Login

Publications  •  Project Statistics

Glossary  •  Schools  •  Disciplines
People Search: 
   
Title/Abstract Search: 

Dissertation Information for Frank DeRobertis

NAME:
- Frank DeRobertis

DEGREE:
- Ph.D.

DISCIPLINE:
- Library and Information Science

SCHOOL:
- SUNY, Albany (USA) (2003)

ADVISORS:
- Donald Ballou
- Thomas J. Galvin

COMMITTEE MEMBERS:
- Lakshmi Mohan
- Neil Vincent Murray

MPACT Status: Fully Complete

Title: Parallelizing object oriented queries with networked PC's: An exploratory study

Abstract: The purpose of the study is to determine if a Microsoft Access query could be generated from an Object Oriented (OO) query, executed in a parallel processing personal computer networked environment, and to what extent the response time is affected by the techniques as compared to a non-parallel environment.

This study tests the algorithm that transforms OO queries into Access SQL queries using databases on three networked PCs. The databases were formatted across the three machines using horizontal data mapping techniques. Each query was selected and executed automatically five times to develop an average response time for each query. The baseline response time was captured on a single, non-horizontally mapped database. The average response times of the parallel queries were compared against the baseline results.

Results show that Access queries can be created from OO queries and that response time is impacted to some degree because of extra processing, system anomalies, the code to handle the transformation of the queries and the mapping of the data. Information Technology organizations using OO techniques on large systems and want to move to smaller, less expensive PC environments to take advantage of parallel processing techniques can use this approach. Careful mapping of the relational data to promote parallel processing of OO queries in a personal computer environment is required.

The development and design focused around the ability to create an algorithm that would be able to identify OO queries, parse the query into logical sub queries and then send these queries to selected machines for execution. The study showed that some queries performed as expected while others did not perform as well when compared to the baseline results.

Practitioners can develop strategies for building a parallel PC environment supported by the algorithm and techniques developed in this research. Recommendation for future research include studies to determine the effect of vertical data mapping and whether update capabilities will have an adverse affect on the algorithm and therefore, on the overall system throughput and response time.

MPACT Scores for Frank DeRobertis

A = 0
C = 0
A+C = 0
T = 0
G = 0
W = 0
TD = 0
TA = 0
calculated 2008-01-31 06:13:32

Advisors and Advisees Graph