Frequency-Adaptive Join for Shared Nothing Machines - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Parallel and Distributed Computing Practices Année : 1999

Frequency-Adaptive Join for Shared Nothing Machines

Résumé

Although many skew-handling algorithms have been proposed for simple join operations, they remain generally inefficient in the case of $\theta$-join and in the case of multi-join. A new method for self-balancing equi-join operations on shared-nothing (SN) machines is proposed here. It offers deterministic and near-perfect load balancing through flexible control of communications in intra-transaction parallelism. The new algorithm mixes a balanced data-distribution strategy with pure hash-join. It is suitable for $\theta$-join operations and Its predictably low join-product- and attribute-value skews make it suitable for repeated use in multi-join operations. Its tradeoff between balancing overhead and speedup is analyzed in the BSP (Bulk-synchronous parallel) computing model. The scalable model predicts a negligible join product skew and a near-linear speed-up in any combination of selectivity, skew and number of processors. This prediction is confirmed by a series of tests.
Fichier non déposé

Dates et versions

hal-00081345 , version 1 (22-06-2006)

Identifiants

  • HAL Id : hal-00081345 , version 1

Citer

Mostafa Bamha, Gaetan Hains. Frequency-Adaptive Join for Shared Nothing Machines. Parallel and Distributed Computing Practices, 1999, 2, Number 3, p. 333-345. ⟨hal-00081345⟩
62 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More