A heuristic approach to file placement in distributed systems

In the Distributed Computing System environment, an efficient and effective means of resource allocation and monitoring is difficult to achieve. A highly available and reliable system rarely exist. Since each processing element is independent from each other and can be geographically distributed, th...

Full description

Saved in:
Bibliographic Details
Main Author: Huang, En-Hsin
Format: text
Language:English
Published: Animo Repository 1991
Subjects:
Online Access:https://animorepository.dlsu.edu.ph/etd_masteral/1370
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: De La Salle University
Language: English
id oai:animorepository.dlsu.edu.ph:etd_masteral-8208
record_format eprints
spelling oai:animorepository.dlsu.edu.ph:etd_masteral-82082021-02-03T03:17:18Z A heuristic approach to file placement in distributed systems Huang, En-Hsin In the Distributed Computing System environment, an efficient and effective means of resource allocation and monitoring is difficult to achieve. A highly available and reliable system rarely exist. Since each processing element is independent from each other and can be geographically distributed, the redundancy of the tasks and resources are permissible and widely observed. It is therefore important to let the processing element access the nearest available resources so as to minimize communication cost. When physical link failure occurs, system could consequently be partitioned and resources could be inaccessible to the processing elements. It is therefore important to locate the still possible accessible resources to ensure the successful execution of processing elements. This paper presents a conceptual model with algorithms for computing a near optimal and a reliable solution and performing reliability analysis of distributed task and distributed system based on graph-theoretic formalism, matrix and set operation. Formal proof of correctness of algorithm is conducted and sensitivity analysis is performed to examine various relationship among availability, degree of duplication and network topologies in the Distributed Systems environment. First, a model is constructed to represent the respective types of topologies, and the initial maximum degree of redundancy of data file is determined. A set of feasible allocation with corresponding query cost is computed. To determine the distributed task reliability and the distributed system reliability of the computed solutions, a set of minimum spanning tree is constructed and a terminal algorithm is applied. With the algorithm presented here, the ability for quick access to resources and maintenance of the system's reliability will greatly improve the overall system performance. 1991-01-01T08:00:00Z text https://animorepository.dlsu.edu.ph/etd_masteral/1370 Master's Theses English Animo Repository Electronic data processing -- Distributed processing Algorithms Heuristic programming Computer Sciences
institution De La Salle University
building De La Salle University Library
continent Asia
country Philippines
Philippines
content_provider De La Salle University Library
collection DLSU Institutional Repository
language English
topic Electronic data processing -- Distributed processing
Algorithms
Heuristic programming
Computer Sciences
spellingShingle Electronic data processing -- Distributed processing
Algorithms
Heuristic programming
Computer Sciences
Huang, En-Hsin
A heuristic approach to file placement in distributed systems
description In the Distributed Computing System environment, an efficient and effective means of resource allocation and monitoring is difficult to achieve. A highly available and reliable system rarely exist. Since each processing element is independent from each other and can be geographically distributed, the redundancy of the tasks and resources are permissible and widely observed. It is therefore important to let the processing element access the nearest available resources so as to minimize communication cost. When physical link failure occurs, system could consequently be partitioned and resources could be inaccessible to the processing elements. It is therefore important to locate the still possible accessible resources to ensure the successful execution of processing elements. This paper presents a conceptual model with algorithms for computing a near optimal and a reliable solution and performing reliability analysis of distributed task and distributed system based on graph-theoretic formalism, matrix and set operation. Formal proof of correctness of algorithm is conducted and sensitivity analysis is performed to examine various relationship among availability, degree of duplication and network topologies in the Distributed Systems environment. First, a model is constructed to represent the respective types of topologies, and the initial maximum degree of redundancy of data file is determined. A set of feasible allocation with corresponding query cost is computed. To determine the distributed task reliability and the distributed system reliability of the computed solutions, a set of minimum spanning tree is constructed and a terminal algorithm is applied. With the algorithm presented here, the ability for quick access to resources and maintenance of the system's reliability will greatly improve the overall system performance.
format text
author Huang, En-Hsin
author_facet Huang, En-Hsin
author_sort Huang, En-Hsin
title A heuristic approach to file placement in distributed systems
title_short A heuristic approach to file placement in distributed systems
title_full A heuristic approach to file placement in distributed systems
title_fullStr A heuristic approach to file placement in distributed systems
title_full_unstemmed A heuristic approach to file placement in distributed systems
title_sort heuristic approach to file placement in distributed systems
publisher Animo Repository
publishDate 1991
url https://animorepository.dlsu.edu.ph/etd_masteral/1370
_version_ 1712574911549014016