Content area

Abstract

With the increasing complexity of computer-based applications, efficient access to high volumes of data became a necessity. Databases with tens or hundreds of Gigabytes of items are common, and applications reaching Terabyte volumes will not be unusual. Parallel Database Systems are expected to provide the required support to data intensive applications. There are, however, limits on the achievable scalability of current parallel database algorithms and topologies.

This work addresses such limitations, with emphasis to architectural features and the problem of data skew on parallel database systems. We propose an architecture which extends the features of the shared-nothing architecture, widely adopted for current parallel database applications. We also propose a new characterization of data skew which captures distinct types of imbalance, and present two data partitioning strategies to deal with this problem in a parallel system. These strategies are employed as components of parallel algorithms to implement external sorting and relational database operations. The execution time of such algorithms, obtained from analytical models derived from their implementations, is used as a measure to evaluate the proposed architecture.

Details

1010268
Identifier / keyword
Title
Performance and scalability of parallel database systems
Number of pages
306
Degree date
1994
School code
0117
Source
DAI-B 55/10, Dissertation Abstracts International
ISBN
979-8-209-09574-3
University/institution
University of Maryland, College Park
University location
United States -- Maryland
Degree
Ph.D.
Source type
Dissertation or Thesis
Language
English
Document type
Dissertation/Thesis
Dissertation/thesis number
9508040
ProQuest document ID
304093627
Document URL
https://www.proquest.com/dissertations-theses/performance-scalability-parallel-database-systems/docview/304093627/se-2?accountid=208611
Copyright
Database copyright ProQuest LLC; ProQuest does not claim copyright in the individual underlying works.
Database
ProQuest One Academic