首页> 外文会议>International conference on algorithms and architectures for parallel processing >Shedder: A Metadata Sharing Management Method across Multi-clusters
【24h】

Shedder: A Metadata Sharing Management Method across Multi-clusters

机译:Shedder:跨多集群的元数据共享管理方法

获取原文

摘要

With the increase of large number of data, multi-clusters structure has been adopted for data storage. Therefore, the demand for file metadata sharing across global scale clusters is quickly rising. However, today's software can not provide features to reflect these desires well. In this paper, We propose and develop a metadata sharing management method called Shedder. First Shedder can allow customized multi-clusters metadata sharing structure. Next, Shedder can provide highly efficient global synchronization for all clusters. Finally, Shedder allows customized user view generated from global namespace. Our evaluation for Shedder shows that Shedder provides low latency for global synchronization. Dynamic transformation from global namespace to customized user view also has low time cost for different size of workloads.
机译:随着大量数据的增加,多簇结构已经被用于数据存储。因此,对跨全球规模的群集共享文件元数据的需求正在迅速增长。但是,当今的软件无法提供能够很好地反映这些需求的功能。在本文中,我们提出并开发了一种称为Shedder的元数据共享管理方法。 First Shedder可以允许定制的多集群元数据共享结构。接下来,Shedder可以为所有群集提供高效的全局同步。最后,Shedder允许从全局名称空间生成自定义的用户视图。我们对Shedder的评估表明,Shedder为全局同步提供了低延迟。从全局名称空间到自定义用户视图的动态转换对于不同大小的工作负载也具有较低的时间成本。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号