Parallel input/output architectures for multiprocessors
Annapareddy, Narasimha Reddy Lakshmi
This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.
Permalink
https://hdl.handle.net/2142/22611
Description
Title
Parallel input/output architectures for multiprocessors
Author(s)
Annapareddy, Narasimha Reddy Lakshmi
Issue Date
1990
Doctoral Committee Chair(s)
Banerjee, Prithviraj
Department of Study
Electrical and Computer Engineering
Discipline
Electrical and Computer Engineering
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
Engineering, Electronics and Electrical
Computer Science
Language
eng
Abstract
This thesis looks at several aspects of solving the Input/Output problem. The increasing processor speeds and the emerging multiprocessor machines have improved the processing power of a system at a tremendous pace. This has not been followed by corresponding improvements in other parts of the system, mainly the I/O subsystem. This could lead to a huge imbalance in the system and, if not corrected, the gains made in the processing power will be of little use as the problem solving speed will be determined by the slower I/O operations.
A number of issues in solving the I/O problem are addressed in this thesis. The effects of I/O on problem solution are studied by tracing a number of applications on a multiprocessor. This study is aimed at gaining an understanding of how I/O accesses can be more efficiently done, how I/O systems need to be organized to effectively address the I/O problem and to explore compiler techniques to reduce the I/O problem. Since the I/O device speeds are not improving dramatically, the I/O power can significantly be improved only through parallel organizations. An extensive evaluation of these parallel organizations is presented. Different parallel I/O (disk) organizations are evaluated in different workloads. A methodology for interfacing I/O processors with a multiprocessor system is presented. This methodology is shown to have a number of advantages over the previous approaches. When data are distributed on multiple disks, fault tolerance becomes a problem. A new technique is presented for achieving fault tolerance in a multiple disk system. A number of compiler techniques are also proposed to reduce the I/O problem by suitable transformation of programs. A number of open problems in this area are highlighted.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
