The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
L'évolutivité et la disponibilité sont les caractéristiques clés des systèmes de bases de données parallèles. Pour réaliser l'évolutivité, de nombreuses méthodes d'équilibrage de charge dynamique avec placement de données et structures d'index parallèles sur une infrastructure parallèle sans partage ont été proposées. La migration des données avec un placement partitionné par plage à l'aide d'un Btree parallèle est une solution. La combinaison du partitionnement de plage et des réplicas déclustés en chaîne offre une haute disponibilité (HA) tout en préservant l'évolutivité. Cependant, le traitement indépendant des données principales et de sauvegarde dans chaque nœud nécessite des temps de basculement longs. Nous proposons une nouvelle méthode pour le traitement composé de répliques chaînées dégroupées à l'aide d'un Btree parallèle, appelé Fat-Btree. Dans la méthode proposée, un seul Fat-Btree fournit des chemins d'accès aux données principales et de sauvegarde de tous les éléments de processeur (PE), ce qui réduit considérablement le temps de basculement. De plus, ces chemins d'accès se chevauchent entre deux PE voisins, ce qui permet un équilibrage dynamique de la charge sans migration physique des données en redirigeant dynamiquement les chemins d'accès. De plus, ce traitement composé améliore l'utilisation de l'espace mémoire pour permettre un traitement d'index avec une bonne évolutivité. Des expériences utilisant PostgreSQL sur un cluster PC de 160 nœuds démontrent l'efficacité de la haute évolutivité et de la disponibilité de notre méthode proposée.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Min LUO, Akitsugu WATANABE, Haruo YOKOTA, "A Compound Parallel Btree for High Scalability and Availability on Chained Declustering Parallel Systems" in IEICE TRANSACTIONS on Information,
vol. E94-D, no. 3, pp. 587-601, March 2011, doi: 10.1587/transinf.E94.D.587.
Abstract: Scalability and availability are the key features of parallel database systems. To realize scalability, many dynamic load-balancing methods with data placement and parallel index structures on shared-nothing parallel infrastructure have been proposed. Data migration with range-partitioned placement using a parallel Btree is one solution. The combination of range partitioning and chained declustered replicas provides high availability (HA) while preserving scalability. However, independent treatment of the primary and backup data in each node requires long failover times. We propose a novel method for the compound treatment of chained declustered replicas using a parallel Btree, termed the Fat-Btree. In the proposed method, a single Fat-Btree provides access paths to both the primary and backup data of all processor elements (PEs), which greatly reduces failover time. Moreover, these access paths overlap between two neighboring PEs, which enables dynamic load balancing without physical data migration by dynamically redirecting the access paths. In addition, this compound treatment improves memory space utilization to enable index processing with good scalability. Experiments using PostgreSQL on a 160-node PC cluster demonstrate the effectiveness of the high scalability and availability of our proposed method.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E94.D.587/_p
Copier
@ARTICLE{e94-d_3_587,
author={Min LUO, Akitsugu WATANABE, Haruo YOKOTA, },
journal={IEICE TRANSACTIONS on Information},
title={A Compound Parallel Btree for High Scalability and Availability on Chained Declustering Parallel Systems},
year={2011},
volume={E94-D},
number={3},
pages={587-601},
abstract={Scalability and availability are the key features of parallel database systems. To realize scalability, many dynamic load-balancing methods with data placement and parallel index structures on shared-nothing parallel infrastructure have been proposed. Data migration with range-partitioned placement using a parallel Btree is one solution. The combination of range partitioning and chained declustered replicas provides high availability (HA) while preserving scalability. However, independent treatment of the primary and backup data in each node requires long failover times. We propose a novel method for the compound treatment of chained declustered replicas using a parallel Btree, termed the Fat-Btree. In the proposed method, a single Fat-Btree provides access paths to both the primary and backup data of all processor elements (PEs), which greatly reduces failover time. Moreover, these access paths overlap between two neighboring PEs, which enables dynamic load balancing without physical data migration by dynamically redirecting the access paths. In addition, this compound treatment improves memory space utilization to enable index processing with good scalability. Experiments using PostgreSQL on a 160-node PC cluster demonstrate the effectiveness of the high scalability and availability of our proposed method.},
keywords={},
doi={10.1587/transinf.E94.D.587},
ISSN={1745-1361},
month={March},}
Copier
TY - JOUR
TI - A Compound Parallel Btree for High Scalability and Availability on Chained Declustering Parallel Systems
T2 - IEICE TRANSACTIONS on Information
SP - 587
EP - 601
AU - Min LUO
AU - Akitsugu WATANABE
AU - Haruo YOKOTA
PY - 2011
DO - 10.1587/transinf.E94.D.587
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E94-D
IS - 3
JA - IEICE TRANSACTIONS on Information
Y1 - March 2011
AB - Scalability and availability are the key features of parallel database systems. To realize scalability, many dynamic load-balancing methods with data placement and parallel index structures on shared-nothing parallel infrastructure have been proposed. Data migration with range-partitioned placement using a parallel Btree is one solution. The combination of range partitioning and chained declustered replicas provides high availability (HA) while preserving scalability. However, independent treatment of the primary and backup data in each node requires long failover times. We propose a novel method for the compound treatment of chained declustered replicas using a parallel Btree, termed the Fat-Btree. In the proposed method, a single Fat-Btree provides access paths to both the primary and backup data of all processor elements (PEs), which greatly reduces failover time. Moreover, these access paths overlap between two neighboring PEs, which enables dynamic load balancing without physical data migration by dynamically redirecting the access paths. In addition, this compound treatment improves memory space utilization to enable index processing with good scalability. Experiments using PostgreSQL on a 160-node PC cluster demonstrate the effectiveness of the high scalability and availability of our proposed method.
ER -