检索结果-内蒙古大学图书馆

parallel group-by query processing in a cluster architecture

COMPUTER SYSTEMS SCIENCE AND ENGINEERING 2002年第1期17卷 23-39页

作者： Taniar, D Rahayu, JW Monash Univ Sch Business Syst Clayton Vic 3800 Australia La Trobe Univ Dept Comp Sci & Comp Engn Bundoora Vic 3083 Australia

SQL queries in the real world are replete with group-by clauses. These queries are often used for strategic decision making due to the nature of group-by queries where raw information is grouped according to the designated groups and within each group aggregate functions are normally carried out. In strategic decision making, parallelization of these queries becomes necessary to speed up query response time. The trend in parallel computing is now shifted to cluster architectures. In this paper, we propose a technique for parallelization of group-by queries in cluster architectures. The need for a specialized algorithm on cluster architectures is raised due to the fact that most existing work on parallel group-by query processing is based on a shared-nothing architecture, and furthermore, they cannot be efficiently implemented in cluster platforms.

关键词： group-by queries parallel query processing parallel aggregate query processing parallel databases cluster architectures parallel database architectures

来源：评论

学校读者我要写书评

暂无评论

parallel SORT-HASH OBJECT-ORIENTED COLLECTION JOIN ALGORITHMS FOR SHARED-MEMORY MACHINES

引用

parallel Algorithms and Applications 2002年第2期17卷 85-126页

作者： David Taniar - E-mail: David.Taniar@infotech.monash.edu.au[a] J. Wenny Rahayu[b] [a] School of Business Systems Monash University Clayton Vic Australia [b] Department of Computer Science and Computer Engineering La Trobe University Bundarra Vic Australia

Collection join queries are join queries based on collection attributes (i.e. non-atomic attributes), which are common in object-oriented databases. We have identified three different kinds of collection join queries, namely;cullection-equijoin,collection-intersectjoin, andsub-collectionjoin. In this paper, we propose parallel join algorithms for these three collection join query types based on a combination of sort and hash methods, which we callparallel sort-hash, collection join algorithms. The proposed join algorithms play an important role in parallel object-oriented query processing, due to their superiority over the conventional join methods which are usually in a form of relational division, and also the inefficiency of the original join predicate processing. In our implementation of these algorithms on a shared-memory machine, we show that the combination between sort and hash methods is proven to be better than the conventional sort-merge and nested-loop based parallel join processing

关键词： parallel query processing parallel join Object-oriented join queries Collection types Object-oriented databases parallel databases

来源：评论

学校读者我要写书评

暂无评论

Dynamic memory allocation strategies for parallel query execution 02

Dynamic memory allocation strategies for parallel query exec...

引用

Proceedings of the 2002 ACM symposium on Applied computing

作者： Franck Morvan Abdelkader Hameurlain Université Paul Sabatier 118 route de Narbonne 31062 Toulouse cedex France

ISBN: (纸本)9781581134452

In the decision support queries which manipulate large data volumes, it is frequent that a query constituted by several joins can not be computed completely in memory. In this paper, we propose three strategies allowing to assign the memory of a shared-nothing parallel architecture to operation clones of a query. The performance evaluation of the three strategies shows that the strategies which favor the operation clones using a lot of memory obtain a better response time than the strategy which favors the clones using little memory. The main contribution of this paper is to take into account the available memory sizes on every processor and to avoid allotting the same processor to two operation clones that must run in parallel.

关键词： performance evaluation optimization parallelism dynamic memory allocation parallel databases

来源：评论

学校读者我要写书评

暂无评论

An adaptive algorithm for mining association rules on shared-memory parallel machines

引用

DISTRIBUTED AND parallel databases 2001年第2期9卷 99-132页

作者： Cheung, DW Hu, K Xia, SW Univ Hong Kong Dept Comp Sci & Informat Syst Hong Kong Hong Kong Peoples R China Tsinghua Univ Dept Automat Beijing 100084 Peoples R China

Mining association rules from large databases is very costly. We propose to develop parallel algorithms for this task on shared-memory multiprocessor (SMP). All proposed parallel algorithms for other paradigms follow the conventional level-wise approach: they need as many iterations as the length of the maximum large itemset. To make matter worse, they impose a synchronization in every iteration which would cause serious I/O contention on shared-memory parallel system. An adaptive asynchronous parallel mining algorithm APM has been proposed for SMP. All processors generate candidates dynamically and count itemset supports independently without synchronization. Two optimization techniques have been proposed for the reduction of database scanning and the number of candidates. The algorithm APM has been implemented on a Sun Enterprise 4000 shared-memory multiprocessor with 12 nodes. The experiments show that the optimizations have very good effects and APM has a substantial lead in performance over other proposed level-wise algorithms.

关键词： data mining parallel databases association rules parallel mining parallel computing shared-memory multi-processors

来源：评论

学校读者我要写书评

暂无评论

A hypergraph based approach to declustering problems

引用

DISTRIBUTED AND parallel databases 2001年第3期10卷 269-288页

作者： Liu, DR Wu, MY Natl Chiao Tung Univ Inst Informat Management Hsinchu Taiwan

parallelizing I/O operations via effective declustering of data is becoming essential to scale up the performance of parallel databases or high performance systems. Declustering has been shown to be a NP-complete problem in some contexts. Some heuristic methods have been proposed to solve this problem. However, most methods are not effective in several cases such as queries with different access frequencies or data with different sizes. In this paper, we propose a hypergraph model to formulate the declustering problem. Several interesting theoretical results are achieved by analyzing the proposed model. The proposed approach will allow modeling a wide range of declustering problems. Furthermore, the hypergraph declustering model is used as the basis to develop new heuristic methods, including a greedy method and a hybrid declustering method. Experiments show that the proposed methods can achieve better performance than several declustering methods.

关键词： declustering data allocation parallel databases hypergraph high performance systems

来源：评论

学校读者我要写书评

暂无评论

Analysis and comparison of declustering schemes for interactive navigation queries

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2000年第5期12卷 763-778页

作者： Chen, CM Sinha, RK Telcordia Technol Morristown NJ 07960 USA Ciena Corp Cupertino CA 95014 USA

Declustering schemes enable parallel retrievals of raster-spatial data by allocating data among multiple disks. Previous work in the literature has focused on static range queries. In this paper, we focus on interactive navigation queries, which exhibit a new class of data access patterns. We analyze and compare the performance of previously known declustering schemes from the perspective of navigation queries. These schemes include a class of latin-square-based schemes, Disk Modulo, Fieldwise Exclusive-OR, Hilbert Curve Allocation Method. and a random assignment scheme. We show that, unlike the case of range queries, disk module outperforms other schemes and gives nearly optimal performance for navigation queries. In addition, we propose a new scheme under the constraint of bounded window size-a common constraint in practice due to resource limitation such as monitor resolution or memory size. Through extensive analysis, we establish guidelines on how these schemes can be tuned to provide guaranteed performance under the constraint of bounded window size.

关键词： declustering navigation queries parallel databases high performance spatial databases

来源：评论

学校读者我要写书评

暂无评论

Workfile disk management for concurrent mergesorts in a multiprocessor database system

引用

DISTRIBUTED AND parallel databases 2000年第3期8卷 279-296页

作者： Wu, KL Yu, PS Chung, JY Teng, JZ IBM Corp Thomas J Watson Res Ctr Yorktown Hts NY 10598 USA IBM Corp Programming Syst San Jose CA 95161 USA

This paper studies workfile disk management for concurrent mergesorts ina multiprocessor database system. Specifically, we examine the impacts of workfile disk allocation and data striping on the average mergesort response time. Concurrent mergesorts in a multiprocessor system can creat severe I/O interference in which a large number of sequential write requests are continuously issued to the same workfile disk and block other read requests for a long period of time. We examine through detailed simulations a logical partitioning approach to workfile disk management and evaluate the effectiveness of datastriping. The results show that (1) without data striping, the best performance is achieved by using the entire workfile disks as a single partition if there are abundant workfile disks (or system workload is light);(2) however, if there are limited workfile disks (or system workload is heavy), the workfile disks should be partitioned into multiple groups and the optimal partition size is workload dependent;(3) data striping is beneficial only if the striping unit size is properly chosen;and (4) with a proper striping size, the best performance is generally achieved by using the entire disks as a single logical partition.

关键词： parallel databases workfile disk allocation I/O parallelism prefetching concurrent mergesorts data striping

来源：评论

学校读者我要写书评

暂无评论

Load balancing for parallel query execution on NUMA multiprocessors

引用

DISTRIBUTED AND parallel databases 1999年第1期7卷 99-121页

作者： Bouganim, L Florescu, D Valduriez, P INRIA Rocquencourt France

To scale up to high-end configurations, shared-memory multiprocessors are evolving towards Non Uniform Memory Access (NUMA) architectures. In this paper, we address the central problem of load balancing during parallel query execution in NUMA multiprocessors. We first show that an execution model for NUMA should not use data partitioning (as shared-nothing systems do) but should strive to exploit efficient shared-memory strategies like Synchronous Pipelining (SP). However, SP has problems in NUMA, especially with skewed data. Thus, we propose a new execution strategy which solves these problems. The basic idea is to allow partial materialization of intermediate results and to make them progressivly public, i.e., able to be processed by any processor, as needed to avoid processor idle times. Hence, we call this strategy Progressive Sharing (PS). We conducted a performance comparison using an implementation of SP and PS on a 72-processor KSR1 computer, with many queries and large relations. With no skew, SP and PS have both linear speed-up. However, the impact of skew is very severe on SP performance while it is insignificant on PS. Finally, we show that, in NUMA, PS can also be beneficial in executing several pipeline chains concurrently.

关键词： parallel databases query execution load balancing NUMA synchronous pipeline execution engines

来源：评论

学校读者我要写书评

暂无评论

Optical implementation of a constant-time multicomparand bit-parallel magnitude-comparison algorithm using wavelength- and polarization-division multiplexing with application to parallel database processing

引用

OPTICS LETTERS 1998年第17期23卷 1372-1374页

作者： Detofsky, A Choo, PY Louri, A Univ Arizona Dept Elect & Comp Engn Tucson AZ 85721 USA

We present a word- and bit-parallel magnitude-comparison architecture that permits multiple comparands to be compared with multiple relations in constant time. The proposed magnitude-comparison algorithm uses a novel polarization and wavelength-encoding scheme to achieve a fast, scalable realization. Distinctive features of the proposed architecture include (1) the use of a multiple-wavelength encoding scheme to increase processing parallelism and (2) multiple-comparand word- and bit-parallel comparison with an execution time that is independent of the data or word size. (C) 1998 Optical Society of America.

关键词： wavelengths bit parallel parallel databases execution time MULTIPLEXOR

来源：评论

学校读者我要写书评

暂无评论

Multiattribute indexing with m-Q-trees 6

Multiattribute indexing with m-Q-trees

引用

6th Euromicro Workshop on parallel and Distributed Processing (PDP 98)

作者： Polo, A Manzano, JC Salas, M Sanchez, M Jurado, E Barrena, M Hernandez, J Martinez, JM Univ Extremadura Dept Comp Sci E-06071 Badajoz Spain

ISBN: (纸本)0818683325

A wide class of multidimensional indexes employs a recursive partitioning of the data space as the kd-tree does. In this paper we present the m-Q-tree as a multidimensional data structure that can achieve the maximum degree of 2(m) children in every node (where m is the number of index attributes) and a maximum of only one underflow page per node. We describe the m-Q-tree, and give searching and inserting algorithms. In order to develop a solution for building the in-Q-tree, we define and use a conceptual tool, called prefix graph, which permits us to manage the regions associated to all sons of every node. The proposed algorithm is of order O(m). Finally, we present the results of a series of rests which indicate that the structure performs well. m-Q-tree gives a general technique for declustering data in a parallel database. We propose m-Q-tree as a new general access method which permits the exploitation of the potential parallelism of all relational operations, in addition to favour the execution of complex queries, including different kinds of conditions over several attributes for one or more relations.

关键词： parallel databases multidimensional data structures multiattribute indexes range queries spatial databases

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：