collaborations with different open-source communities that put this vision into practice. We will also discuss the community operations of GraphAr within the Apache Incubator – how we build a diverse contributor base, practice the Apache Way, and plan the roadmap toward graduation. This session is designed for developers, data engineers, and open-source enthusiasts interested in graph technologies, data lake architecture, or community building under the Apache umbrella.

Speakers:

Xiaokang Yang: Apache GraphAr (Incubating) Committer

Xiaokang Yang is a Committer of Apache GraphAr (Incubating), an open-source project that provides a standardized file format and multi-language libraries for graph data storage in data lakes. He is currently a Master’s student at Northeastern University (China) and an intern at Alibaba, with a research focus on graph computing. He has authored two research papers in the graph domain. He is passionate about graph data management and building open-source communities following the Apache Way.

Zhe Wang: Apache GraphAr(incubating) PPMC

Master student at JLU, interested in graph store and graph AI.