Distributed Graph Generation Using JDM Replication

Graph substructures play a critical role in defining network characteristics, making it essential for graph generators to accurately replicate these substructures to produce realistic graphs. However, the increasing rate of data in networks poses significant challenges for graph generation. To address this, a novel distributed algorithm is proposed, which employs JDM (Joint Degree Matrix) replication across distributed servers. This approach utilizes the vertex cut partitioning across multiple commodity computers. Experiments are conducted to evaluate the algorithm's runtime performance, demonstrating its effectiveness in handling large-scale networks while preserving realistic substructures. The results highlight the potential of this method to meet the growing demands of scalable and efficient graph generation.