Scholarlink

<p>The graph partitioning problem is increasing with the emergence of Big Data. Handling tremendous volumes of graph data requires an efficient graph processing system and especially a high-quality graph partitioning approach(s) to cope with graph application needs. However, all graph partitioning algorithms do not consider graph data volumes during graph partitioning. As a result, graph processing systems experience an imbalance in their workload and a decrease in system performance. For this purpose, we designed our distributed framework for high-quality graph partitioning including the volume metric. Also, it is created for scalable, high-availability, and fault tolerance. Using real-world datasets, we show that VF-Hammer performs a good graph partitioning quality and achieves better performance results against state-of-the-art graph partitioning.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Distributed framework for high-quality graph partitioning

  • 期刊论文

  • 发表日期：

    2025年10月7日

查看全文
  • Chayma Sakouhi,
  • Abir Khaldi,
  • Henda Ben Ghezala

摘要

The graph partitioning problem is increasing with the emergence of Big Data. Handling tremendous volumes of graph data requires an efficient graph processing system and especially a high-quality graph partitioning approach(s) to cope with graph application needs. However, all graph partitioning algorithms do not consider graph data volumes during graph partitioning. As a result, graph processing systems experience an imbalance in their workload and a decrease in system performance. For this purpose, we designed our distributed framework for high-quality graph partitioning including the volume metric. Also, it is created for scalable, high-availability, and fault tolerance. Using real-world datasets, we show that VF-Hammer performs a good graph partitioning quality and achieves better performance results against state-of-the-art graph partitioning.