On Parallel Hash-Join Processing with Skewed Data

スポンサーリンク

概要

• 論文の詳細を見る

• When data are uniformly distributed, parallel hash-join algorithm scales up well. However, the presence of data skew can cause load imbalances among the processors, significantly deteriorating its performance. Within the last years, there has been a growing interest in addressing the problem of join product skew, where the different join selectivitics on processors lead to an imbalance on the number of output tuples. Here we present a dynamic skew handling algorithm which detects and rebalances unexpected join product skews at run-time in a shared-nothing database environment.

• 一般社団法人情報処理学会の論文
• 1994-09-20

著者

• Kitsuregawa Masaru

  University Of Tokyo

• Harada Lilian

  Fujitsu Laboratories

関連論文

• Relational Algebra Machine GRACE
• Join Strategies on Grid-Files
• Join Strategies on Multi-Dimensional C1ustered Relations
• On Parallel Hash-Join Processing with Skewed Data
• Pipeline Stage Based Dynamic Load Balancing for Right-Deep Multi-Joins
• An Efficient Query Execution Plan for Multi-Way Joins in Shared-Nothing Database Environment
• Evaluation Results of Multi-Way Joins in Shared-Nothing Database Environment : The Case for Left-Deep and Right-Deep Execution Query Trees

スポンサーリンク