Cardiff University | Prifysgol Caerdydd ORCA
Online Research @ Cardiff 
WelshClear Cookie - decide language by browser settings

Quality of service assessment over multiple attributes

Al-Dossari, Hmood Zafer 2011. Quality of service assessment over multiple attributes. PhD Thesis, Cardiff University.

[thumbnail of U585497.pdf_no sigs.pdf]
Preview
PDF - Accepted Post-Print Version
Download (14MB) | Preview

Abstract

The development of the Internet and World Wide Web have led to many services being offered electronically. When there is sufficient demand from consumers for a certain service, multiple providers may exist, each offering identical service functionality but with varying qualities. It is desirable therefore that we are able to assess the quality of a service (QoS), so that service consumers can be given additional guidance in se lecting their preferred services. Various methods have been proposed to assess QoS using the data collected by monitoring tools, but they do not deal with multiple QoS attributes adequately. Typically these methods assume that the quality of a service may be assessed by first assessing the quality level delivered by each of its attributes individ ually, and then aggregating these in some way to give an overall verdict for the service. These methods, however, do not consider interaction among the multiple attributes of a service when some packaging of qualities exist (i.e. multiple levels of quality over multiple attributes for the same service). In this thesis, we propose a method that can give a better prediction in assessing QoS over multiple attributes, especially when the qualities of these attributes are monitored asynchronously. We do so by assessing QoS attributes collectively rather than indi vidually and employ a k nearest neighbour based technique to deal with asynchronous data. To quantify the confidence of a QoS assessment, we present a probabilistic model that integrates two reliability measures: the number of QoS data items used in the as sessment and the variation of data in this dataset. Our empirical evaluation shows that the new method is able to give a better prediction over multiple attributes, and thus provides better guidance for consumers in selecting their preferred services than the existing methods do.

Item Type: Thesis (PhD)
Status: Unpublished
Schools: Computer Science & Informatics
Subjects: Q Science > QA Mathematics > QA76 Computer software
ISBN: 9781303222719
Date of First Compliant Deposit: 30 March 2016
Last Modified: 25 Jul 2022 16:14
URI: https://orca.cardiff.ac.uk/id/eprint/55108

Actions (repository staff only)

Edit Item Edit Item

Downloads

Downloads per month over past year

View more statistics